Connell-From Observables To Unobservables PDF

FROM 0BSERVABLES
TO UNOBSERVABLES
IN SCIENCE AND
PHILOSOPHY .
Richard J. Connell
University Press of America, Inc.

Lanham New York Oxford
Copyright 2000 by
University Press of America, Inc.
4720 Boston Way
Lanham, Maryland 20706
12 Hid's Copse Rd.

Cumnor Hill, Oxford OX2 911
All rights reserved

Printed in the United States of America
British Library Cataloging in Publication Information Available
Library of Congress Cataloging-in-Publication Data
Connell, Richard J.
From observables to unobservables in science and philosophy I
Richard 1. Connell.
p. em.
Includes bibliographical references and index.
I. Knowledge, Theory of. 2. Science-Philosophy. 3.
Empiricism. 4. Inference. 5. Concepts. I. Title.
BDI81.C667 2000 121--dc21 00-023498 CIP
ISBN 0-7618-1663-1 (cloth: alk. ppr.)

ISBN 0-7618-1664-X (pbk: alk. ppr.)
Q'"'
'V The paper used in this publication meets the minimum
requirements of American National Standard for Information
Sciences-Permanence of Paper for Printed Library Materials,
' ANSI 239.48-1984
Contents
Preface
Chapter 1 The Problem 1
Chapter 2 Observables-What They Are 7
Chapter 3 The Fundamental Cognitive Relation 21
Chapter 4 The Similarity-Dissimilarity Relation 33
Chapter 5 More on Analogies and Metaphors 43
Chapter 6 Signs, Symbols and Words 59
Chapter 7 The Necessity of Myth 81
Chapter 8 Theory-Laden Observation 99
Chapter 9 The A Priori and Innate Ideas 115
Chapter 10 Regularities, Experience, Experiment 121
Chapter 11 Hypotheses and Reality 133
Chapter 12 The Mathematical Character of Physics 145
Chapter 13 General Considerations on Philosophy 157
Chapter 14 Philosophical Principles: Preliminaries 177
Chapter 15 Natural Philosophy 185

Chapter 16 Definitions 195
Chapter 17 Analytic Propositions 205
Chapter 18 Recapitulation 225
Bibliography 227
Index 231
Preface
A while ago Richard Rorty said to philosophers 1 that they serve only
a very limited function in society. Because they have failed to live up to
their claims to solve systematically the problems that are concerned with
the various kinds of human knowledge (regarded, since Descartes, as
philosophy's proper domain), they are now left with the improvement of
human conversation as their sole function, which they exercise in part by
acting as gadflies, stimulating discussion to see whether various "theories"
"hang together." No one, Rorty says, ought to think that philosophy
actually understands anything about the real world, the world "out there."
Whether or not we agree with Rorty, we certainly must admit that
philosophy is experiencing a malaise, a sign of which is the way
philosophers speak. Somewhere Thomas More remarks that lawyers are
able to take a proposition with a very plain sense and after talking about
it for a time make it so obscure that no one knows any longer what it says.
And it would seem that lawyers are not a singular species in this regard,
for too many philosophers appear to share the same "talent," in evidence
of which one need only look at the journals, filled as they are with the
obscurity that comes from complex and prolix language, from the
multiplication of jargons, and from the use of words in senses removed
from those of ordinary speech. However, many problems arise not merely
from the obscure use of language but instead from philosophers such as
the deconstructionists who claim that words do not put us in touch with
an objective world, claiming that the only reality we have is that which we
form from language. This is indeed a sign of an unhappy condition of the
modem intelligence ..
When we read newspapers, magazines, and many books being
published today we can see, upon examination, that the authors do not
know what the mind is supposed to do and how it is to be governed. The
problem has many facets, not the least of which is the post-modernist-as
it is called-abandonment of objective sources and standards for the
mind's judgments and opinions. Of course there are serious philosophers
From Observables to Unobservables
who do not maintain that the only reality we know is the one we create
and who seek to discover how things really are. Some of them are in the
field of the philosophy of science, and a number of the topics we shall
take up here are of interest to them. Still, we cannot undertake to translate
their language and to expound upon their many expressions; nor can we
discuss even those topics they treat which stand in some relation to our
own. Were we to do so we would at the very least obscure our work. And
so we wish to say that we shall presuppose a realist position in philosophy
and then go on to consider in detail where the intelligence gets its starting
points in observation and how it moves from them. More specifically, our
basic issue is to show how the mind moves from sensible characteristics
to realities beyond sensation; that is, to show how the mind moves from
observables to unobservables not only in its basic operations but also how
in particular ways this movement is realized in systematic disciplines, that
is, in science and philosophy. We shall not take up here the realist anti-
realist debate, but instead we shall attempt to cast some light on realism
by the procedures we mentioned just above. As far as we know this has
not until now been done; so now let us continue.
The contrast between intellectual and sensory knowledge caused some
of the pre-Socratics to deny the reality of sensible properties. Later
Descartes put philosophy inside the mind when he started from his cogito,
in contrast to Hume who confined philosophy to observables-impre-
ssions of the senses-when he claimed that ideas, which presumably
belong to the intelligence, are but fainter copies of impressions,
precluding thereby the mind's coming to understand anything unob-
servable. Since that time philosophy has not been able to return to the
objective world, a world that has been abandoned to the natural sciences
and to a lesser extent to those that are social. As a consequence realist
philosophy, despite the statements of many who claim that the mind must
start from observation, does not seem to know where and how to begin;
and so as we said we propose to show how the mind comes to understand
unobservables from observables: that is the principal issue. Once we have
completed that task, and one we have talked about the sciences, we shall
say a little about what philosophy is and attempt to argue that it is first of
all a systematic discipline that bears on the real world and so depends
upon ordinary experience and observation both for its starting points and
for its measure. We then can describe philosophy in a general way as
systematically considered common experience, and it must of course
originate in observation.
Preface
At this point we shall quote two different sources outside philosophy

that differ from one another, each tending to show that currently many
philosophers do not appear to know how to be empirical. Our first source
tells us that
It was for many centuries a central belief that the whole world,
from the motion of falling bodies to the existence of God, could
ultimately be understood by pure thought alone. Thus philosophers
and theologians labored mightily to construct "proofs" of the
existence of God; thus thinkers such as Aristotle and Plato declared
that nature worked according to certain principles, and sought to
deduce from those principles the phenomena of the natural world,
relegating to a distant second place empirical tests of whether the
world behaved as their theories required it to. 2
Apparently Lindley thinks that before the days ofNewton academics paid
little attention to observation, to sensory data as the foundation for its
conceptions of nature. But few minds in the history of the human species
have been more empirical than Aristotle, however difficult that may be to
see. And despite some of his theories, many of Plato's dialogues have
solid anchorings in sensible data. But let us look at the Introduction in a
second source where the authors say the following:
Our chapters have, in the main, been constructed in such a way that
they can be read independently, and the notes and references are
collected together accordingly. Scientists with no interest in the
history of ideas can just skip the chapters in which they are
discussed. Likewise, non-scientists can avoid mathematics
altogether if they wish. One last word: the authors are cosmolo-
gists, not philosophers. This has one very important consequence
which the average reader should bear in mind. Whereas philoso-
phers and theologians appear to possess an emotional attachment
to their theories and ideas which requires them to believe them,
scientists tend to regard their ideas differently. They are interested
in formulating many logically consistent possibilities, leaving any
judgment regarding their truth to observation. 3
Philosophers, these cosmologists think, are attached to their theories on

the basis of emotions, which require the philosophers to believe-not know
From Observables to Unobservables
or see on the basis of evidence-what they hold. If that is true, then it is

difficult to understand how philosophers can be rational, much less
empirical. Thus our efforts to show how the mind starts from observables
would appear to be worth doing, especially since groundng philosophical
issues empirically is more difficult than reading the indicators of
measuring instruments. As the reader will know, the observational and
therefore empirical basis of the natural sciences consists in reading
measuring instruments, about which two observers reading the same
instrument will not disagree. But the observational, empirical starting
points of philosophy are more obscure, and one must ask, how does one
identify them? That, we think is a central difficulty from which much
confusion arises.
It would seem, too, that philosophy's language ought to be (relatively)
simple and straightforward, not obfuscating, a language which allows
ordinary experience to present its problems more clearly. Consequently
we have attempted to make our presentation plain and uncomplicated in
its terminology. Of course philosophy cannot exist without its own
vocabulary, but there is more than one way to bring it about. Our
procedure here will be to begin with the ordinary meanings convention
bestows on language, developing from these ordinary meanings-when
necessary-more technical terms for our philosophical use.
The reader will find that what we do here is relatively spare C but not
devoid c of references to other authors who have spoken on some of the
issues we discuss. The references are spare partly for economy but more
importantly because the issues we take up have not, as far as we know,
been treated in the same way and at the same length by any modern or
contemporary philosopher.
*****
As was suggested above, the present essay is divided into two main
parts. Part I deals with the way the mind knows unobservables through
observables. and this part is itself subdivided into two, the first of which
is concerned with the foundational issue, that is, with our mode of
knowing unobservables through observables, while the second is
concerned with some properties or consequences of that mode of
knowing. Certain characteristics of language, together with the genesis of
myth before science, are two consequences of the mind's having to know
unobservables through observables. Furthermore, during our discussion
of a fundamental property of language, we revive a position of Rudolf
Carnap's, making it our own but defending it on the basis of what we have
Preface
shown about the empirical origins of concepts.

Part II discusses regularities and their different roles in providing
principles or starting points for science and philosophy. More particularly
it shows first how the hypotheses-the principles-of the sciences are
related to observation, and then it goes on to show how philosophy
derives its principles, its starting-points, from ordinary experience. Such,
then, in the main is the essay we shall present here, outlined in the main.
Finally, we talk about theology and how it depends on observables and
why miracles are necessary.
*****
Nowhere, it seems, are men more socially dependent than in the life
of the intelligence, and we think it important to acknowledge our
dependence in a public way. Should the reader find the pages that follow
to be of some value to him, we would have him understand that we owe
much to the people with whom we work in the department of philosophy
at the University of St. Thomas, but especially to J.M. Hubbard, whose
careful criticisms and professional suggestions have saved us some
embarrassing missteps. We wish to acknowledge, too, all the other
members of Academe, in philosophy and in other disciplines, who must
go unnamed but whose writings and remarks have contributed to our
understanding in ways we cannot detail. Everyone is especially indebted
to his teachers, and in our case three especially played a major role: the
late Charles DeKoninck, the late Maurice Dionne, and the late Eugene
Babin, all of Laval University in the City of Quebec.
Needless to say, the faults are our own and ought not to be assigned to
anyone else.
University of St. Thomas

Richard J. Connell
St. Paul, Minnesota
Notes
I. Philosophy and the Mirror of Nature (Princeton University Press, 1979).

2. David Lindley, The End of Physics, Basic Books (New York: Harper Collins
Publishers, Inc., 1993), p. 8.
3. John D. Barrow, Frank J. Tipler, The Anthropic Cosmological Principle
(New York: Oxford University Press, 1986.)
I I
I
II
Chapter 1
The Problem
A. Introduction
For a realist observables are the foundation of all that we know about
unobservables, and our dependence on them is manifest in the external
expressions of the intelligence. Not surprisingly, then, observables and
their cognitive relation to unobservables have been topics of discussion
among philosophers. The problem that arises from this relation has been
posed in several ways: one can ask how we relate observables to
theoretical entities, or he can ask about the ground for the distinction
between observation and inference, or he can ask without qualification
how we know unobservables. But these different questions indicate that
the issues are not viewed in exactly the same way by all, a state of affairs
that requires us to present our own case and, we hope, with a minimum of
entanglements. As we said in our Preface, a treatment of every aspect of
the problem is impossible, and obtaining a commonly accepted set of
philosophical starting points for its resolution is at the very least difficult.
No one should find it strange, then, if we first delineate the starting points
that will guide our discussion, and to set the stage for that delineation let
us first contrast a philosophic position with the position of the ordinary
man. The contrast will be instructive because of the light it will cast on
our starting points.
B. A contrast
W. V. 0. Quine thinks himself an empiricist despite looking on physical

objects as cultural posits that are not definable in terms of experience:
As an empiricist I continue to think of the conceptual scheme of science

2 From Observables to Unobservables
as a tool, ultimately, for predicting future experience in the light of past

experience. Physical objects are conceptually imported into the situation
as convenient intermediaries-not by definition in terms of experience, but
simply as irreducible posits comparable, epistemologically, to the gods of
Homer. For my part I do, qua lay physicist, believe in physical objects and
not in Homer's gods; and I consider it a scientific error to believe
otherwise. But in point of epistemological footing the physical objects and
the gods differ only in degree and not in kind. Both sorts of entities enter
our conception only as cultural posits. The myth of physical objects is
epistemologically superior to most in that it has proved more efficacious
than other myths as a device for working a manageable structure into the
flux of experience. 1
To the ordinary man Quine's position appears strange, for the ordinary
man regards physical objects as real and inseparably related to experience:
physical objects are realities that are known through direct contact. For
him, to think of physical objects as conceptually postulated entities
differing only in degree from Homer's gods is equivalent to denying what
is most obvious, namely, the reality of the world "out there." But as the
reader knows, that is essentially what many philosophers question, if not
actually deny. They tend to spend their professional lives inside their
heads, lamenting their inability to escape.
Currently the word "empirical," which one might take to imply a
contact with extrinsic realities, is often employed to mean something that
is only psychological, something that appears internally but cannot be
regarded as putting us in contact with realities of any sort. Such a position
is either skeptical or solipsistic, and so it would seem that we may point
to two statements that represent the content of this position: ( 1) we know
sensory impressions, not things as they are "out there" (2) observables
(impressions) do not lead the mind to unobservable realities. The reader
will recognize Hume and Kant in these propositions, we shall,
nonetheless, present thumbnail sketches of what they say about
observables and unobservables.
C. Historical antecedents
Hume told us that ultimately all our ideas originate from impressions
of the senses, and Locke before him had held a similar view. Ideas, Hume
said, are but fainter copies of impressions, which means that our
understanding is limited by the capacity of things to produce sensory
The Problem 3
impressions. An object that cannot produce an impression cannot be

known, and we are under illusion if we think we have grounds for
asserting its existence. Thus if one is serious about such a position, he will
have to deny the reality ofunobservables.
Immanuel Kant, on the other hand, admitted that we know unobser-
vables, but his position involved a radical solution to the question about
how that occurs:
The peculiar characteristic of a science may consist of a simple difference

of object, or of the sources of knowledge, or of the kind of knowledge, or
perhaps of all three conjointly. On these, therefore, depends the idea of a
possible science and its territory. First, as concerns the sources of
metaphysical knowledge, its basic concept implies that they cannot be
empirical. Its principles (including not only its maxims but its basic
notions) must never be derived form experience. It must not be a physical
but metaphysical knowledge, namely, knowledge lying beyond experi-
ence. It can therefore have for its basis neither external experience, wich
is the source of physics proper, nor internal, which is the basis of
empirical psychology. It is therefore a priori knowledge, coming from the
pure understanding and pure reason. 2
Unlike Hume, Kant held that we do have knowledge that is not

empirical, knowledge that is essentially of unobservables and that he
called "metaphysical." Its principles, he said, cannot be derived from
experience but lie beyond it and are a priori, stemming from the pure
understanding and the pure reason. In other words, Kant took "metaphys-
ical" to describe the knowledge in the knower, not the realities known. In
contrast, the Greek founders of philosophy took "physical knowledge" to
be knowledge of physical, observable realities and "metaphysical
knowledge" to be knowledge of trans-physical realities, the phrases
indicating the kinds of realities known, not the character of the act of
knowing. Thus we see an "evolution" of the problem. The Greeks held
that we know unobservables through observables; Hume held that we had
ideas only of observables; Kant held that we knew unobservables but not
by starting from experience or observation. And so the direction of the
evolution is clear: from "out there" to "in me."
Earlier in our own century, the logical positivists endorsed a position
similar to Hume's, holding that meaning could be determined only
through observation; that is, they held that meaning could be determined
solely through the observables in terms of which something was defined
and that the observables themselves were the meaning. The positivists,
I 1
who had some sound philosophical principles, nonetheless effectively

denied the reality ofunobservables, a position they could not sustain. Not
all of the positivists were alike, however, for Rudolf Camap granted the
existence ofunobservables that had empirical manifestations. Currently
words that signify something beyond observation are not for that reason
denied meaning, though the problems bearing on how we know
unobservables have not disappeared. And so if we grant the reality of
unobservables, we may formulate the problem as follows: if the mind is
empirical, taking the origin and the standard of measure of what it knows
from experience, then how is it able to move beyond observables to
unobservables? In Hume's terms, how are "impressions" able to be the
source of concepts that are something other than washed-out copies of
sensory impressions that direct the mind to something beyond the
impressions themselves?
We must not overlook the fact that currently a number of philosophers
maintain that observation is theory-laden, retaining in some measure the
a priori that derives from Kant. And though these philosophers grant a
certain priority to observables, that priority is relative, for their position
makes observation to be theory-dependent, either (1) by claiming that
"observation sentences" are meaning-dependent and so therefore is
observation; or (2) by making the distinction between the observable and
the non-observable a matter oflinguistic convention. The text that follows
illustrates this last point;3
What is observable, in the pragmatist view, is what a community of

speakers and inquirers takes to be observable. The predicates appealed to
as the means of establishing whether truth conditions have been fulfilled,
as the empirical test or warrant of truth claims, therefore determine the
observation language for a framework of inquiry ... On such a relativistic
view of what constitutes "observable," and therefore what sorts of predi-
cates will be taken to name a statement as an observation statement,
observation and observability tum out to be framework-dependent. What
this comes to, in the way of a linguistic pragmatism, is that any predicates
whatever may be taken as basic predicates if, in some conceptual frame-
work or in some context of linguistic use, such predicates are taken as
uninterpreted for that framework. The one "real" framework of observa-
tion proper becomes steadily more elusive the further such an analysis
proceeds. For a logical reconstruction of what distinguishes observation
from interpretation, the so-called basic-predicates are merely those which
function as primitive terms or undefined terms, as the place where the
empirical buck-passing stops.
The Problem 5
It does not matter whether the view one holds is precisely that
described by W artofsky or whether it is a variant that shifts the ground
from language to concepts, making the business of being theory-laden
conceptual rather than linguistic. In either case, what counts as an
observable is a matter of choice, assuming that no claim is made about a
theory imposing itself necessarily upon the intelligence. Such positions
stand in plain contrast to those that regard "the boundary between what
can be observed and what must be inferred [to be] largely determined by
fixed, architectural features of an organism's sensory perceptual
psychology," as Jerry Fodor claims. 4 Moreover, the reader will recognize
that what we have said so far bids fair to introduce many philosophic
problems that have something to do with the issues related to foundatio-
nalism and the primitive "givens" of philosophical discourse. For the
moment, however, we shall pass them by, since our aim here is to find a
starting point that the reader will accept. By so doing, we shall be able to
leave many issues aside for a later time.
The starting point we think acceptable is the following: the recognized,
systematic disciplines regularly appeal to facts both to generate problems
and to measure proposed solutions. In other words,facts are commonly
admitted to have priority over both the interpretations made of them and
the inferences drawn from them, and it matters not whether the systematic
discipline is a natural science, a social science, or listing it separately,
history. Therefore we are justified in starting from the proposition
presented just above.
D. What will come
In the chapters of Part I we shall discuss the following main issues: (1)
how we know unobservables through observables (chapters 2-5); (2) what
some of the properties are of the intelligence that are the consequence of
its empirical character (chapters 6-7); (3) the independence of observation
on theory, that is, it is not theory-laden (chapter 8); (4) the impossibility
of innate ideas (chapter 9). In Part II, we discuss ( 1) the relation of science
to its principles; (2) the relation of philosophy to its principles and to
experience and observation. We need not say that a number of ancillary
issues will be discussed along with the main problems.
Notes
I. From a Logical Point ofView (Cambridge, Massachusetts: Harvard University
Press, 1953), p. 44. We wish to note here that Quine uses "myth" in the sense of
"fiction," and as we shall see later that is a distortion of its proper sense.
2. Prolegomenoa to Any Future Metaphysics The Library of Liberal Arts (New
York:The Bobbs-Merrill Company, Inc., I 950).
3.Marx W. Wartofsky, Conceptual Foundations of Scientific Thought (New York:
The Macmillan Company, 1968), p. I I 7.
4 "Observation Reconsidered," Philosophy of Science, 51, 1984) p. 25.
1.
:I
Chapter 2
Observables-What they are
A. Introduction
That the factual has cognitive priority over what is not factual is
admitted in every systematic discipline (except perhaps mathematics,
because of the latter's special character), and that provides us with an
appropriate starting point for our considerations on observables. It is
generally admitted that the "raw material" of scientific activity consists
of regularities that provide the facts relevant to the discipline's
investigative considerations. Ultimately the aim of the investigations is
either to interpret the facts or to infer something from them, an aim that
plainly grounds the observation/inference distinction; or as we prefer to
say, it grounds the observable/unobservable distinction. Stating the point
another way, every scientific, systematic discipline takes pains both to
discover its problems in the facts and to measure its solutions by them.
Thus it seems appropriate to begin our considerations by reviewing what
is meant by the word "fact."
B. The meaning of "fact"
Because every systematic discipline aims at understanding some

category of reality, it starts its investigations by fixing upon the facts and
subsequently either inferring something from them or interpreting them,
a point which we have already made. But though facts are related both to
inferring and interpreting, these two activities are not identical, and in
order to avoid any ambiguity we shall make their difference more evident
through some illustrations.
Were someone to look at the moon expecting it to be full only to
discover that a shadow covered it, he would very likely form a statement
such as "The moon is in a shadow." Having done that, he would then
8 Observables and Unobservables
probably draw the conclusion "Therefore an opaque body is interposed

between the sun and the moon." Further considerations allow him to
eliminate other bodies, and so he concludes that the earth is the body
interposed between the sun and the moon. Now this latter statement has
to be inferred as a deductive conclusion from the first statement because
it cannot be confirmed directly. That is to say, the observer, by merely
looking, cannot determine that the earth is between the sun and the moon.
We should note, however, that the proposition can function another way;
it can become a premiss of an argument that says "Because the earth is
interposed between the sun and the moon, the moon is in a shadow," and
though there is an inference involved in this movement of the mind, the
proposition that is the premiss is an interpretation or explanation of the
observed phenomenon represented in the conclusion. Under different
aspects, then, the statement "The moon is in a shadow" is both a premiss
and a conclusion. When it is a premiss, another proposition is inferred
from it, a proposition that is an interpretation of the "The moon is in a
shadow" when this latter is inferred from it. It can never be said, of
course, that "The moon is in a shadow" is an interpretation of the
statement that says "The earth is interposed between the sun and the
moon."
Because "The moon is in a shadow," signifies an observable effect,
the natural and ordinary procedure is first to infer the existence of the
cause from the empirical proposition, then after that is done it can be
restated as an explanatory premiss. Once the existence of the cause has
been inferred, we may use the causal statement as a premiss to explain the
fact, the empirical regularity, and in so doing the mind goes full circle,
though the procedure is not circular. It is not circular in a way that makes
it fallacious because the argument from the effect shows only that there
is such or such a cause, whereas the argument from the causal premiss
shows why the effect has to occur, given the cause. 1 Because these are two
different formalities or aspects under which the same proposition is
considered, the procedure is not circular in a sophistical manner. But let
us look at another illustration.
If a policeman were asked to narrate the facts of an automobile
accident, he probably would describe the position of the vehicles
involved, the dents in fenders, the black marks on the pavement, as well
as other observables he might detect. Were he to add, however, that the
accident happened because one driver was drunk, he would either have
inferred his statement from the data, or he would have introduced an
Observables-What They Are 9
interpretation, both ofwhich involve more than a reporting of facts. The

facts might indeed indicate drunkenness, but should the policeman say in
a court of law that the driver was drunk, he would be told that he had gone
beyond his function, beyond his competency as a witness. 2 Even had he
limited his remarks to saying "The driver had been drinking," he still
would have gone beyond the facts (beyond observation) by drawing a
conclusion from them. Thus grasping facts on the one hand and inferring
from them or interpreting them on the other are different activities. But let
us return to our main consideration and look at the principal dictionary
listings of "fact."
Originally "fact" came into English from the Latin "factum," which
means "... that which is done or made and has been achieved or finished."
Webster's (Second International) lists the English cognate as having a
sense similar to the Latin original: "A thing done, a deed." Things that
have been finished actually do exist; so we can see why Funk and
Wagnells' Standard College Dictionary lists as the frrstmeaning of"fact":
"Something that actually exists or has actually occurred; something
known by observation or experience to be true or real: a scientific fact."
For systematic considerations this is the most important sense the word
has because that which we observe or experience is known directly
insofar as nothing else stands between the knower and that which he
knows. (Here we use the word "know" in a generic sense to include
sensation as well as understanding.) Therefore in the considerations that
follow, "fact" will ordinarily signify: that which is observed. But some
other important senses must be noted too.
Webster's also tells us that "fact" can signify "An occurrence, quality
or relation the reality of which is manifest in experience or may be
inferred with certainty" [emphasis supplied]. The latter part of the
description provides an important new sense, namely, "... that reality
which can be inferred with certainty from what has been observed." In
short, this meaning of the term refers to something beyond the realm of
observation, though derived from it. At this point we ought to note that
statements which signify facts in either of the first two senses may
themselves be called "facts" or be described as "factual"; and care must
be taken to keep this meaning separated from the others. Statements are
linguistic expressions or conceptions of the mind and are therefore
different from the realities with which they put us in touch.
But the notion offact is not yet clear enough because we all know that
observation or experience can be twofold. Upon observing an external
reality, a shape, say, we become simultaneously aware not only of the

shape but of our act of observing it as well. At the very moment we are
aware of a color, shape, sound, motion, etc., we are also aware of our
activity of sensing one or the other of these properties. In a similar way,
when we infer the existence of something from a fact or facts we can
become aware not only of certain objects of thought but also ofthe act of
inference within us. Thus we have two kinds of awareness, and we know
two kinds of things at once: an external reality, and our own mental and
emotional states. The first of these two different ways of being aware we
call "external observation" or "external experience," and the second we
call "internal observation or "internal experience," referring thereby to the
operations themselves that are within us. We now have one more meaning
of "fact" according to which our own cognitive and emotional states are
private facts to which we alone have direct access.
If someone objects that the cognitive states within us often present
illusions and therefore ought not to be regarded as facts, we may reply
that although the content of our thoughts and imaginings (a leprechaun,
say) or the object of our emotions (a fear of dying of spinach) may be
fanciful, the cognitive and affective states as properties of the subject are
real, even though their content or object is not. And it is upon the former
that internal observation bears insofar as it reveals private facts to us. For
that reason it does provide another sense of the word.
To summarize. In the foregoing paragraphs we have delineated and
ordered a few notions that are known by everyone because they are part
of ordinary experience. We noted that systematic enterprises center on
facts, which, in the primary sense of the term, are external observables.
The systematic, scientific process consists, most broadly described, either
in drawing inferences from facts or in explaining them, and in this way
facts are both the origin of our investigations and their final measure.
Facts are not to be identified either with inferences from them or with
interpretations of them. The conventional signification of "fact" is
confined in its primary sense to something observable, while in its second
sense "fact" includes what is inferred with certainty from an observable.
Neither sense of "fact" includes the act of interpreting. But now we turn
to ordinary experience to show that we do know some unobservables.
I!
C. We do know some unobservables
There is no dpubt that systematic investigations (not to mention

Observables-What They Are II
ordinary conversation) concern themselves with many things that are

unobservable. When we consider the kinds of natural phenomena the
natural sciences investigate, we discover that ordinarily the causes which
account for the observable regularities that are the object of concern are
unobservable. Falling bodies, the movements of planets, stars, galaxies
and the like are explained by an unobservable gravity. Electrical
attractions and repulsions, diffusions of gases, propagation of sound over
distances, chemical reactions, osmotic diffusion of liquids through
membranes, growth of plants, repairing of wounds in animals, and many
other events are all accounted for by causes that escape observation, either
because of circumstances or because they are unobservable in principle.
By cause we mean that which is responsible for the existence or occur-
rence ofanother, and all the illustrations above show events that require
something beyond themselves that are responsible for them. And things
as well as phenomena have unobservable causes; for example, the active
source which brings about the generation of one organism from another.
Moreover, the description given above of cause, though general, is clear
enough to indicate the sort of reality to which the term "cause" is applied;
and we can see that in a way unobservables are more substantive to the
natural sciences than the observables from which the sciences originate.
We are not saying anything new when we note that the behavioral
sciences are also immersed in externally unobservable realities. Unless
one thinks he can throw into the intellectual wastebasket things such as
attitudes, prejudices, emotions, habits, skills, and the like, hardly an
explanation of human behavior can be found that does not involve
unobservables. In the past some philosophers and some psychologists
have chosen to ignore the interior, unobservable determinants of human
actions, putting aside human introspection, even to the point of
disregarding the very existence of these (externally) unobservable states
or activities. At present, however, it is no longer gauche to accept the
existence of these externally unobservable determinants of human
behavior. And so on the basis of inductive procedures, we see that some
realities are commonly admitted to be observable, while others are
commonly admitted to be unobservable; and that is where we must begin.
D. The source of our knowledge of unobservables
When we reflect, we readily see that the unobservables the sciences

talk about are known through their observable manifestations, that is,
through facts. Electrical currents are known through their heating and
mechanical movements; magnetic fields are recognized through
deflections of compass needles and other similar motions; a disturbance
in a planet's orbital revolution tells us of an unseen source of gravity; a
fast pulse indicates excitement or perhaps an infection; insulting speech
a hostile attitude or prejudice, and so it goes. The point, then, seems
evident enough: observables provide evidence for unobservables, and that
is what we mean when we say that unobservables are known through their
observable manifestations.
Only for the sake of completeness need we mention that the natural
and social sciences are not the only enterprises to use observables for
manifesting unobservables. Imaginative literature uses metaphors and
symbols made from them to tell about human plots and emotions; myths
employ images to represent unsystematized ideas about the origins of
things; parables tell about life beyond the domain of experience. In sum,
the human mind universally erects its unobservable edifices on observable
foundations.
But not all talk about unobservables is regarded as respectable.
Philosophy, it seems, tends to specialize in unobservables, sometimes not
making much sense, so that early in this century some philosophers
denounced a large part of the philosophic enterprise as fictitious.
Metaphysics, they told us, was nonsense; and they were inspired largely
by the unfounded, unverifiable statements that some practitioners of
philosophy made. And though at present this view is not widely accepted,
it does signal a difficulty that comes from philosophy's preoccupation
with unobservables, especially those removed from observable
manifestations. It behooves philosophers, therefore, to pay very careful
attention to their empirical starting points, as well as to the procedures by
which they argue from them. More particularly, the empirical starting
points of philosophy are obtained from words that signify either
observable properties or things endowed with such properties. But our
point needs expansion:
'I I
I, It was not an empty remark RudolfCamap made some years ago when
he wrote the following words:
ll1
I
In its mythological use the word [God] has a clear meaning. It, or parallel
words in other languages, is sometimes used to denote physical beings
which are enthroned on Mount Olympus, in Heaven or in Hades, and
which are endowed with power, wisdom, goodness and happiness to a
greater or lesser extent. Sometimes the word also refers to spiritual beings
which, indeed, do not have manlike bodies, yet manifest themselves

nevertheless somehow in the things or processes of the visible world and
are therefore empirically verifiable ... .In its metaphysical use, on the
other hand, the word is deliberately divested of its reference to a physical
being or to a spiritual being that is immanent in the physical. 3
We do not intend to bring back all that logical positivism endorsed, but
it is important to stress that the lines cited above do make a fundamental
point. No attempt will be made here to give an exegesis of Camap's
positions; instead we shall note only that what he says above is
legitimately interpreted to correspond with the regularity which we have
been at pains to describe, namely, that reflection shows us to know
unobservables through observables. Given that regularity, one has reason
to say that the name "God" is truly meaningless if it is divested of every
and any reference however indirect to observable manifestations. His
name cannot refer to something that is in all ways beyond experience; but
this is not to say that observation of physical manifestations is always
adequate to showing the full character of an unobservable entity, even of
many physical causes, much less God. Indeed our knowledge of
unobservables suffers limitations, a point about which more will be said
later.
We may now consider our point to be adequately established, namely,
that we know unobservables through the medium of observables, as a
result of which we may ask a very fundamental question: what is an
observable?
E. A definition of observable
Currently the distinction between observable and unobservable

realities is more frequently, if not always, discussed as a distinction
between observational and theoretical terms; and as one author says:
Those who have defended an 0/T distinction have sought to defend it as

representing a difference in kind and not a difference in degree .... In
what follows we first establish that no distinction in kind can be drawn
between 0-terms and T-terms. None the Jess, a rough pragmatic
distinction of degree can be established and that, it turns out, is all we
need. 4
According to Newton-Smith, there is no distinction in kind between

observable and unobservable, only a difference of degree. Furthermore,
Newton-Smith tells us (with an appropriate reference) that according to

Camap an observation term corresponds to an observable quality, and he
quotes Hempel as saying that observables are indications on measuring
instruments, colors, odors, etc. But then he goes on to add:
This characterization is entirely unenlightening unless it is supplemented

by an account of what it is that makes a quality an observable one. 5
And that, he thinks, cannot be done. In his view, definitions of observable

fail. For instance, if one were to define an observable quality as "one
whose presence or absence can be detected by using our actual perceptual
faculties unaided by instruments," we encounter difficulties, one of which
becomes apparent in the example of force. We cannot always detect
forces by feeling them, and if we require that observable qualities always
be detectable, then even color-predicates are not observable. That is the
essence of his objection.
It hinges, however, on an equivocation; for some qualities are
"detectable" (observable or sensible) by reason of what they are intrinsi-
cally, in their own character, in which case all instances of the qualities
are detectable. The opposite of this kind of detectability is the
undetectable that is not intrinsically and by its own character able to be
detected or sensed, in which case every instance of such a quality is
undetectable. On the other hand, something can be detectable because of
favorable circumstance accompanying the instances and undetectable
because of the lack of them. For example, I cannot detect the color of
something in a box when the box is closed, nor can I detect something
that is too small to be seen with the naked eye, though it is visible through
a magnifying glass or a microscope. As for a force that I cannot sense, the
equivocation at issue has to do with the fact that what I sense in such a
case is the effect ofthe force on me, not the force itself. Thus with such a
distinction in mind, it becomes plain that a proper definition requires an
observable to be detectable only in the first sense of the term, for a
general defmition pertains to the common nature but not to the circum-
stances of individual instances. Furthermore, the relation that instruments
have to observables is of secondary concern, for obviously observables do
not depend on instruments for their general nature or for their function as
the source of our knowledge of regularities. Furthermore, if we see with
an instrument, a microscope say, we still see; that is we detect what is
under the microscopic insofar as it is colored. Discussion of such issues,
then, need not concern us here.

Newton-Smith also claims that "force" is a T-term and yet is
observable. And to the objection which says that we observe the effects
of the force and not the force itself, the latter being inferred from what we
feel, he replies that "the distinction between directly sensing the presence
of something and inferring it on the basis of its effects is problematic."
But apart from the fact that his statement is wholly gratuitous, it begs the
question; for if observable is defmable, then the defmition has to serve as
a basis for determining what is directly sensed and what is inferred.
Another example he proposes that is supposed to tell against the
definability of observable is the gene, which originally was a purely
theoretical entity and which now, he says, has been observed. But granted
that something may have been seen with instruments, the issue is not
thereby settled; for whether or not what was observed was the gene itself
is still the question. 6
Having made these introductory remarks, let us now approach the
issue by first noting that there are certain properties which everyone
admits are sensed and therefore observable, no matter what theoretical
position on observables and unobservables he may hold, and no matter
what his position on "where" observables exist. Historically, philosophers
have recognized two classes of observable properties; and over the
centuries the two classes have acquired two sets of names. The Greeks
called one class "proper sensibles," which includes color-light, sound,
odor, flavor, and the principal subdivisions of tangible qualities: hot-cold,
wet-dry, hard-soft, heavy-light. All of these properties are stimuli,
adequate stimuli, as they are sometimes called; and each one is related to
a specifically different sensory apparatus: light-color to sight, sound to
hearing, etc. Since the time ofGalileo, however, such properties have had
another name, which is "secondary qualities."
The Greeks named the second class of observable properties "common
sensibles" because each ofthem could be known by more than one sense.
They include size (magnitude, extension), shape, motion, rest, and
multitude (of physical things). To them we may add time, location
(sometimes called position) and (bodily) position, such as sitting,
standing, etc., as secondary (for reasons we shall give later) members of
the genus. The more modem name for the principal members of this class
is "primary qualities." Now the official data of physics and chemistry
belong to the second category; that is, the sensibles which provide the raw
data of the natural sciences are first of all coincidences of magnitudes the
magnitude of the measured, and the magnitude of the measure. 7 For

example, the length of a column of mercury alongside the scale on the
thermometer is such a coincidence, the length of the mercury being the
measured magnitude, the scale the measuring. Similarly, the deflection of
a needle through an arc that indicates voltage or current is the coincidence
of the arc of the needle, which is a magnitude, with the scale on the dial.
Of course digital counters have replaced many such instruments, the
number displayed on the counter functioning as a substitute for the scale.
But whether the number is obtained from a scale or a digital counter, the
official data-of physics especially-measure degrees of "intensity," the
more or less of some physical qualities or substances. Needless to say, the
instruments can also represent multitudes of entities that are counted. But
let us now take up our definition.
Defming observable requires us to note first that experience reveals
that the relation of the senses to the realities which they detect is a relation
of a passive (in a certain manner) recipient to an active stimulus. As the
word itself indicates, a stimulus is an agent that is necessary for a sensory
apparatus to perform its function, which is not to say that a stimulus
produces the operational capacity in the organism. That clearly is not the
case, since the active ability to sense stems from the constitution of the
organism. We see, then, that in different ways the sensory apparatus must
be both active and passive, its passivity belonging to it in relation to the
stimulus-object. The nature of this passivity becomes more apparent when
we consider in a more particular way the relation of the senses to the
realities-the objects-they know. 8 Specifically, no animal can control
what it sees, hears, smells, etc., except by moving its attention from one
object to another. Because its environment is already there and must be
taken as it is, an animal can control what it knows through its sensations
only by shifting its attention from one reality to another. Hence, we see
that our sensory response is determined by the object-property; in other
words, the character, the nature of the sensation is brought about by the
stimulus, that is, the object-property. The sensation of seeing red is not the
same as seeing blue; the sensation of tasting something bitter is not the
same as tasting something sour; hearing high C is not the same as hearing
low C. In short, sensations differ qualitatively from one another, and their
character or nature is determined by the action of the stimulus-object. To
repeat: under normal, natural circumstances, whether we see red, hear low
C, taste a sweet flavor, feel a moving object, see a spherical shape, etc.,
depends on the stimulus. In sum, the application of the sensory power to
an object is actively determined by the animal, but the nature of its

response is not.
We may now state our definition more formally. A sensible trait or
property, an observable, is that which determines the character (the
quality, the nature) of a sensation. And let it be noted that the definition
reports what experience justifies, for we see that observable and stimulus
do go together; that is, every observable is a stimulus and every (sensory)
stimulus is an observable. 9 Furthermore, because observable is defmed by
its agency, by its character as a principle rather than by its character as a
term, we have not begged the question. And though the defmition is a
simple and obvious one, it does tell us how to distinguish observable
properties from those that are unobservable. Let us illustrate.
Everyone is aware that red apples and green apples give rise to
different sensations, as do high notes and low notes, dry shoes and wet
shoes, hot and cold pans, cubical and spherical bodies, what is moving,
what is at rest, etc. On the other hand, the non-violinist who becomes a
violinist by acquiring a real skill does not produce different sensory
impressions before and after he has acquired his art; so the skill of
violin-playing, although very real, is unobservable. (The sounds that issue
from the violin before and after are of course observably different.) Two
pieces of iron, one magnetic and the other non-magnetic, make the same
impression on the power of sight, hearing, etc. Stated another way, a piece
of non-magnetic iron does not later introduce any difference or variation
into a sensation as a consequence of its being magnetized. But we need
not multiply our illustrations, for we are primarily concerned with the
general nature of the observable and its opposite, the unobservable.
It ought now to be plain that the distinction between observable and
unobservable is not merely a difference of degree, for there are many
properties that are incapable in principle, under any circumstance or set
of conditions, to stimulate the senses. No one will ever observe a
magnetic field without first acquiring a sensory capacity the human
species does not now have. Nor is there anything inappropriate about
limiting the term "observation" to the exterior sensory capacities, for that
is what the common language first means by the word. An internal
sensory power, one that bears directly on sensory states rather than on
properties of things, is obviously of a different nature; and "observation"
as applied to its activity takes on an extended meaning. Furthermore,
"inference," which is well understood to belong to conclusions generated
from premisses, obviously does not occur in the power of sight or hearing
I I
. II'
II
I
or taste, etc., since the senses do not give rise within themselves to
movements that go from one property to another; or to put it another way,
they do not form propositions. There is, then, nothing in the least
"problematic" about the distinction between observable and unobservable,
and the failure to keep it in mind cannot but produce confusion. (The
principle of distinction employed above obviously starts from the
proposition that different passivities require different agencies to actuate
or actualize them. An ability to be magnetized must be actualized by an
actually existing magnetic field, which is to say that ability and agency
must be proportioned to one another. 10)
F. Asummary
Our procedure in this chapter consisted in first pointing out certain

traits that are commonly admitted as being observable or sensible, after
which we noted that the senses are passive in relation to their stimuli,
which are what they also know. Having obtained these starting points, we
then arrived at the definition of an observable as that which determines
the character or quality of a sensation, from which it becomes plain that
other cognitive capacities are involved when we interpret the data of
experience. The number and character of these capacities is not, however,
a matter to be considered here since it is unnecessary for what we are
doing.
Notes
!.The conclusion-become-premiss is not an hypothesis, an issue we shall discuss

at length in Chapter 17.
2. Smelling alcohol on a driver's breath and a subsequent blood-alcohol test are
facts that alii ow a more convincing conclusion.
3. "The Elimination ofMetaphysics Through Logical Analysis of Language," in
Logical Positivism, ed. A. J. Ayer (New York: The Free Press, 1959), p. 66.
4. W. H. Newton-Smith, The Rationality ofScience (Boston: Routledge & Kegan
Paul, 1981 ), p. 23.
5. Ibid., p. 24.
6. This example, however, depends upon the distinction between substance and
property, which we consider in Substance and Modern Science (Notre Dame,
Indiana: University of Notre Dame Press, 1988) and we presuppose those
considerations here. Thus, we may postpone for a while the discussion of the
observability of substances.
7. For those who may not be familiar with this point, see Sir Arthur Eddington,
New Pathways in Science, (Cambridge: At the University Press, 194 7), c. I,
section III.
8. Hereafter we shall call that which is attained by any operational power its
"object"-that will be the meaning of the term.
9. It is important to note that an observable property-color, say-must not be
confused with an emotional stimulus. In one person the olor red may recall the
very pleasing scent of roses, whereas in another it may recall the horror of are
different and hence not the result of the color alone.
10. This principal blood from a serious wound. Clearly the emotional responses
e will be discussed at length in a later chapter.
1'
--~------------------------------------------------------ ---
I"
.\.
j
;.
Chapter 3
The Fundamental Cognitive Relation
A. The precise formulation of the problem
Though observables provide our only evidence forunobservables, they

do not always point to them. Compass needles do not always tell of
nearby magnetic fields, body temperatures do not always indicate
infections, facial color does not always suggest strong emotion,
expression does not always reveal attitude. Hence if observables
sometimes give rise to an understanding ofunobservables and sometimes
do not, then on those occasions when they do, they must modify the
sensory reaction differently; only something extrinsic to observables
themselves can account for the differences, which means that the question
now is: precisely what do we detect in an observable when it modifies the
sense power so as to point to an unobservable?
The answer is evident enough, for if an observable is to affect the
sense one way at one time and another way at another, then there must be
a difference in the sensible trait itselfat the different times. That is to say,
the manifestation of the unobservable by the observable comes about
because of a modification in the observable that points to the
unobservable. Furthermore, the possible differences or modifications are
only three: (1) the observable has undergone some variation,
intensification, or attenuation; (2) the observable has just come into
existence; (3) the observable has just gone out of existence. (This last
alternative applies to those properties that are stable and are not constantly
attenuated all the while they exist. An example of the latter is a hot body
that gradually cools by radiation.) In short, a change in the observable is
the evidence for the existence of the unobservable, as for example a rise
in temperature indicates an infection, a shift of the compass needle from
its normal position indicates a magnetic field, cessation of heartbeat tells
of death. Some time ago a deviation from the predicted orbital motion of
Uranus led to the discovery of the planet Neptune, and a similar

discrepancy in the motion of Neptune led to Pluto. A flow of particles
against a concentration gradient 1 led biologists to find a transport system
in the membranes of cells; and so we might continue. Wherever we look,
we see that observables are manifestations of unobservables only when
the observables are modified in one of the three ways listed above.
In addition to what happens to a single trait, a variation in a correlation
of several sensibles can also point to an unobservable; and this sort of
indicator will be instanced in the following chapter. But not only are
unobservable realities known through observables, so are non-entities,
absences, negations. For instance, the mind's knowing non-red, non-sphe-
rical, darkness, and silence depends upon its first knowing red, spherical,
light, and sound. Before continuing, however, we must examine the
different kinds of opposites, for they are not all the same and they affect
the senses differently.
The first kind of opposite is that which is most obvious, namely
contradiction, and it is represented by an affirmative term and its
negation, as for example, man and non-man, sighted and non-sighted,
skilled and non-skilled. Here the negative term of each pair removes
entirely that which is signified by the affirmative term, and therefore it,
the negative term, can be predicated of anything that is different from that
which is signified by the affirmative term. A rock is non-man, non-sight-
ed, non-skilled; so is the number two. Clearly this kind of "contradictory
opposition" is an opposition of realties, of things "out there"; it is not an
opposition of propositions, which are either conceptual entities in the
mind or linguistic signs of those entities. Contradictories in reality are,
however, the opposites on which contradictory statements or concepts
depend.
A second kind of opposition we call "privative opposition." Here one
term signifies something affirmative or positive while the other signifies
its absence, and to that extent this kind resembles contradictory
opposition. But a privative term can be predicated only of the same kind
of subject as the affirmative term that is paired with it. In other words, a
privative name signifies a negation or absence in a definite, determinate
subject, one example of which is blind and another unskilled. Neither of
these removes entirely what is signified by its corresponding affirmative
term. Blind can properly be said only of those animals which are endowed
by nature with a power of sight (however much it may be used
metaphorically of other things), and unskilled can be said only of humans.
The Fundamental Cognitive Relation 23
We may not say that stones are blind or that trees are unskilled, although
we may say that stones are non-sighted and trees are non-skilled. Thus the
negative term in privative opposition can be said only of those things that
ordinarily are endowed with that which the affirmative term signifies at
a time when the latter can be expected to be present. We do not consider
newly born puppies to be blind, for they are not mature enough to
exercise the operation of seeing. Here again, then, differences in reality
are responsible for the differences of signification between contradictory
and privative opposition.
There is another kind of opposition in which the terms are opposed as
extremes within a category of positive realities. For instance, to succeed
and to fail divide a common category as extremes, for both success and
failure are kinds of activity; one cannot fail if he does not try. Kind and
cruel are opposites of this sort too, since a cruel act is not the same as no
act. Such opposites, because both are within the same genus of reality, are
related as the perfect to the imperfect; that is, one always lacks something
of that which the other possesses, which means that the differences that
define each are related as affirmative and negative. One difference will
remove what the other posits; one opposite will explicitly deny a part of
what is contained in the notion of the other. Hard and soft are examples
of this sort of opposition, for a soft body differs from a hard one in
offering minimal resistance to a penetrating agent yet the soft body does
not lack all resistance to such an agent. Opposition of this sort is called
"contrary opposition," and like the others, it, too, reflects differences in
realities.
The last kind of opposition involves pairs of terms of which one refers
to the other in its understanding or definition and both of which are
positive in character; for instance, teacher implies student and father or
mother implies child. Similarly, whole implies part and ruler the ruled.
Now although both of these terms are positive, pairs of terms of this
description are not said of the same thing at the same time or in the same
respect and for that reason are included under the notion of opposition;
but it is important to emphasize that neither term is negative, neither
implies the absence of the other reality but instead requires its
Eimultaneous existence, each in its own entity. 2
Now the point of importance in this discussion of opposites is that
knowing one of the two precedes knowing the other. To illustrate: the
mind's knowing non-red, non-spherical, darkness, and silence depends
upon its first knowing red, spherical, light, and sound. Even when one of
two opposites is privative and not purely an absence, the

negation-containing part must still be known intellectually through the
other to which it is correlated. If someone goes into a dark closet, he must
shine a light in order to see both white and black billiard balls, which
means that both colors are positive properties. The quality of a surface
which is called "black" is truly a color and must be defmed using color
(something positive) as its genus or category. Yet a black surface reflects
much less light than a white one, and the definition of the color black will
contain within it a negation-either wholly or partially of that which
formal to color-the reflection of light. In a similar way ,failure must be
defined as an act (something positive) that does not attain its goal
(something negative). Darkness, on the other hand, is defined as an
absence of light and silence as an absence of sound; hence the category
of each of these is negation. (The difference between privation and
negation is important and should be kept in mind.)
That the mind must define one opposite through the other when one
of them is or contains an absence of the other is not hard to see; for in the
measure something is a non-entity or absence, it is incapable ofaffecting
the senses. 3 We are aware of absences by trying to see, trying to hear, etc.;
and because only actualities can be stimulating agents, negations can be
understood only through the medium of positive properties that can
function as stimuli. This explains, too, why we sometimes confuse
non-beings with realities, for though the concept-content of a negation is
distinct from the concept-content of something positive, the mode in
which the mind conceives negations is the same as the mode in which it
conceives positive realities. When we say "Socrates is blind," and
"Socrates is sighted," we understand blindness in the same mode in which
we understand sightedness, that is, in the mode of a common attribute.
Again: the way in which blindness signifies is the same as the way in
which sightedness signifies, and that is our point. We do not understand
blindness to be an existing, real positive common attribute as we do
sightedness. Put another way, negations and privations are not existing
attributes common to many things, though their mode of presentation in
the intelligence is the same as that of positive, existing attributes. Here we
should note that some philosophers, confusing the mode of conceiving
with the content of conceiving, have made much of "nothing. "4 But
having said these things, we must now turn to an important principle that
is buried in what we have said about knowing unobservables through
observables.
B. An unexpressed premiss
Whenever we move from an observable to an unobservable, we reason

from the one reality to the other; and the inference we make is founded
upon our understanding of a proposition that we do not at the time
articulate to ourselves. That is to say, a variation in an observable leads us
to an unobservable precisely because we spontaneously understand that
the observable property cannot vary itself, bring itself into existence, or
take itself out of existence. (To maintain that it could come to be by itself
would be to maintain that non-existence as such can become something
existing.) Because properties cannot modify themselves they must be
modified by something else; and every human being who has had the
normal use of his intelligence throughout a significant part of his life,
whether he has lived in a sophisticated or a primitive society, has inferred
the existence of unobservables on such grounds. Indeed, without our
accepting, either implicitly or explicitly, that there is a cause-effect
relation between the observable and unobservable, we could never move
beyond the observables themselves. All our understanding would be
Humean. However, that unexpressed premiss is always in the background,
and it would seem to tell us that the cause-effect relation is truly the first
unobservable the mind spontaneously attains, on account of which it plays
a primordial role in our learning procedures. But this remark shows that
a general question about the observability or unobservability of relations,
of which cause and effect are only an instance, necessarily arises; so let
us consider this question.
If a nail is placed near an electromagnet when the power to the magnet
is turned off, then merely by watching (or listening to, or feeling, etc.) the
nail an observer cannot tell when the power goes on and the nail receives
a magnetic impression. Since we cannot detect the coming to be of the
magnetic field by sensing, it is unobservable and the cause-effect relation
between magnet and nail is too. Speaking more generally, because a
relation is a proportion of some kind of one reality to another, the relation
itself is not observable, even though the realities that are referred to one
another are themselves observable. To repeat: no relation as such modifies
the sensory apparatus, though many properties upon which relations are
founded do modify it. Even when a cold body is heated by being put in
contact with a warm one, we observe only the hot and cold qualities in the
bodies, together with the warming of the one and the cooling of the other.
Had one warmed the other and then been observed by us only at
equilibrium, we could not, as with the magnet, tell which was cause and
which effect. Thus, to know one reality as cause and another as effect
requires that the mind go beyond observation.
The relations involved in location, bodily position, and time are
perhaps less clearly unobservable, and in the preceding chapter we did
classify them secondarily as common sensibles. But when we locate a
body, we do so by comparing it to another body that serves as a point of
reference. For instance, when we say that the chair is next to the table, the
tree is north of the house, the lamp is on the desk, etc., we describe one
body by comparing it to another to which we give some temporary or
permanent priority; and the reference of the one to the other is the
relation. Furthermore, because the relation requires certain real (and in
this case observable) properties and proportions in the things themselves,
it clearly is not a fiction, though it clearly is a "minimal reality."
Bodily position is like location in regard to the relations it involves. If
we were to say that a man or a horse or some other organism is sitting or
standing or lying down and then were to describe one or the other of these
positions, we could not escape referring some parts of the organism's
body to others. (A perfectly homogeneous body cannot have the
characteristic of bodily position as do organisms with clearly
differentiated appendages.) The sizes and shapes of the parts are
observable, of course, and they are the foundations of the relations of
position. Nonetheless, the relations themselves do not variegate the
sensory response. Similar remarks can be made about time, which because
it involves measurement, implies a relation of measure to measured. But
some arguments will help make our case.
The reader will probably concede that relations are indivisible, for we
cannot take away from them some part or fraction; either the "whole"
relation exists or nothing. For instance, we cannot take away a part of the
relation next to or north of Yet if we consider a visual field made up of
a chair next to a table, the visual field can be reduced by a part. If the
chair is removed, only a part of the field is gone, but in removing the chair
the entire relation of next to has been done away with. Certainly we
cannot explain the relation's removal by claiming that the whole of it
belongs to the chair; for that would mean the table contributed nothing,
which is clearly not the case. For the moment, however, let us assume that
the relations mentioned above are observable, and if they are, they can be
expected to be visible. The sense of sight attains all the common
sensibles, and that is what such a relation must be assumed to be. 5 It then
follows that when two objects are placed next to one another, the relation
they acquire is a new observable property which must affect the senses as
long as the objects are next to one another; and when they are not next to
one another, then the observable property ceases to be. Stated more
precisely, since the relation of next to belongs to both objects and not to
one alone, each related object must affect the sense differently than it does
when one is not next to the other.
Let us now assume that the objects possessing the next to relation are
a table and a chair and that a screen is put in front of the chair so that we
no longer see it. The screen does not destroy the relation; therefore it
continues to exist. But if someone unseen by us removes the chair, the
relation will be destroyed; and thereafter the table should affect our visual
sense differently than it did when the chair was next to it, assuming, as we
did, that the relation is observable. But plainly the table does not affect
our sense differently, and observation will not tell us whether the chair
has been removed or not; in this respect the relation is not different from
that of cause-effect. So however observable the shapes and dimensions
upon which spatial relations are founded, we have to say that the relations
themselves are not. Again, a visual or other sensory field may contain a
multiplicity of observables that are foundations for relations, but the
relations themselves do not affect the senses. On the other hand, location,
position, and time are relations that are so close to observation by reason
of the sensible properties upon which they are founded that we often
speak of them as ifthey were observable. So it would seem that although
the linguistic convention ought to be recognized, such things as time,
position, and location are not observable in the strict sense of the term.
We have argued at length in another essay6 that, contrary to Hume,
substance is real and that it does indeed function as a substratum for
properties, though it is certainly not an "unknowable somewhat." We must
concede, however, that substances are unobservable, which means that
they too must be known through observables. Words such as "gas,"
"mineral," "fluid," "water," "gold," "electron," etc., signify stuffs or
substances; and when we distinguish one from the other, we do so through
their observable manifestations, that is, through their observable
properties or effects. From a distance a jar may appear to contain water,
and that might be what the stuff is; yet upon coming closer and smelling
its odor- and tasting its flavor-which also is not the stuff, we discover
an odor and a flavor that we do not associate with water but with alcohol.
Hence, if water is one stuff and alcohol another, we see very readily that
we do not sense or observe either; for if we did, we would have no need

of the properties to identify them, and the substances themselves would
modify the sensory reactions. From this we also see why we sometimes
err in regard to judgments about stuffs, for we make precipitous
inferences from inadequate observations. Neophyte prospectors have been
deceived by fool's gold because the observable properties common to the
true and the false stuffs occasioned an erroneous inference from their
inadequate observations. In the essay to which we referred above (note 6),
"alive" is shown to describe a category of substances; yet once again the
substance itself does not introduce a difference into sensory reactions.
When we determine whether a man, say, is alive, we do so on the basis of
heart beat, breathing motion, stiffness of limbs, temperature, etc., all of
which are observable properties. Of course many of them can be absent
for a time without our knowing whether the organism is alive or dead, and
mistakes about such states occur with some frequency when the "vital
signs," as we call them, are neither clearly present nor clearly absent.
Hence we repeat: an attribute which describes the interior of a substance
is unobservable and therefore intellectually knowable only through
observable manifestations.
C. Observables and necessary propositions
It seems appropriate to digress for a moment and comment on

Immanuel Kant's position on necessity in propositions. As with Hume
before him, Kant could not reconcile the necessity he attributed to certain
propositions with the obvious contingency of sensory data. He sought to
solve his problem by maintaining that the intelligence itself contributed
the necessity. Certain propositions, he said, were a_.priori in the sense that
they issued from the mind itself. That was how he attempted to solve his
problem.
But one has no need for such an enterprise. If we consider that natural
entities begin to exist after first not existing, we realize that they can come
to be only through the activity of a cause proportioned to them. Such an
active cause has a necessary relation to the coming to be of the effect in
the sense that if the effect is to come to be, then necessarily the active
cause must function. Now such necessity belongs to real things, and it
clearly justifies the mind's forming necessary propositions to signify it. In
short, whatever has come to be in nature came to be necessarily through
the action of an agent cause. Now although our words do not constitute
anything like a full discussion of Kant's position, they do suffice to show

that necessity in propositions can come from real things through our
understanding of contingent observable entities.
We perhaps should say too that the difficulty the pre-Socrates had in
recognizing that sensible properties are more than appearances might very
well be explained on the same grounds. What existed for them was a
fundamental reality-substance or being-that allowed no room for change
and hence for the reality of properties. In their minds what the intelligence
confronted took priority over the data of the senses, whereas if they had
understood the dependence of one on the other perhaps their positions
would have been different.
D. Observables and ordinary language
In the strictest sense of the term only certain properties are observable;
yet a broader sense is commonly employed, and we must be careful to
take note of it. Ordinarily we say that salt and iron and other stuffs are
observable, and we speak that way because the stuffs have observable
properties. Similarly, we say that cats and dogs are observable things,
though strictly speaking only their properties are sensed. We do not truly
see a cat or a dog; so there is, then, an extended use of the term
"observable" or "sensible" according to which any thing or stuff that has
observable properties is itself said to be observable, and the extended
sense provides an economical way of speaking. Only a pedant would
insist on the strict sense of the term in ordinary conversation. Nonetheless
it is imperative to stress that this usage does indeed involve an extension
of the terms "observable" and "sensible" that must be kept in mind when
one discusses substances in a technical way.
There is still another way in which we use "observable." We say, for
instance, that we have observed the growth of an animal or plant, or that
we saw someone sneer, smirk, or do something similar; and these are
cases in which "observed" or "saw" truly signifies an observable property
(increased dimensions in the case of growth, shape of the mouth in the
cases of sneer and smirk). Yet in addition to the observable properties, the
words "sneer," "smirk," etc., signify relations which the properties have
to unobservable, interior causes). For that reason the meanings of these
words and others like them are complex in the sense that the words
signify both an observable and a relation of the observable to its
unobservable cause. Terms of this sort are many, and that leads us to
another point.
When we speak of apprehending sneers, smirks, and the like, we often
use the word "perceive, "7 which commonly signifies something grasped
that is over and beyond the observable itself. Whenever we consider an
observable and something known through an observable we have to
employ a number of cognitive capacities, and we have to order what we
apprehend through them. Thus we must take care not to confuse a union
that comes from an order with a unity that is absolute and implies the
activity of only one operational power. Such a confusion can lead to our
assigning to sensory powers that which does not truly belong to them,
which seems to us to be the main reason for the difficulties about
unobservables.
E. Recapitulation
For the mind to move from an observable to an unobservable, it must

first understand that observables do not modify themselves. Consequently
at the same time that the mind comes to know an unobservable through
an observable it also recognizes the reality of the cause-effect relation
between the realities, which makes this relation the first unobservable.
Additional considerations allowed us to see that in general relations are
unobservable. Furthermore, "observable" and "sensible" have several
meanings, which must not be confused. Nor ought a union among distinct
cognitive powers be taken for the absolute unity of one operational
capacity.
Having said these things, we may now leave behind our discussion of
the fundamental cognitive relation by which the mind knows that
unobservables exist. The mind does, however, make use of another very
important relation, a relation that allows us to know something about what
unobservables are; and to it we must now tum.
Notes
!.Of course the particles themselves were not observed, but measurable effects
were.
2. Currently "opposed" is used improperly to mean "distinct from," as in "He
watched the game as opposed to working on his taxes," rather than "He watched
the game instead of working on his taxes." It is true that the two activities cannot
occur simultaneously and so in that respect they embody the principle of non-
contradiction. But they do lack the determinate relation that opposition taken more
properly implies. At the least, such a use of "opposed" is bad English.

3. The opposites that are contraries, each of which is positive, of course affect the
senses independently of one another; and the one member of the pair that lacks
something of the formality which is more fully realized in the other is nonetheless
able to stimulate the senses in the measure it is positive. Yet as our actual practice
shows, the mind does not understand the lesser of the pair perfectly except in
comparison to the more actual, which is able to stimulate the sense more strongly.
4. We suggest the reader see the entry under this title in the Encyclopedia of
Philosophy.
5. If the relation is not observable by sight, then a case will have to be made for
the sense which detects it; and that case will have to be made by those who claim
the relation is observable.
6. Substance and Modern Science, Center for Thomistic Studies (University of
St. Thomas, 3812 Montrose Boulevard, Houston, Texas)
7. The words "perceive" and "perception" have recently come to be commonly
used to indicate that one takes not stand on whether what one is talking about is
true or false. As so used the words claim only that the one talking apprehends
something in the way he indicated, whether it is that way in reality or not. This,
however, is not the sense of "perceive" that we just mentioned.
Chapter 4
The Similarity-Dissimilarity Relation
A. Introduction
Whenever the mind goes from an observable to an unobservable, its

use of a cause-effect, principle, 1 sign, or other relation as a vehicle for
moving to the unobservable is reflected in words that have a multitude of
related yet distinct meanings, one of which is the first or fundamental
signification. "Health" and its derivatives, "healthy" and "healthful," are
well-known illustrations, and "cheerful" is similar in the way it signifies.
"Health" first tells us of a state or physical quality in an animal or plant,
but when we say medicine is "healthy," we do so not because health is in
it but because medicine is a cause of the health in the animal. When we
say urine is "healthy," we do so because it is a sign of health. Similarly,
"cheerful" first signifies a mood in us, while a room is "cheerful" only to
the extent that it disposes us to such a mood, and we may say the same
about music. The sound of someone's voice, however, is "cheerful" in the
measure that it is a sign of his interior mood. Thus the extensions of the
word "cheerful," like those of "health," depend on a cause-effect relation
in the things the word signifies. To repeat, then, the many meanings of
words that are related to one primary meaning are the consequence of the
mind's having gone from something observable to something else that is
related in one of the ways above to the observable. Yet there is another
way mind goes from one to the other.
Often we use observables to show something about unobservables at
times when we do not make use of a cause-effect relation. Poems, novels,
and other works of imaginative literature employ metaphors, similes,
verbal symbols, and the like to talk about realities that escape observation
yet are not causes of something else of which the work of literature
speaks. Although metaphors, similes, symbols and literary figures
generally do not depend on cause-effect relations, we must nonetheless
ask what a work of literature would be that confined itself to the

properties that stimulate the senses, never surpassing them to get at
something not observable? Nor should we forget that the sciences,
although metaphors are not their proper device, make use of comparisons
and analogies that do not always depend on cause-effect relations. In
short, this other way of using observables is what we wish to consider
now.
B. Analogies
When we are close to someone striking a nail, our seeing and hearing
the blow appear simultaneous to us and do not reveal any difference in the
time at which we see the impact and hear its sound. But when we stand at
some distance we first see the hammer strike the nail and only several
seconds later do we hear its impact. On such occasions we are aware of
a difference in the time of our apprehension of the two sensible properties,
a difference that prompts us to infer that sound takes time to travel. We
also infer (having excluded other alternatives) that the propagation of the
sound depends upon an unobserved movement of air which is between us
and the object struck.
But a question arises about the sort of movement that is responsible for
sound transmission; and because the physicist cannot infer the nature of
the movement directly from the difference in time or from the character
of the sound, he is obliged to construct an hypothesis which postulates the
nature of the propagation. Furthermore, to form his hypothesis, he must
make use of sensory impressions he already possesses, as well as
whatever intellectual understanding he has already acquired from them.
Using these starting points, he then proposes that sound is a wave motion.
If we now ask from what the notion of wave is first obtained, we
obviously must say from the waves of a lake or the sea, which are the
waves most familiar to us and which are part of everyone's ordinary
experience. (Of course, other observable waves might provide starting
points, for example, the oscillation of a plucked string.) Then the physicist
compares this familiar motion to the transmission of sound, with a view
to modifying or adapting his original conception of wave to this quite
different phenomenon. In doing so he first notes the characteristics of
observable waves that the unobservable sound propagation must be
assumed to resemble, namely, the medium is disturbed, the disturbance
has a periodicity, a speed of propagation, etc. Having done that, he
The Similarity-Dissimilarity Relation 35
postulates the differences that must exist in order to account for what is
special to sound, for example, that its waves are not propagated in a plane
but in three directions from the point of origin. In short, the characteristics
which the wave motion of a body of water has that sound propagation
cannot have must be denied of sound, and so negations play an essential
role in understanding the nature of these waves.
Modem physics first conceived light, too, to be propagated by a wave
motion; and the relation of our understanding of both sound and light
waves to ordinary experience is brought out well by Christian Huygens
in the passage below: 2
Light, then, is transmitted in some other way, a comprehension of which

we may get from our knowledge of how sound moves through the air.
We know that sound is sent out in all directions through the medium of
the air, a substance invisible and impalpable, by means of a motion that is
communicated successively from one part of the air to the next; and as this
movement has the same speed in all directions, it must form spherical
surfaces that keep enlarging until at last they strike the ear. Now there can
be no doubt that light likewise reaches us from a luminous substance
through some motion caused in the matter lying in the intervening
space-for we have seen above that this cannot take place through
transmission of matter from one place to another.
If, moreover, light requires time for its passage-a matter we shall discuss
in a moment-it will then follow that this movement is caused in the
substance gradually, and therefore is transmitted, like sound, by surfaces
and spherical waves. I call these "waves" because of their likeness to those
formed when one throws a pebble into water, which are examples of
gradual propagation in circles, although from a different cause and on a
plane surface. 3
From this passage we see that the points we have already made about
how we go from an observable to a description or definition of an
unobservable are confirmed by the actual historical record as outlined by
Huygens. That record provides us an illustration ofhow we spontaneously
use observables to come to an understanding of what a certain unobser-
vable is; and in this case there is no cause-effect relation between the
observable water waves and the unobservables sound and light waves that
are defined through them.
That light is more difficult to come to understand than sound is easily
seen, and the reason is that the observables-the facts-from which we

must start are fewer. Percy Bridgman makes the point very well when he
says:
Practically all our thinking about optical phenomena is done in terms of

an invention, by means of which these phenomena are assimilated to those
of ordinary mechanical experience, and so made easier to think about. To
realize that invention has been active here, we must think ourselves back
into that naive frame of mind in which experience is given directly in
terms ofsensation [emphasis supplied]. The most elementary examination
of what light means in terms of direct experience deals with the things
lighted. This fundamental fact is never modified by the most complicated
or refined experiments that have ever been devised [emphasis supplied];
from the point of view of operations, light means nothing more than things
lighted [emphasis hisV
Unaided sensation reveals two kinds ofbodies, the first of which are those
that are lighted and are visible by themselves. Such bodies are said to be
"luminescent," and ordinary language simply calls them "lights." Physical
theory then extends the word "light" to name that which travels, that
which is propagated from the luminescent body, and plainly that is an
extended sense of the term.
The second kind ofbodies, those that are not visible by themselves, are
not luminescent, and in order to be visible they must be in the more or less
proximate presence of a light. Instead of luminescence, such bodies have
a colored (or colorless) surface by means of which they are made visible.
Just how the lighted body makes the colored body visible the physicist
must explain, and it is here that we see how little data he has. Ordinary
observation shows us that a body which is not visible by itself can be seen
only if it is in the presence of another body that is luminescent. Inside a
windowless closet, no non-luminescent body is visible; but introduce a
light, and it can be seen, from which we infer (instantaneously, of course)
that the luminescent body somehow makes the non-luminescent body
visible by acting on it. And to our inference that light is active we must
add some others:
Now experience shows that these things lighted may stand to each other
in varied relations; in attempting to reduce these relations to order and
understandability we make a certain invention. This is prompted by
several cardinal experimental facts: in the first place, things lighted have
a simple geometrical relation to each other, in that screens placed on
-
.........
....- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
straight lines between the lighted objects may suppress the illumination of
one or the other and themselves become illuminated. This leads to the
concept of rectilinear beams of light, which is no more than a description
of the geometrical relation between lighted objects. Then we have the
experimental fact of the asymmetrical relation of the lighted objects,
described in terms of sources and sinks. Finally, we have the discovery
made at a much later stage, and not possible until physical measurements
had reached a high refinement, that light has properties analogous to the
velocity of material things. This was first discovered in connection with
astronomical phenomena in the shift of the time of eclipse of Jupiter's
satellites and in aberration, but was later found to hold for purely
terrestrial phenomena, in that a beam of light reflected from a distant
mirror does not return to the source until after the lapse of a time interval
that can be measured with means sufficiently refined. 5
When it comes to saying that light travels, the physicist depends upon a
correlation of observables, on the basis of which he must fashion a
satisfactory account of the nature of light propagation, making use of
ordinary experience; and it is not hard to see that the analogies he
employs are founded on a similarity-dissimilarity relationship between an
observable and an unobservable that are not related as cause and effect.
Additional testimony to the value of analogies is provided by the
physicist's use of fluids flowing in pipes to manifest something about
electrical currents, and here, too, the observable is not the cause of the
unobservable:
To this day, in teaching elementary electricity we liken electric circuits to

hydraulic circuits of water taps, flowmeters, pressure gauges ... to
correspond to generators, switches, ammeters, voltmeters .... Like many
uses of analogy in teaching, this does make things easier for the beginner
to understand. 6
But physics is not the only science that employs analogies:
The genetic language (language of genes) is very similar in principle to a

written language, like English, which consists of a set of symbols forming
the letters of an alphabet. The letters are arranged in linear sequences to
form sentences. The 26 letters ofthe English language can be arranged in
sequence to form words which vary in length from single-letter words ..
. to very large words ... Words can then be assembled into sentences of
varying lengths. Unlike English, the genetic language contains only four
letters, corresponding to the four bases A, G, C, and T. The words in the
genetic language are the base sequences which correspond to specific

amino acids. As we shall see, each word in the genetic language is only
three letters long, i.e., consists of a sequence of three bases. Thus, word
size is constant. ... 7
Regarding the text cited, let us note that the order which the authors
followed in manifesting their point is an order the mind must employ.
That is to say, the mind starts by noting the similarity between an
unobservable and an observable and then goes from there to the
differences between the two. Given the origin of human knowledge from
sensible traits, the mind can do nothing else.
C. The necessity of analogies
The kind of mental process described and illustrated above is

employed everywhere in the sciences, and one philosopher speaks of its
role as follows:
[The] history of theoretical science supplies plentiful examples of the

influence of analogies upon the formation of theoretical ideas; and a
number of outstanding scientists have been quite explicit about the
important role models play in the construction of new theories .... 8
Analogies, Nagel tells us, are widely employed. 9 If, however, someone
argues that analogies are helpful but not necessary, we wish to say 1) that
because no one can escape comparing like to unlike, and 2) because
everyone must know unobservables through observables, it follows that
analogies are not only helpful but are indispensable. To read the nature of
an unobservable from an observable, we have no alternative but to use the
similarity/dissimilarity relation.
There is reason to note that the similarity relation is also used on those
occasions when we can infer from an observable effect something about
the nature of its unobservable cause. For instance, if we notice that the
moon is eclipsed, we infer from the shadow-the effect-that an opaque,
spherical body has been interposed between the moon and the sun.
However, in other cases that are unlike the eclipse, a likeness can be so
remote, so extended, so tenuous that we speak of the more manifest as
only "suggesting" the nature of another. One ought not always to think
image or copy when he hears "like" or "similar," especially in a systematic
context, for the unobservables of nature ordinarily are not images or
copies of observables. To repeat: at times the distance between the one

and the other is so great, or the character of the unobservable is so
obscure, that there is no similarity sufficiently evident for the mind to
make use of it. Furthermore, when the term "sign" is taken strictly, 10 it
provides no useful likeness, and therefore it can show only that the
unobservable exists.
When we say that knowledge ofunobservables involves making use
of similarities and dissimilarities, we do indeed imply that every definition
of an unobservable involves a negation. This point is well made by
Thomas Aquinas when he comments on a text of Aristotle's:
He [Aristotle] says that a point-which is a certain sign of division

between the parts of a line-and everything else that is a division between
the parts of a continuum [as an instant between periods of time, etc.], and
likewise anything whatever that is both potentially and actually indivisible
[as is a point] is "shown," that is manifested, to the intelligence "as a
privation," or in other words, through the privation of continuity and
divisibility.
The reason is that our minds start from sense experience, and therefore
sensible things are the first to be intellectually apprehended. But all
sensible things have magnitude, hence the point and the unit can only be
negatively defined. For the same reason whatever transcends these
sensibles can be known by us only negatively ....
The same is true of things that are known through their opposites, for
example evil or black, both of which are related to their opposites as
privations. One of two contraries is always imperfect or privative in
relation to the other. He [Aristotle] then adds, as though by a reply, that
the intelligence in a certain way knows each of these through its opposite,
namely evil through good and black through white. 11
His point is that whether an unobservable is known through an analogy

or by some other relation it has to an observable as cause, effect, sign, or
principle, the unobservable will nonetheless contain a negative element
in its definition. That is a limitation we cannot escape.
Perhaps we should add one more point having to do with the relations
of principle and sign. When we considered the word "cheerful," we saw
that it has several senses, those that are secondary being related to one
first, primary meaning derived-in this case-from internal experience.
The primary sense of"cheerful" refers to one of our moods, while its later
extensions refer to exterior realities, one of them disposing us to be

cheerful, and another functioning as a sign of the mood. In such cases
there is no question of an analogy or similarity, but only of a cause-effect
or sign relation.
Although we need not labor the issue, in some instances internal
experience provides a starting-point for knowing something external that
is less accessible to the mind than what is internal. In some measure we
know animals better than plants, principally because we have internal
experience of activities they also have. Animals see, hear, feel pain, etc.;
and we are very much aware that they resemble us in this respect. Plants,
on the other hand, are not accessible in this way; internal experience can
reveal little about them.
D. A recapitulation
In this chapter we have discussed the second relation on the basis of

which the intelligence can know unobservables through observables. It is
the similarity-dissimilarity relation, which founds analogies and can be
used even where there is no cause-effect relation between the
unobservable and the observable. The similarity-dissimilarity relation is
also used when an effect allows us to know something about its cause on
the basis of a resemblance to the cause. Such resemblances, however, are
often tenuous and they ought not to be thought of as images.
Notes
1. Principle is more generic than cause; for every cause is a principle, but not
every principle is a cause. Principle has the notion of a beginning; and a journey,
for example, has a beginning, but its beginning is not its cause. Similarly, a point
is a beginning of a line, but it is not a cause of the line.
2. We may ignore for our purposes here the particle-like properties light has been
discovered to have.
3. The text cited is from "The Wave theory of Light," in Classics of Modern
Science, ed. William S. Knickerbocker (Boston: Beacon Press 1962), p. 54.
4. The Logic ofModern Physics, (New York: The Macmillan Company, 1927),
p. 150.
5. Percy Bridgman, ibid., p. 151.
6. Eric M. Rogers, Physics for the Inquiring Mind (New Jersey: Princeton
University Press, 1960), p. 150.
7. Arthur J. Vander, James K. Sherman, Dorothy S. Luciano, Human Physiology,
(New York: McGraw-Hill Book Company, 1970), p. 103.
8. Ernest Nagel, The Structure ofScience, (New York: Harcourt, Brace & World,
Inc. 1961), p. I08.
9. As the reader knows, "model" is another name either for an observable or for
a prototype invented by the imagination with the aid ofthe mind and put together
from the sensory data of ordinary experience.
IO. When "sign" is taken strictly, it excludes "image," "symbol," and "likeness."
II. Commentarium in Aristotelis Librum de Anima (Rome: Marietti, 1948), Lib.
III, teet. II, nn. 757-9.
Chapter 5
More on Analogies and Metaphors
A. Our knowledge of internal states
In an earlier chapter we considered the difference between external

and internal experience, pointing out that external experience is an
awareness of the properties of exterior realities, whereas internal
experience bears on our own cognitive and affective states, the two kinds
of observational activity being very distinct and independent of each
other. Yet their independence is not absolute. Investigations carried out
by Jean Piaget on children indicate that is so, for in one of his works he
reports the results of inquiries which aimed at determining the
understanding children have of some realities of ordinary life. Among
other things, he discovered that when children are young, they tend to
identify thought with speech and/or hearing, and the mind with the mouth
and/or hearing. The following are answers that are typical of those given
to the interrogators:
Mont.[the numbers that follow each name tell the age] (7;0 months): "You
know what it means to think? _ Yes _ Then think of your house.
What do you think with? _ The mouth. _ Can you think with the
mouth shut? _ No. _ With the eyes shut? _ Yes _ With the
ears stopped up? _ Yes. _ Now shut your mouth and think of your
house. Are you thinking?_ Yes _ What did you think with? _
The mouth.
Pig(9;6,backward): "You know the word 'think"'?_ Yes_ What

does it mean, to think? _ When someone is dead and you think of
them. _ Do you think at school? _ No _ And here? (we were in
the school office)._ Yes, I think because you have asked me things.
_ What do you think with? _ The mouth and the ears. _ And do
babies think?_ No _ Does a baby think when its mother talks to it?
_Yes_ What with?_ The mouth ...
Barb(5;6) "You know what it means to think? _ When you can't

remember something, you think. _ What do you think with? _ The
ears. _ If you were to stop them up, could you think? _ Yes ...
no."
We may summarize Piaget's results by quoting his own lines:
... we have traced three distinct stages, the first of which is easily
distinguishable from the other two and appears to contain a purely
spontaneous element. During this stage children believe that thinking is
"with the mouth." Thought is identified with the voice. Nothing takes
place either in the head or in the body. Naturally, thought is confused with
the things themselves, in the sense that the word is a part of the thing.
There is nothing subjective in the act of thinking. The average age for
children of this stage is 6. 1
The responses reproduced above are not isolated, Piaget tells us, but
are indeed representative of the young, which must not be taken to
suggest that we can all obtain the same sort of responses. On the contrary,
Piaget maintains that considerable skill is required to make sure the child
is responding to the questions according to his own mind and not
according to what he has picked up from his environment, nor according
to what he thinks the interrogator might like. But when investigations are
carefully made, the results do indicate that children fail to recognize that
their internal states are distinct from the objects known through those
states. Children do not understand internal states as internal states; their
minds require time to develop sufficiently to make distinctions. So given
the priority of observables over unobservables, we are hardly astonished
to find that children do not at first distinguish thought itself from thought's
sensible manifestation-speech-and the mind from the mouth's
organ-the mouth.
That in our understanding of observation we should find the object to
have priority over the activity itself becomes more intelligible when we
consider the principle at issue, a principle illustrated by how we
distinguish one artistic activity from another. We would instruct anyone
not knowing the difference between painting, say, and sculpturing by
pointing out to him the works each art produces. As a painting itself
differs from a piece of sculpture, so the art of painting (the mental quality)
differs from the art of sculpturing. Thus we distinguish activities from one
More on Analogies and Metaphors 45
another through their objects, and the same is true of the operations that
are passive in relation to their stimulus-objects; for were we in doubt
about whether an organism had a sense of hearing or a sense of sight, we
would settle the question by exposing the organism to the object-stimuli.
In sum, experience of our own behavior shows us that we know the
external realities which are the objects of our activities before we know
the activities themselves, which is a fundamental trait of an empirical
intelligence. Though we are a bit ahead of ourselves, we wish to say that
ultimately everything the mind knows, even about itself, it knows through
objects that the senses apprehend.
Focusing now on the results ofPiaget's investigations of children, we
may say that only after a time do we come to see that our interior states
or activities are distinct from the objects they attain; yet knowing what the
internal states are is even more dependent upon external realities, a point
we consider central and one we shall now attempt to show.
B. Internal states: how we obtain their definitions
Consider how ordinary language uses the word "sensation." According

to Webster's (2nd ed.), "sensation" signifies: "Orig., a state of
consciousness produced by an external object or by a change in the body;
also, that mode of mental functioning referred to immediate stimulation
ofthe bodily organism, including seeing, hearing, smelling, etc." Ordinary
language provides either "state" or "function" as category or genus-terms;
for both of them appear in the definition quoted above. And we may add
that Webster's also indicates that "function" signifies: "Action, activity;
doing; performance"; and daily experience confirms the use of these
words to describe sensation. Philosophers usually prefer "state," while
psychologists and biologists usually prefer "response" or "function," both
of which are words for a doing, an activity. We, however, in what follows
shall most frequently use the terms "activity" and its synonym "action" in
talking about cognitive, mental states.
Activities that are physical and observable are the first the mind
knows: we cut and plane boards, heat water, mold clay, mix cement, paint
houses, bend pipes, throw and hit balls, etc. And nature is responsible for
similar doings: the sun warms the earth, the wind blows seeds, glaciers
move boulders, waves erode shores, the moon raises tides, etc. In all of
these, the agent to which we relate the action brings about a real, physical
modification in a receiving object that is outside the agent itself; that is
what the word "activity" signifies. Were the sun the only body in
existence, it would be hot, yet it would do no heating. The sun would also
possess a gravitational mass, but it would do no attracting. In sum,
convention imposes the words "activity" and "action" upon the physical
modification in one body that is brought about by a quality or motion in
another; that is the primary meaning of these words.
In contrast, our activity oflooking at the moon, for example, does not
modify the moon nor any of its properties; nor does anything happen to
the piano when we listen to it. Even when we feel warmth or wetness or
softness or heaviness, the touching sensation itself does not modify the
object that is touched. Though a physical contact (and hence a physical
modification) is necessary for tactile sensations to occur, the contact is not
the sensation or operation, but the condition of the operation's exercise.
We could repeat our descriptions in the paragraphs above using "state"
in place of "action," "activity," or "function"; and in so doing, we would
begin with a term that is more general as well as equivocal, which would
require that we separate those states that are correlated with physical
modifications from those states that are not so correlated. Thereafter,
however, the procedure would be similar. And though "action," "activity,"
and "operation" appear the better terms, we shall not make an issue of
conventions. Whether one uses "state" or "activity," the notions signified
are the same. However, as Aristotle pointed out long ago, the activity we
call "sensation" is not the same as the kinds we have used as illustrations
above. With our external sensory apparatus we cannot detect other
people's (or our own) sensations; for plainly the movements of eyeballs,
fingers, etc., although necessary, are not the sensations themselves, which
are "psychic activities"; and the fact that sensory operations bear such a
name suggests that they differ from physical actions in important ways.
Still, physical and psychic activity do resemble one another; for we know
from eliciting and controlling our sensory operations that they require
both a "doing," a "sensing," as well as an object toward which the doing
or sensing is directed.
We may now come to the point: understanding sensation as an activity
obliges us to describe it as lacking something inherent in physical action:
psychic actions do not modifY the objects toward which they are directed.
Thus we must introduce a negative element into the definition of such
sensory states or actions because they differ from the actions that we first
know: those that have observable effects. To recapitulate: we define
sensation first by naming it an "activity" because of its similarity to
physical action, and then we introduce a distinguishing negative element

that puts it into another category; and in so doing we have used an
analogy like that of wave by employing observable actions as the starting
point for defining sensation. 2 Having said all this for the sake of showing
how we define unobservable sensations, we may add that our internal
cognitive and affective states are even more remote from observation and
therefore are even more difficult to defme properly. They are one step
further removed from external observables.
What we said and illustrated above about the use of analogies in
defining unobservables was confirmed very succinctly some years ago by
the following words:
Words are human artifacts, meaningless save as our associating them with
experience endows them with meaning. The word "swarm" is initially
meaningful to us through association with such experiences as that of a
hovering swarm of gnats, or a swarm of dust motes in a shaft of sunlight.
When we extend the word to desks and the like, we are engaged in
drawing an analogy between swarms ordinarily so-called, on the one hand,
and desks, and so forth, on the other. The word "molecule" is then given
meaning derivatively: having conceived of desks analogically as swarms,
we imagine molecules as the things the desks are swarms of. 3
Quine's words need no exegesis from us.
To summarize: we know that internal states exist directly by

awareness within us, but we know what they are through the medium of
the evidence provided by external observables. Yet even knowing that
they are within presupposes we first know the world "out there," which
means we can say only with qualification that our knowledge of our
internal states is "subjective." On that account Descartes and others like
him are wrong-headed when they attempt to infer the character of the
external world from internal observations which bear on their own interior
states. Cogito ergo sum is unintelligible unless one knows at least in a
minimal way that thinking is an activity directed toward an object; and in
knowing that much he will have spontaneously employed the notion of
activity obtained from the world of observation. In short, there is no way
philosophical idealism can avoid surreptitiously introducing at the
beginning that which it claims either not to know at all, or to know by
inference from what is within the mind.
C. Metaphor: its necessity
As we have seen, the human mind makes use of likenesses in

describing and defming unobservables, a use that stems from the mind's
having to derive its concepts from observables. But the fundamental role
oflikenesses requires some emphasis if its full import is to be appreciated.
Turning to Ernest Nagel once again, we read:
Common speech is full of expressions that initially were employed in a

more or less conscious metaphorical sense, though many of them have
well-nigh lost their original meanings and are currently used in what is in
effect a literal manner.... The widespread use of metaphors, whether they
are dead or alive, testifies to a pervasive human talent for finding
resemblances between new experiences and familiar facts, so that what is
novel is in consequence mastered by subsuming it under established
distinctions. In any event, men do tend to employ familiar systems or
relations as models in terms of which initially strange domains of
experience are intellectually assimilated. 4
In his own words Nagel confirms the central points we have made above,
but we must add two qualifications. First, words that are no longer used
metaphorically do not "lose" their original meanings. On the contrary, the
original meanings must be retained if the secondary meanings are to be
clearly understood without a substitute meaning being supplied at the
moment. We have argued that what actually happens is that the word
ceases to be a metaphor through acquiring another literal meaning.
Second, the use of metaphors is not only a "pervasive talent," it is also a
necessary one with which we are not able to dispense. Let us elaborate.
Because we must know unobservables through the medium of
observables, our initial transference of the name of an observable property
or entity to something that is unobservable has to be a metaphor. For
instance, the first step the physicist takes towards providing a category
term for sound propagation is to call it a "wave"; yet as long as he rests
there, as long as he goes no further in his description (his understanding),
his predication of wave is metaphorical; for up to this point he
understands only a likeness that has not yet been circumscribed.
Moreover, as the reader knows, metaphor has often been defined as a
compressed simile, which indicates its character very well; and the first
step the physicist takes amounts to saying that sound propagation is like
a wave. That there may be no noticeable time lapse before he moves from
understanding sound as a likeness to understanding it as distinct in certain
respects does not affect the issue; for until he (mentally) sees that sound
propagation differs in certain determinate respects, he has not gone from
the metaphor to the analogy.
Whenever we use a metaphor, we assign a name without assigning the
definition to that to which the name is joined. For instance, the statemenJ
"John is a moose," attaches the name but not the defmition or description
of moose as a four-legged, antlered mammal, etc., to John. Consequently,
the plain intent of the metaphor is to tell us that John is like a moose.
Similarly, when we first call sound a "wave," we are saying that it is like
a wave; and in both the case of the moose and the case of the wave, we do
not consider the named entities with respect to what is proper to them. (Of
course, "moose," unlike many metaphors, will probably not acquire a
literal meaning in relation to men. Furthermore, metaphors can be used
when there is no attempt to delineate the proper nature of the thing
described, as is usual in imaginative literature.) So in summary, whenever
an unobservable receives its first description from an observable, whether
internal or external, we have done no more than present an uncircum-
scribed likeness. Not until we have assigned a distinguishing characteristic
that separates the unobservable from the observable does the name receive
an extended literal sense. Metaphors, then, are an essential element in
human cognition. Stated another way, the human intelligence must begin
with what many call the "concrete."
D. Some additional points
Pierre Duhem accused English physicists of being tied to physical

models in a way he considered inappropriate:
In the treatises on physics published in England, there is always one

element which greatly astonishes the French student; that element, which
nearly invariably accompanies the exposition of a theory, is the model.
Nothing helps us better understand how very different from ours is the
manner in which the English mind proceeds in the construction of science
than this use of the model. ... These abstract notions of material points,
force, line of force, and equipotential surface do not satisfy his [the
Englishman's] need to imagine concrete, material, visible, and tangible
things. "So long as we cling to this mode of representation," says the
English physicist, "we cannot form a mental representation of the

phenomena which are really happening." It is to satisfy this need that he
goes and creates the model.
The use of such mechanical models, very far from facilitating the
understanding of a theory by a French reader, requires him, in many cases,
to make a serious effort to grasp the operation of what is often a very
complicated apparatus, as described to him by the English author. Quite
an effort is required in order to recognize the analogies between the
properties of this apparatus and the propositions ofthe theory that is being
illustrated. This effort is often much greater than the one the Frenchman
needs to make in order to understand in its purity the abstract theory
which it is claimed the model embodies. 5
The model, the image constructed in the imagination and sometimes

in material representations, is the mark, Duhem says, of the English mind,
an issue which he discusses at considerable length in the work cited. But
his comparison and exactly what he means is not our concern here, except
insofar as it allows us-no matter what Duhem may have in mind-to
point to some extremes, neither of which is a genuine alternative for the
human intelligence.
When he describes the penchant English physicists have for physical
models, he could be taken to mean that they think their theories must
contain conceptions that are exact likenesses of the observable models
they have constructed. In other words, ifDuhem is construed to mean that
the conceptions of the mind cannot transcend sensory representations,
then he would make English physicists adherents, knowingly or
unknowingly, to Hume's position on concepts as fainter images. At the
other end of the comparison, he could also be taken to mean that the
French and German physicists have purely intellectual conceptions of
physical realities, purely intellectual in the sense that their content bears
no trace whatever of observables. If Duhem's position is so construed,
then he has indeed presented two extremes, both of which are to be
avoided. A theory that is purely abstract in the sense that its notions have
no relation whatsoever to observables is not truly possible for the human
intelligence; and let it not go unnoticed that such a position comes to what
the positivists and analysts described as a meaningless "metaphysics." We
have been told that much that passes for metaphysics is meaningless
because it is empty of reference to the empirical; and so if metaphysics is
nonsense, then so is a "purely formal" theory of the natural world. But if
the mind must derive its concepts of the unobservable from the
-""'-.------------- ---- --
observable, then a total divorce of its conceptions ofunobservables from

observables is impossible in any discipline; and whether we consider our
theories to be "purely formal" or not does not matter. Moreover, what we
have just said is implicitly conceded by those who propose the purely
formal to be that which is common to all realities. However "thin" the
concept of that which is universally common, it nonetheless bears a
relation to the real world, and is not totally divorced from it. The concept
is not totally empty.
Thus, the reasonable position is one which maintains that our
conceptions of unobservables will to some extent resemble observables
and to some extent will not; and it needs no saying that at the far end of
our cognitive spectrum some notions of unobservables will contain only
faint traces of their most direct observable origins. Currently particle
physicists who are seeking a unified theory are hunting for more
elementary particles to unite the forces, particles for which they
admittedly cannot suggest adequate images. And we must be careful not
to exclude entirely the concomitant use of internal sensory impressions
with concepts. Through our intelligence we deal with arguments and
definitions and the ingredients of each, yet we can hardly avoid a
concomitant use of internal sensory capacities, as for instance, the
imagination. We must, however, take care not to confuse one with the
other, not to substitute an image for a definition or an argument; and that,
we might say, is often not an easy task.
Considering once again the extremes described above, the error we are
more likely to make is to substitute pictures for concepts. That is, we are
more likely to be inordinately bound to observables than we are to try to
transcend them altogether, a point that some years ago was well made by
Moritz Schlick:
The connection between theory and reality was formerly always conceived
of as though the symbols occurring in the laws of nature represented
simple magnitudes, or quantities, which could either be immediately
perceived, or could at least be regarded as being of the same nature as such
magnitudes or quantities (e.g., a length of Ill OOm). Thus, in Newtonian
mechanics, the fundamental concepts represented by lines in space, time,
and mass, were three terms of which the meaning seemed to be derived
immediately from sensory imagery. All three are combined in the concept
of motion, which is equivalent to temporal change in the spatial position
of a mass. Motion is that process in which the basic requirement for
knowledge appears to be fulfilled in a pictorial manner -namely, as the
perception of the constant element in change. That which is moved-the
mass-fills the role of substance, and remains unchanged in sense-percep-

tion. And yet, something does change-namely, the position. The whole
process seems to be perfectly clear and visually conceivable; and this is
the only reason for the predilection for mechanical explanations, and for
the desire of earlier physicists to reduce their science to mechanics. Hence
also, the peculiar extension in meaning of the word "mechanism." 6
The point to be made from the citation above is that the mechanical view,
the representation of unobservables wholly in terms of observables, is an
inclination we must be careful to discipline, otherwise we seriously distort
the intellectual enterprise.
E. A recapitulation
In this chapter we illustrated the use of analogy and observables in our

understanding of our own sensory and internal states. We argued that we
must employ exterior observables both to know that our internal cognitive
and affective states are distinct from their objects and to know what the
states are. We also noted that metaphors are a necessary first step in
corning to know what unobservables are precisely because we must first
know the latter through similarities they have to observables. As a
consequence of this dependence on observables, we must be careful to
avoid two extremes, one of which consists in thinking we can separate our
concepts altogether from observables, and another of which consists in
thinking that concepts cannot transcend observables in any measure.
At this point we wish to note that in the past some have had the
practice of thinking that in using the Aristotelian terms of "actuality" and
"potentiality"-or simply "act" and "potency"-they were doing
metaphysics in the Aristotelian sense. Because the investigation of a
discipline's subject, a subject that does not imply a material in the
constitution of the entities it considers, uses terms such as these
extensively, it does not follow that every such use is metaphysical; and we
hope that what we said above about analogies and how what they mean
begins in the world of observation makes it plain that "actuality" and
"potentiality" and synonymous terms first of all get their sense from the
world of nature through an analysis of change and changeable entities.
Again: their entire first meaning is derived from an analysis of natural
substances and their properties, and, we cannot go beyond the world of
nature in a systematic, scientific way until the existence of unobservable
immaterial substances has been established.
F. Epilogue: experience and Phenomenology
The philosophical "movement" known commonly as Phenomenology,

which now seems to belong more to the past than the present, represented
itself as an attempt to return to the phenomena of experience as the basis
and measure of its philosophizing. Nonetheless a claim phenomenolgists
made about experience deserves consideration for the sake of making as
clear as possible our position on observation.
In his work on the history of Phenomenology, Herbert Spiegelberg 7
devotes the last section of his two volumes to a discussion of that which
is common to all phenomenologists, namely, the phenomenological
method. He tells us that
Ever since Husserl's phenomenological manifesto, "Philosophy as a

Rigorous Science," the reclamation of the immediate phenomena under the
watchword "To the things themselves" (Zu den Sachen) has been the
leitmotif of phenomenological research.
According to Spiegelberg, the aim of phenomenology is to take

philosophy out of the mind and put it once more in the world of real
things, and at first sight phenomena would seem to be extra-mental-
perhaps even observable-traits of real things, since phenomenology
... expresses a revolt against an approach to philosophy that takes its point
I I of departure from crystallized beliefs and theories handed down by a
tradition which only too often perpetuates preconceptions and prejudg-
II ments. This negative part, the identification and deliberate elimination of
!
I I theoretical constructs and symbolisms in favor ofthe return to the unadul-
'.
terated phenomena, is by no means a simple and easy affair; it takes a
determined effort to undo the effect of habitual patterns of thought and to
return to the pristine innocence of first seeing.
The positivists are right in refusing to honor the checks of a high-sounding

metaphysical terminology unless they can be paid off, at least in principle,
in the specie of concrete data....
The watchword "To the things themselves" has primarily a positive

objective, bids us to tum toward phenomena which had been blocked from
sight by the theoretical patterns in front of them ... 8
One has to agree that the goal phenomenology sets for itself is the
right one, and the foregoing words do appear to say that phenomenology
has abandoned the enterprise of examining to the exclusion of everything
else, what is within the cognitive powers. There are, however, serious
doubts as to whether it has; and since it claims to find its origins in real
things, it will be beneficial to examine its claims for the purpose of
throwing more light on what is genuinely an empirical origin of
knowledge.
Spiegelberg tells us that the phenomenological method has seven
major steps, each of which has to be considered if one is to understand the
nature of the method. The first of the steps is called "investigating
particular phenomena," and the first sub-consideration under it is
"phenomenological intuiting." After some descriptive comments and an
apology for the term "intuiting," Spiegelberg turns to the example of
force, which he uses throughout his discussion to illustrate the points at
issue. Having paid attention to some confusions to be avoided and some
restrictions on the use of the term "force," Spiegelberg then says the
following:
Two situations seem to me to reveal forces with particular clarity: one is

the active use of our own force, the second the undergoing of the impact
of a foreign force. Both situations occur in the context of the experience
of our own body. However, it is important to realize that the body here
involved is not that of biological science with its bones, muscles, sinews,
joints, and nerves; for of these, oftheir number and locations, we have no
direct awareness. The body in which the experience of force is localized
is the subject of a special phenomenology, which describes merely the
body as it is given to our direct awareness and in the way it is given. True,
this experienced body and the "scientific" body overlap. But it is in the
non-overlapping sphere that the phenomenon of force makes its
appearance. 9
The focus of his text is clear in the sense that he claims that
phenomenology is concerned with force (and other such realities) insofar
as they issue from or affect us and so are the object of our internal
awareness. In other words, phenomenology does not address phenomena
that are external realities taken as such; instead it addresses phenomena
according to their mental or emotional appearances. Stated another way,
it addresses phenomena that are "experiences" insofar as the term signifies
an awareness rather than the reality on which the awareness bears. Thus,
it would seem that phenomenology, despite its assertions to the contrary,

has not left the domain of what it calls "experience." For it phenomena are
subjective; no distinction between the internal and the external actually
plays a philosophical role, even though at times allusions are made to a
world that is independent of the knower. The remainder of Spiegelberg's
description of force confirms what we have said.
Spiegelberg appears to be aware of the objections that are likely to be
made to his position, and to counter them he asks the question "Does
phenomenology explore only subjective phenomena"? He replies by
denying that it does, maintaining that
If by "subjective" we mean what is accessible only to reflective introspec-

tion, then phenomenology is certainly not completely subjective. For its
descriptions deal not only with the subject's side of experience, with his
acts and dispositions which can become thematic only in a reflective
return upon himself, but at least as much with those contents of his acts
which confront him as the objects of his experience and which do not
require any reflective tum. Thus colors, melodies, and specifically those
"forces" which we experience in our own lived body appear, as it were, in
front of us. No particular direction, inward or outward, is prescribed by
the essential nature of phenomenological intuiting. 10
Spiegelberg's attempt to defend the objective nature of phenomen-

ology would seem to fail. For him, the objective is the content of our
initial cognitive acts, which he sets in opposition to reflection. But a
realist will immediately reply that "content" is ambiguous and can mean
( 1) that which is known insofar as it is within the cognitive power yet is
distinct from the activity whose content it is; or (2) the objective reality
outside the knower that is actually attained. Only the second meaning of
"objective" gets the cognitive powers outside themselves. 11 1t would seem,
then, that in effect Spiegelberg collapses the distinction between the
external object and our awareness of it.
Additional steps in the phenomenological method are "phenomenolo-
gical describing" and "investigating general essences," to name only two.
But when one looks closely at the discussion of these and other steps, he
sees at once that a confusion is present that cannot help but prevent the
mind from focusing on the data. The confusion consists in the failure to
distinguish external sensory activities from those of internal sensory
powers as well as from the activities of the intelligence, and then it merits
more attention than is in order here.
The point we wish to make from what is said above is that we possess
a number of cognitive powers that are distinct and often difficult to
distinguish from one another, which means that unless we are careful we
can make erroneous statements about what pertains to external
observation. Often without realizing what we do, we assign what internal
cognitive powers know to the external object, and in so doing we go
beyond the immediate external observables, which is a confusing state of
affairs, to say the least. (Later, in a chapter on theory-laden observation,
we shall have more to say about this sort of thing.) There is, however, one
more point we wish to make about phenomenology, a point about its aim
of bracketing existence, of abstracting from existence and leaving it in
abeyance.
According to Spiegelberg the "minimum meaning" of the "phenome-
nological reduction" can be stated as follows:
The underlying idea of this metaphor [bracketing] is that we are to detach

the phenomena of our every day experience from the context of our naive
or natural living, while preserving their content as fully and as purely as
possible. The actual procedure of this detachment consists in suspending
judgment as to the existence or non-existence of this content. This by no
means implies that we deny or even doubt its existence to the extent of
writing it off, as Descartes had done. Eventually we could, and I might
even add, we should, return to the question of existence, although Husserl
himself never did so explicitly after developing his transcendental
idealism. To this negative or "bracketing" aspect of the reduction
corresponds as its positive complement the possibility of concentrating
exclusively on the non-existential or essential content, the "what," of the
phenomena. It is in connection with its positive aspect that Husser)
expected the phenomenological reduction to open up entirely new
dimensions for phenomenological research. 12
By way of comment let us say frrst that all of our ordinary knowledge
as well as that which is systematic begins with an existing world. We start
with entities known to exist, and we examine them; hence it is impossible
to abstract from their existence except in a qualified sense of the word.
Could anyone truly think that a newly born child's first concepts are of
possible realities? It is true that to the extent our systematic investigations
consider regularities, to the extent they consider classes and what is
common to many individuals, to that extent our investigations abstract
from the existence of the individuals themselves but not from the general
existence of the species. When a biologist investigates sharks, for
example, though their differences attract his attention, his principal

reflections leave aside the particular dimensions of this or that individual
specimen-as well as the peculiar singularities of its properties-in favor
of a consideration limited to that which is common. And though we do
abstract from singular existence of individuals, we do not abstract from
existence in general. That which is the object of our attention is common
to actually existing entities, and therefore it is not something purely
possible. In sum, we may say that to speak of bracketing existence,
thereby claiming that we suspend belief in the existence of the object of
our considerations, is to mislead, for we cannot suspend belief in regard
to the existence of objects of our initial cognitions and reflections because
that is what we first know about them. We can "suspend belief' only in
regard to those objects which cannot be clearly seen to be implied by real
entities, entities that are not knowable by observation.
Notes
I. The Child's Conception of the World, trans. Joan and Andrew Tomlinson
(Totowa, New Jersey: Littlefield, Adams & Co., 1965), pp. 39-40.
2. Although our description of sensation is accurate as far as it goes, we do not
pretend that the definition we have outlined above is complete. Enough has been
given, however, to make our point.
3. W. V. 0. Quine, "Posits and Reality," reprinted in Theories and Observation
in Science, ed. Richard E. Grandy (Englewood Cliffs, N.J., Prentice-Hall, Inc.,
1973 ), section II.
4. Op. cit., p. 107.
5. The text that follows is from The Aim and Structure ofPhysical Theory (New
York: Atheneum, 1962), pp. 66-71.
6. From Philosophy of Nature, trans. Amethe von Zepplin (New York:
Philosophical Library, 1949.), p. 27.
7. The Phenomenological Movement (The Hague: Martinus Nijhoff, 1960.) The
quotations we shall present here are all from Volume II, Part 5, Section XIV, The
"Essentials of the Phenomenological Method."
8. Ibid., pp. 656-8.
9. Ibid., p. 662.
10. Ibid., p. 691.
Chapter 6
Signs, Symbols and Words
A. Introduction
Having discussed the two relations on the basis of which the mind
moves from an observable to an unobservable, we must talk now about
signs and symbols as properties and as manifestations of this mode of
knowing, for we all recognize that the function of signs and symbols is to
direct the mind to something not directly and immediately apprehended
by it. To begin properly, however, we must first point to some
representative instances of signs and symbols.
Many things that we observe function as signs: black clouds tell of a
coming storm, leaves turning color forecast an approaching winter, the
smell of smoke warns of a fire, a sharp tone of voice declares the boss's
irritation. Physicians note the symptoms (a synonym for "signs") of their
patients and diagnose their diseases. Psychologists observe the abnormal
behavior of their clients and read their emotional disturbances, and so
both physicists and psychologists make use of signs. More prosaic affairs
depend on signs too, for all of us in the course of performing our daily
tasks read multitudes of signs. Yet not everything one observes is a sign.
The grass in the lawn, the bush next to the house, the squirrel in the yard,
the automobile parked by the curb-all these ordinarily have no signifying
function. Now all of this is familiar, for in essence we are repeating what
we said earlier when we gave evidence to show that we know
unobservables through observables. The repetition is desirable, however,
because we shall say something about how artefactual signs are made. 1
First, then, let us note that the genus or category according to which
sign must be defined is observable. Moreover, we must emphasize that the
process of making artefactual signs (and symbols) will have to imitate, if
the device produced is to perform properly, the mode according to which
the intelligence knows unobservables through observables. Symbols of
course have something in common with signs, and the two names are
often applied to the same objects. We hear man described as a "symbolic
animal" or a "symbol-using animal," more often perhaps than we hear him
described as a "sign-using animal." But we ought not to be blind to the
differences between "sign" and "symbol" when the things to which they
are applied differ. "Symbol," for instance, is applied to artefacts but rarely
to things in nature; so one important use of"symbol," is to indicate certain
signs of human manufacture.
Other words, too, are employed as more or less synonymous with
"sign," the more common of them being "symptom," "signal," "token,"
and "indicator." Such a plurality of nearly synonymous names should tell
the reader that although the word "sign" may be used in preference to the
others, the definition of sign, not its name, must be our primary concern.
B. What signs are2
As we have already seen, signs, in the strict sense of the word, are
observable realities; they are not images in the sense or concepts in the
mind, as some of the textbooks say. Furthermore, not everything, not even
all organisms, make use of signs-only men and certain animals, a point
that need not be labored. It is also clear that signs are starting points,
never end points, of a cognitive process from which the mind moves to
the knowledge of other things; and so only an extended use of the word
would permit one to apply "sign" to an unobservable, for an unobservable
can never be an absolutely first starting point. Once again, then, a sign is
by nature an observable reality through which the mind comes to know
something else, something not present to the senses, whether that which
is not present is in principle unobservable, or whether it is unobservable
only because of some circumstance.
Photographs, paintings, statues, and the like are of course not called
signs, though they certainly direct the mind to unobservable realities; and
the reason they are not signs is that they are all images. Convention does
not permit us to call a likeness a "sign" of the absent object it resembles.
As soon as one compares the things that convention does call "signs" to
things that are likenesses, he sees the difference between a sign and what
is not a sign, even though the non-sign may have the function of directing
the mind's attention to something that is unobservable. Briefly: a sign is
"qualitatively" distinct from that which it signifies, whereas a likeness, as
the name tells us, requires some measure, however remote, of qualitative
"! Signs, Symbols and Words 61
~'
t..
:i
similarity, and when the likeness directs the mind to something beyond
the reach of sensation, it often goes by the name of "symbol." Thus not
only does "symbol" suggest human manufacture, it also-even though it
can substitute for "sign"-suggests that in symbols some measure of
likeness is ordinarily the source oflight for the mind. 3 Literary images and
metaphorical devices are said to have "symbolic meaning," of which
parables and myths are perhaps the primary instances. 4 Let us add, too,
that a word such as "signal" usually signifies some practical activity to be
performed, a use to which "symbol" is not often put. 5 Having said those
things, we are now in a position to see that sign is to be defined according
to the following formula: a sign is a sensible reality that by reason of the
impression it makes on a sensory power draws the mind's attention to
something different from the sensed object, something that is not itself
able to produce a sensory impression (at least at the moment). As so
defined, a sign seems to be any observable that points to an unobservable,
but that is not quite right, for we ordinarily do not use "sign" and
"observable" coextensively. We do not say, for instance, that a moving
compass needle is a sign of a magnetic field, nor do we say that a shadow
on the moon is a sign of an eclipse. The reason of course is that these are
per se effects and they point directly to the cause. Signs, however, are
usually incidental-not per se-effects. In a somewhat similar fashion, an
image such as a photograph or painting leads the mind to an unobservable
(at the moment at least) reality, but it is not called a symbol because the
likeness is an image. Thus just as a per se effect is not a sign, so an image
is not a symbol. But speaking more generally, symbols, to the extent that
they are more remote likenesses of that to which they point, obviously
differ from signs in not being incidental effects of the unobservable to
which they point. Insofar as a symbol is a likeness but not an image it
does not have a cause-effect relation of the sort typical of signs.
From what has been said we now know that signs can be classified
according to the two different sources from which their signifying
function is derived. Some are of human origin, having been brought about
by the human intelligence and will, and others are independent of human
causal activity and function as signs whether we want them to or not. The
former may be called "arbitrary" signs to denote their human origin,
whereas the latter are obviously "natural." Black clouds, smoke, animal
tracks, glacier marks, etc., are natural signs, while traffic signals, many
trademarks, and most of all language, are "arbitrary."
The distinction between the two kinds can be approached, however,
in another and somewhat different way by noting that a systematic

investigation requires correlations or regularities as its starting points. In
the present case the correlation occurring between a sign and what it
signifies can be due on the one hand to human intelligence and volition
and on the other, to nature, which means that in one case the correlation
results from habits or customs, and in the other it results from something
natural. But language is perhaps the most universal and important
arbitrary sign, and so to it we now tum.
C. Language as a sign
We have already remarked that language is an artefact because words

are fashioned by the human mind for the purpose of signifying what it
understands. And when the mind sets out to make signs for its thoughts,
it uses as its raw material the human voice, which is naturally capable of
signifying the affective states called "emotions." Pitch, tone, intensity,
modulation-these are all ingredients of screams, moans, crying, squeals,
and the like; and when we produce them we tell the outside world about
our fears, pains, sorrows, pleasures, etc. Nor do men have to learn the
pitch, tone, etc. of their voices to express their feelings; nature gives them
tliat ability. Thus one may say that the sounds of the human voice, which
antecedently and by nature signify emotional states, are the materials from
which linguistic signs are made.
The function of words is not, however, to signify emotions in the way
the natural voice does. Even words such as "fear," "hate," "desire," etc.,
represent emotions not as felt but as thought about, not as something
subjectively undergone, but as objects of reflection. 6 Furthermore, words
that signify objects that cause emotions do not signify the emotions that
are caused by the objects. To be sure, words can be expressed in tones
concomitantly signifying feelings that accompany the concept, the
concomitant significations stemming not from the word as such but from
the voice as a natural indicator of affective states. Thus the concomitant
expression of emotion, however useful it may be, is incidental to the
signification of language whose direct and immediate function is to
signify concepts of realities. Propositions are true or false independently
of their tone, pitch etc., which would not be the case if an emotional state
were per se to a word's signification. Of course people sometimes use
vocal inflections to suggest something over and beyond the signification
of their words, as for example irony; yet ordinarily the listener must
Signs, Symbols and Words 63
construe a proposition put forward in an argument according to its stated

intention, ignoring thereby the indirect suggestions of the vocal sounds,
unless some circumstance points to them.
Words, then, are made from sounds that naturally signify affective
states, which means that in order for speech to represent what the mind
conceives, in order for speech to take on a new and additional
signification, it must undergo a modification. The point is important, for
the production of words imitates the natural process by which observables
point to unobservables. Just as natural sensible realities must undergo
modifications before they can lead the mind to unobservables, so the mind
must modify the raw vocal sounds before they can signify a new kind of
unobservable, namely concepts. Syllables, which are put together from
more fundamental sounds, are made into words, and words are then
arranged into sentences. Now in general the order introduced into
artefactual things is an effect of man's intelligence; and therefore the order
that appears in speech sounds, the property that most obviously
distinguishes language from purely natural utterances, is a general
indicator of intelligence as distinct from emotion. But though the
tendency to arrange and order is native to the mind, convention does enter
into the formation of words, for public acceptance and public use are the
only causes of a word's signifying one thing rather than another. Yet the
arbitrary has limitations, for once words have been assigned to the
realities that are closest to experience, further naming is not well done if
it is purely arbitrary.
The intelligent production of signs avoids the purely arbitrary as much
as possible because the purely arbitrary is inherently unintelligible. 7 Just
as words start with sounds that already have a natural signification and
then are given an additional sign function, so actions or objects having a
natural function or signification-or a signification well established by
usage-can be modified to signify something new. The clenched fist
gesture illustrates the point. The swung fist is a weapon, and the gesture
is but a modification of the natural movement. Consider, too, the
handshake. Whatever its historical roots and genesis, the handshake is
clearly related to the helping hand we extend to someone who has fallen,
who is climbing an obstacle, etc. Consequently it employs the basic
motions of the natural act of helping and merely modifies the movement
in a stylized way to give it the signification of good will that it expresses.
In short, because the handshake is made from a natural but modified
motion, its signification is readily understood. Our point, however, ought
not be mistaken; for our aim here is not to question whether purely
arbitrary signs can be made or whether they are in any way useful. On the
contrary, our intention is merely to note that if in making a sign one starts
with something that has a natural function or established signification
already proportioned in some measure to the new signification to be
assigned, he will achieve superior results. In other words, an act of sign
making that modifies a regularity already possessing a meaning makes use
of the natural mode of the mind. On those grounds some artefactual signs
can be judged to be better or worse than others.
In this connection let us note that according to some scripture scholars,
the account of the origin of the universe in the Book of Genesis consists
of stories that are found in already existing myths about creation. If that
is true, it certainly would illustrate the principle we defended just above,
for the account in Genesis does not merely repeat these other stories but
modifies them according to their new intention, their new signification.
The advantage of this sort of origin of Genesis lies in its starting with
something familiar and commonly known and then undertakes to modify
it, thereby telling its audience that it has a new signification. 8
A more general illustration of the use of observables that occurs in
Scripture, this time an illustration of knowing a cause through its effects,
was given in a lecture by Dominique Barthelemy at the University of
Freiburg in Switzerland. Commenting on the Hebrew name of God that
is translated as "I am who I am," he remarked that although this is a very
legitimate rendering, there is another that is more fundamental and that
also renders the Hebrew accurately and is translated as "I intervene as I
intervene." (Plainly this refers to events that were not to be explained as
natural phenomena and stem from natural laws; on the contrary, their
special character as observables pointed to something beyond the
regularities of nature. In general that is the function of miracles: they are
sensible manifestations of events or phenomena that transcend the
ordinary processes of nature.) This understanding of the Hebrew
expression certainly accords with God's calling himself The God of
Abraham, Isaac, and Jacob, for he intervened in their lives, as the
Hebrews knew, and so he identified himself through those interventions.
That was the "name" he gave them.
The point of these illustrations is that in both of them the ordinary
mode of the human intelligence is observed. Genesis, if composed as
described, takes accounts of creation already familiar to many peoples,
including the Hebrews, and modifies them to signify something new; and
tl.
when God names himself as he does, he makes use of the dependence of

the human intelligence on observables. Only through his intervention in
earthly affairs do we come to know him; and succeeding generations of
Hebrews came to know him through the accounts of what happened to
Abraham, Isaac, and Jacob. But modification must also occur in the
human art of painting.
D. Imitative art
The reader undoubtedly knows that the critics of the imitation theory
of art understand the theory to endorse the view that works of (fine) art
have as their function the copying of nature. An imitation, they tacitly
assume, is a copy, which means that the theory implies the artistic ideal
to be a photograph of the reality portrayed. As a consequence, those who
support, for example, an emotive theory do so by attacking the obvious
weaknesses of such a view; but they misunderstand the position they
decry.
Once again we must draw attention to our lack of expertise in regard
to things artistic, yet we must say something about the imitative theory of
art precisely because its proper (and traditional) understanding entails the
empirical mode of the intelligence. When the Greeks said that the function
of the fine arts is to imitate nature, they did not mean that art should
attempt to produce copies. If that were its function, the artistic
representation would be as a source of light inferior to nature itself. But
even one who is only casually familiar with such things knows that art is
in some way an "improvement" on nature, at least insofar as art addresses
the human intelligence, though it also makes use of the human affective
response. This is not to say that art will strike the mind or affect the
emotions with more vehemence than nature; on the contrary, that often
may not be the case and is not to be expected. Traditionally, art was held
to imitate nature when the artist modified that which nature presented in
order to make some aspect of the object more manifest and so more
intelligible to the viewer. For instance, the function of a portrait painter
would be to show some principal quality or qualities of the character of
the person represented, thereby making them more manifest than they are
in the person himself or herself. To be effective the artist would have to
subdue some observable traits and highlight others; he would "interpret"
nature, portraying something of it in such a way that the viewer could see
and understand nature's works better through the artistic object than he
could without it. Thus, by definition art had to introduce a modification

into the presentations of nature.
The point we intend to make is now obvious. Ifthe function of the fme
arts is to imitate and be representative, and if imitation requires that the
artist modify the observable traits nature uses to manifest its realities, then
artists must follow nature's lead with respect to the way nature draws the
mind's attention to unobservables. Art, like nature, must introduce a
modification into observables to manifest some unobservable (or less
observable) reality that is its principal concern. So once again we see that
the empirical mode of the mind is ubiquitous; wherever the mind operates,
there its mode is present. But now we must look at words, how they relate
to concepts and to their first imposition on observables.
E. Words, concepts, and nonsense
Repeating what has been said above, signs are observables that point
to unobservables; and when the intelligence is itself a cause of signs, it
does well to follow the natural order of the intelligence by modifying
something that already has a public signification or by modifying some
regularity; that is the proper way to introduce the new signification. And
our point that signs ought not to be purely arbitrary is what Plato had said
before us in the "Cratylus," where he argues that signs are not well made
when they are purely arbitrary And this same point is implicit in the
words of a philosopher who was a logical positivist and who died some
years ago. We are speaking of Rudolf Camap, who claimed that
metaphysics is nonsense because it comes from what we might call the
"misuse" of words. Camap says that "A word which ... has a meaning ..
.is said to designate a concept," whereas words that are peculiar to
"metaphysics" designate pseudo-concepts, which is but another way of
saying that they are meaningless. 9 To illustrate his point he quotes from
an author who was his contemporary, but instead of using Camap's
illustration, I would like to provide one of my own choosing but not of
my manufacture, an illustration that in fact comes from an abstract of a
paper given at a philosophical convention:
The paper attempts to address this question of panchronic logos by

incorporating the insights of hermeneutic phenomenology (and related
developments in ordinary language analysis, etc.) into the celebration of
self-symbolizing, the "methodologically self-verifying constitution," of
the "absolute immediacy" of symbolic mediation in and across the models
of objective, theoretical dialectic, the gestures of practical, subjective,

political rhetoric and the preconceptual, connatural individuality of
productive poetic.
Nonsense and pseudo-concepts: that is the right description, and one finds
it difficult to understand how anyone could offer those words as a
description of what he planned to say. Now pseudo-concepts, Camap
claims, are not concepts at all but words that at one time had a meaning
which later they lost. Originally they signified either an observable or
something related to an observable, but they lost that connection when
they fell into the hands of "metaphysicians," continuing nevertheless to
exist as familiar conventional sounds. That is why they have only the
appearance of meaning and are pseudo-concepts. If words have no
connection to concepts, then they also have no connection to reality and
hence have no meaning. To be sure, the same word need not be a
pseudo-concept in all contexts, for its user may have found the word's
original sense and extended it in a very legitimate way.
Were the human intelligence to function in a Kantian, a priori fashion,
then we would be hard pressed to say how a pseudo-concept could come
about; yet once we see how the meanings of words originate in
observation, we readily understand that Camap is correct. Many words
have indeed become meaningless; and much jargon-especially
bureaucratic jargon-is, despite its often mundane subject matter,
"metaphysics" in Camap's sense. Only because it pretends to deal with
familiar affairs through familiar sounds does such jargon differ from the
more pretentious utterances of some academicians; and too often the
bureaucrats do not have clear ideas of what they are saying. But to make
our point more evident let us look again at the several meanings of words
that are related to one primary sense, which has to be the source of any
clarity secondary meanings have.
F. The many meanings of words
The axioms supporting Camap's position appear to be contained in two

propositions: 1) words signify concepts; and 2) through concepts words
signify realities. Now the second axiom depends on the first, for whatever
we think concepts are, we all know that somehow they do put us in touch
with real things, though just how they do that may be obscure. Still, there
is no doubt that they do it, for the human intelligence is certainly not
confmed within itself. Stated another way, if words signify concepts
directly, then they must signify realities indirectly, through the concepts.
Thus Camap correctly used word as his category-term for pseudo-concept
because the word was all that remained of what once was a concept. Yet
when Camap used the term "metaphysics," though he had certain people
in mind, 10 he did declare that not everything done under the title of
"metaphysics" throughout the history of philosophy warranted being
called nonsense. Also we wish to note that the word "metaphysics" is no
longer pejorative in sense as it was when the positivists held sway. But
leaving the term aside for the moment, let us illustrate what we have said
above by looking at the several meanings of words and their relations to
one another. And though there will be some repetition of what we have
said, we think it useful for our considerations on the order of naming.
We need only consult the dictionary to see that most words have
several meanings that ordinarily are related to one another. "Shape," for
instance, means first of all the contour, the configuration, the figure of a
body; and then it is extended from this first sense to mean more generally
a visible form or semblance of a type. In a still further extension, "shape"
comes to stand for any representation, even a phantom, after which it is
once more extended to mean a form of embodiment, as in words. And
there are yet more extensions: form of thought, mode of existence of
something ("the first shape of the writing"); yet even these do not bring
us to the end of the listings.
Another example is the word "form," which in the original Latin had
significations much like those of"shape." (In English "form" has actually
retained its first Latin sense, which is shape or likeness.) Now here, too,
the extensions are many, an important one being the ideal or intrinsic
character of anything, following which come manner or method, conduct
regulated by custom, orderly arrangement, and most generally, mode of
existence. The words "shape" and "form" remind us of "posture," which
first signifies a bodily configuration, and then is extended to mean a
military policy, as in "our defense posture." In each of these illustrations
the extensions are made on the basis of a similarity the second reality
signified has to the first, even though the similarity may in some instances
be remote.
In chapter 4 we talked about "health" and "healthy" as well as
"cheerful," for they are words that have original impositions that differ
from the examples above because both of them refer in their first sense to
something we detect directly by internal experience. When we are ill or
indisposed in some way, we are usually aware of that indisposition-our
lack of health-by how we feel, despite our not knowing exactly what
health is. The word "health" is then extended in the forms of "healthy" or
"healthful" when it is applied to food, medicine, and complexion. "Cheer"
and "cheerful" are like "health" and "healthy," signifying, as they do, a
mood we are in or not in that is known through internal observation. Later
these terms, too, are extended to mean cause of health or cheer, as well as
sign or expression of health or cheer, which can be read, for instance,
urine in the case of health and in the face or voice in the case of cheer. In
these examples the many meanings are related to one that is primary on
the basis of a cause or principle or sign relation. Of course not all words
with several meanings have senses that are related but instead are purely
equivocal; for example, "pen," which signifies a writing instrument and
a small enclosure. That, however, does not concern us here.
G. A summary of the above
The reason for the relation among meanings is not hard to see. If
words signify concepts directly and realities only through concepts, then
the order according to which words signify realities is not the absolute
order the realities have in the world, but the order they have as they are
known by us. If we derive our knowledge of unobservables from
observables, then the order of our concepts is not a one to one mapping
on the real world made according to the world's own relations. We cannot
claim that the mode of presentation in the intelligence is a copy of the
mode of occurrence in the world. So to repeat: the order according to
which we acquire concepts through observables determines the order
according to which words acquire their many related meanings. 11 Thus
realities are knowable by us according to their proximity to or distance
from observation. Furthermore, it follows from what we have said that a
reality which has no direct or indirect relation whatsoever to observable
traits cannot be known by us, as Carnap correctly maintained. We have no
warrant to speak about such "realities," and they present a grand occasion
for self-deception. (In passing we might add that what we have just said
indicates why probative arguments for God's existence present such
difficulties. The intelligence is obliged to start from observables and then
move to the object that is at an infinite distance from observation.)
H. Returning to words: the order of naming

Earlier we noted that Plato tells us in the "Cratylus" that "names ought
to be given according to a natural process ... " and as we have seen, the
natural process begins with an imposition of the name upon an
observable, after which the extensions follow the order in which concepts
are derived. If this order is followed, the first meaning of the word throws
light on those that are extended, and it does so either on the basis of a
cause-effect relation (in a broad sense of "cause") or on the basis of
similarities and dissimilarities. But as Plato's word "ought" says, the
natural order is not obligatory in a rigid way; in some measure, though not
totally, it can be circumvented.
Despite what is natural and ideal, unobservables can be labeled with
words or symbols that are made specifically for them, not exhibiting in
their own first signification a connection with observation. Still, the
natural order cannot be totally circumvented, for words or symbols with
no observable first imposition can be attached to unobservables only by
a procedure that indirectly makes use of observation. The attachments can
be made in several ways: ( 1) by a definition which itself employs words
or symbols with meanings originating in observation; (2) by a context
fully enough elaborated to make the significatum of the name stand out
through familiar and properly established references, a procedure that
requires at some point words which have first impositions taken from
observation; (3) by equating the new name with an established synonym.
Although such indirect attachments are more obscure than those that
result from naturally extending the name of an observable, they
nonetheless allow the attached word to perform its function. Now it might
seem that in this last paragraph we have been speaking primarily of
mathematical symbols, and though they are instances of what we are
describing, they are not the only ones. A certain class of words-a class
that originates in another language-is less easily recognized than are
symbols as being obscure, so they make good candidates for becoming
pseudo-concepts. But let us tum again to Camap for an illustration.
After having quoted certain passages from a "metaphysician," 12 Camap
singles out the word "principle" as having no meaning in the context
which he cites; and in our judgment he is correct. The author who is
quoted does not define the term "principle," nor does the context allow a
meaning to be assigned by comparing the reality signified (none can be
identified) to something observable; nor does the context provide an
equivalent synonym. If the author has anything at all in mind, he certainly
has given insufficient evidence of it. Still, the word "principle" is trouble-
some by itself, and there is a special reason why it is inherently obscure.

"Principle" is not indigenous to English; it has been borrowed from the
Latin, and when it was borrowed, its first imposition, which was an
observable, was left behind. In fact, the English use of "principle" has
been mainly though not exclusively philosophical, and it has been
attached to unobservables that are removed from ordinary internal and
external experience, and its significations have been acquired in an
indirect fashion. Consequently the English "principle" is inherently more
obscure than principium is in Latin. But in the texts which Carnap cites,
the word "principle" cannot be seen to be attached even in this indirect
fashion.
In Latin "principium" means "starting-point," which is what the word
meant to the Romans. And like its Latin counterpart, the English
"starting-point" first signifies the beginning of a sensible, physical motion,
which is to say that the Latin "principium" and the English
"starting-point" are attached by the same observations and are as
synonymous as words can be. Hence if the author whom Carnap criticized
had employed "starting-point" (or its equivalent in his own language), the
confusion would have been less, to say the least; and perhaps the author
himself might have seen that what he had written was vacuous. But let us
now move from the example to the general issue.
The obscurity inherent in "principle" occurs in many words that come
to English from other languages. Whenever we use foreign cognates
without first imposing them on an observable, or without correlating their
English senses with the observable attachments of the original tongue,
then the cognates have meaning in English only insofar as they are
arbitrarily and indirectly attached by definitions or synonyms. At best this
sort of imposition is obscure, which means that if a speaker does not make
his sense clear, then the listener subconsciously fills one in for himself;
and if he fills it in according to his desires, according to his affective
dispositions, then he probably will receive the speaker well. 13 Still, even
the cognates we use regularly and understand well lack the clarity of
native Anglo-Saxon speech. But let us illustrate our point.
We apply the word "to decide" to judgments of the mind, of which we
are aware through reflection. This English verb comes from the Latin
caedere, to cut, together with de, away from. Hence "to decide" comes
from a Latin expression meaning "to cut away from," which is an
observable physical action. And when we see that observable reference,
we understand better what "to decide" means; we see that we cut out one
option, preferring it to another or others.

The English "abstract" is used in a variety of ways to signify at one
time a concept in the mind and at another some immaterial of fictional
reality. But when we use "to abstract" in connection with intellectual
considerations, it means to consider one thing apart from another with
which it is joined in reality. The first meaning of the Latin word from
which "to abstract" comes is "to take away from," which refers to physical
acts in ordinary life. Interestingly, however, academicians, particularly
philosophers, are currently disinclined to use "abstract" and instead speak
of "cuts," or "intellectual cuts," expressions that are plainly clearer.
Because "cut" first signifies a physical activity, its extension to those that
are intellectual starts from an observable prototype, allowing us to see
very well what is meant when someone is said to make an "intellectual
cut." In sum, then, we must, if we wish our terms to signify clearly, attach
names first to observables, either external or internal, and then extend
them on the basis of a relation the new significatum has to the original
observable. Without such a procedure, the word has either an inherent
obscurity or no signification at all. Moreover, this natural process shows
why an ostensive identification of objects is the fundamental way for
learning what words signify.
We ought to add, too, that more and more our use of language is
becoming abstract in a pejorative sense of the word. Bureaucratic jargon
showers down upon us in signs, the media, and in conversation with the
bureaucrats themselves, all of which tends to hide the imprecision of mind
and judgment. We become attached to abstract terms because they are
easier to use than those that are specific, yet that makes more obscure
what ought to be clear. Observable or sensible qualities make a maximum
impact on us through our sensory apparatus, and on that account they tend
to inform the intelligence accurately and put it in touch with the source of
its knowledge, the world "out there." Consider, for instance, the
translation in St. John's gospel where it says that "He dwelt among us,"
and compare it to "He pitched his tent among us." There is little doubt
about which of the two expressions is the more forceful, although in this
case there may be good reason for not choosing the more forceful of the
two. But a second instance does not have the same excuse. Consider the
words of the centurion who said "I am not worthy that you should enter
under my roof," and contrast them to the more recent rendering which
says "I am not worthy to receive you." "Enter under my roor' is a
metaphor for an internal reception when used in the Eucharistic liturgy,
a signification not at all approximated by "receive." This latter term is

abstract enough to signify any kind of reception whatsoever and does not
of itself suggest one that is internal. By abandoning the metaphor much
is lost.
In practice, because we ordinarily use language without paying perfect
attention to the origins of words and their relations, we use them with a
measure of obscurity. Consequently, we usually are not perfectly
connected to the realities that are the source of our intellectual lights, and
it seems that this is one way in which it is true to say that "[intellectual]
starting points are everything" in systematic disciplines, especially
philosophy, a principle more honored in the breach than in the
observance.
I. Etymologies
That knowledge of etymologies is useful, even for native words, is

plain from our considerations above, since a search for etymologies is a
search for a first imposition provided by observation. The word, "water,"
for instance, signifies a stuff; and it comes from the Indo-European root
"wed," which meant "wet," as we are not surprised to hear. 14 Another
example is "sand," also a stuff, whose name comes from the root "bhes,"
meaning "to rub.' Every camper will appreciate this origin.
An attempt we made to predict to ourselves the root for "iron" as "that
which is strong" led us only to a synonym for the word; but "steel" turns
out to come from "stak," and the root means "that which stands firm." One
step away, the prediction is verified.
To repeat, then, etymologies are important for putting us in touch with
the observable world; they lead us back to the ordinary observations of
human life from which basic language derives its meaning. To be sure,
different cultures will show a number of different observable first
impositions because of differing customs and environments. Nevertheless,
because so much ordinary experience is a common human possession,
much of the time etymologies provide a first meaning that we ourselves
can confirm in observation. We also should keep in mind that first or
"nominal" definitions of natural entities-animals, for example -can be
appropriately given in terms of their appearances. 15 Because of the motion
of its mouth, a rabbit might appear to be cud-chewing; and as long as the
definition confines itself to the appearances, not pretending to be
scientific or systematic, no error is involved. But though such definitions
are accurate on their own plane, they are not equivalent to nor replace-
ments for defmitions that are systematic or scientific. Still, such first
defmitions need not be "corrected."
J. That from which and that upon
The general order of derivation of concepts and the corresponding

general order of naming lead to our recognizing that natural substantial
entities receive their names from something other than the entities
themselves. If substances are realities and not fictions, as Hume claimed,
then we understand why they cannot be named directly. Hume's principal
objection to substance was that it gave rise to no (sensory) impression,
and we must concede that substance is unobservable. Our understanding
of substance does, however, start from observables; and we can obtain a
notion of substance only from its observable properties. We should not,
therefore, be surprised to find that substances are named by taking a word
already attached to a sensible property and extending it with modifications
to the stuff or entity to which the observable property belongs.
Some of the etymological roots we considered earlier illustrate the
point we are now making. Consider, for instance, what "water" means. It
certainly does not signify the property wetness but rather the stuff itself
that is wet. Again: although the name "water" is originally taken from a
property, it is imposed upon the stuff, not the property. Perhaps a more
obvious illustration is provided by a word such as "carpenter," which is
taken from the art of carpentry but attached to the person in whom the art
exists. The same can be said of "meat cutter," "electrician," and many
othernames, all of which signify entities through their properties. In short,
the properties furnish the names; the entities receive them.
A different sort of origin for a name is illustrated by the word
"copper," which comes from the Greek, "Kuprios," "Cyprus" in English,
which is the name of a place. The property, which is a location, is perhaps
more obviously distinct than any other from the stuff which is named
through it, providing ready confirmation of the distinction between that
from which a name is taken and that upon which it is imposed. 16 An
interesting though false medieval account of a derivation is illustrative. It
would have the Latin word lapis originate from laedens pedem, injuring
the foot; and an injury to the foot of a man without shoes is an effect of
stones familiar to everyone.
There is a delightful book 17 whose aim is to show how various things
l
have acquired their names, and it provides a sort of scheme of

classification for the properties from which names are taken. The
following list contains one example of each division, as the author gives
them. Place of origin: "spaniel (Spain)"; size: "horsefly"; means of suste-
nance: "anteater"; characteristic sound emitted: "catbird"; shape:
"ringworm"; method oflocomotion: "grasshopper"; color: "hare (related
to hazy, gray)"; odor exuded: "muskrat"; facial expression: "dodo
(Portuguese, doudo, stupid)"; mode of scratching: "raccoon (Algonquin,
arathcone, he scratches with his hands)": mode of excretion: "butterfly,"
whose excrement resembles butter. We could go on, but these illustrations
suffice.
It is desirable to note, however, that linguists-one of my colleagues
tells me-offer two rules of thumb about sources of names: ( 1)
Scandinavian languages tend to name things from their activities; (2)
Latin and Greek tend to name things from their effects. These rules of
thumb are, of course, generalizations that are ordinarily true; and so they
supply additional support for the point above.
Other linguists state another generalization, one that says all words are
metaphors, meaning, it would seem, that one must expect names of
observables to be extended beyond the observables themselves. Such a
maxim, whether or not it holds strictly, is clearly founded upon the mode
of knowing appropriate to an empirical intelligence. The maxim also tells
us how to find the origins of our notions or concepts in the real world. But
certainly it cannot mean that all language is originally poetic or literary;
for metaphors presuppose first impositions that are literal.
Of course in one class of realities we can make no distinction between
that from which its name is taken and that upon which its name is
imposed; that is to say, there is one class for which the two are one and
the same, and that is the class of observables, of sensible properties. Only
when one reality is in some way more obscure than another and known
through it can the two be different.
Now the issue we have been discussing is not without philosophical
importance. For example, some who have disputed about the identity or
non-identity of the "morning star" and the "evening star" seem to have
failed to take the distinction into account. "Morning star" and "evening
star" are not identical as far as that from which the names are taken is
concerned, but they are identical in regard to that upon which the names
are imposed. The names of God provide another illustration; for they are
taken from different sources: his causality of real things (God), his
governance (Lord), his providence (Providence), etc. 18 All the names are
imposed upon one and the same entity, though they are taken from
different effects or activities, and they are synonyms only by reason of
that which they name and not by reason of the sources of the names.
K. Inflected languages
Languages are inflected, but some more so than others. We are

speaking now of those languages which modify a stem to decline nouns
and adjectives, conjugate verbs, etc. That languages should be so
constructed is not surprising when we consider that such inflection is an
imitation of the way the mind first knows unobservables. The
fundamental stems of the words compare to observables as what is first
known, and their endings compare to modifications that tell of relations
the observable has to unobservables (or at least to something more
obscure in some respect) that are subsequently signified. This mode of
constructing language would seem to have advantages for the intelligence
insofar as it makes for a greater unity in the mind's conceptions.
L. The contingency of language
To some it seems that the contingent nature oflanguage is an unhappy

event, and they would like to construct artificial languages to overcome
what they consider the deficiencies of the ordinary tongue. But we think
an artificial language would be a net loss; for it would undoubtedly ignore
the normal order of extending words, which occurs in a felicitous manner
when naming is left to the ordinary intercourse of human affairs. The
contingency of language has an additional advantage insofar as it helps
draw the mind's attention to its own activities and makes it reflectively
aware of its own empirical character. More particularly, the contingency
of language obliges the mind to investigate distinctions between things
bearing the same name. Stated still another way, because words can be
variously used and because they change their meanings and have various
origins, we are forced to trace them to observable starting points and to
examine the relations of the extended meanings to observables. Were the
connections of words necessary, it seems that at a minimum we would be
inclined to be less empirical than the health of the mind requires. Stated
another way, those who want to construct "artificial languages" appear to
want to axiomatize language and make it a priori, which would be a real
mistake.
Another remark seems to be in order. Given the order among the
meanings of words and the character of language as an instrument for
signifying the conceptions of the mind, it is evident that the failure to
discipline the young in their formative years to regard language with care
can be and has been mentally crippling. The first analytic training the
mind receives is in grammatical analysis, and that activity is certainly not
"purely verbal," if anything at all meets such a description. Furthermore,
the value of inflected languages in the formation of young minds (Latin
and Greek in particular) can hardly be exaggerated. A careful
understanding of the conventional regularities of language and their
careful application to a text, subject the mind to the basic disciplines
necessary for its own well being. Grammar involves the mind in the use
of principles, however imperfect they may be as principles because of
their imperfect regularity. However, before we can analyze arguments and
extract them from contexts, we must first be able to understand the
relation of the parts of the sentences in which the arguments are
expressed; and to understand the things signified, we must first
understand the linguistic order; failing the latter, the mind is left in
confusion.
M. Deconstruction: some comments
Not long ago I remarked to a student that deconstructionists regard

words as so indeterminate in sense that the only meaning they actually
have is what we put into them, and as a consequence words cannot put us
in touch with an objective reality. The student then told me that he tended
toward deconstructionism because the words of philosophers seem to
have many different meanings that were not distinguished. Given such
ambiguities and given the modem tendency to look on the mind as
contributing to the concepts it knows, one can hardly be surprised to fmd
that deconstructionists hold that the "reality" we know is one of our own
making. Nor is one surprised to see philosophers playing word games.
Aristotle remarks that the theoretical activity characteristic of philosophy
tends to exceed man's intellectual reach, and we can see why that is so
when we reflect on the difficulties the human intelligence has in
philosophy. But one part of the cure of our present ills is to start carefully
by identifying the observables that words first name and then moving
carefully to the extended senses to see what relation they have to the first,
sensory significations. We shall say more about these difficulties when we

talk about defmitions. We may however, state without hesitation that to
be empirically grounded, philosophy for instance, must trace the origin of
every word in the phrasing of its problems and every word in the
premisses or principles it uses in the problems' solutions. Philosophy must
see the first impositions in observations of the meanings of every one of
these key terms. That is what is required of every non-mathematical, non-
measuring discipline if it is to be empirical.
Notes
I. We use the non-conventional word "artefactual" in preference to artificial,

which too often nowadays is applied to something fictional. "Artefact" is a
legitimate spelling in place of "artifact"; so we do not think we are taking too
many liberties by using "artefactual," and we hope the reader will not object too
strongly to such a use.
2. The reader might see H.H. Price, Thinking and Experience (Cambridge,
Massachusetts: Harvard University Press, 1953), cc. IV-VII.
3. Mathematical symbols are not, of course, likenesses; they tend to be arbitrary,
and on the testimony of more than one mathematician, are lost from memory
shortly after they are no longer used.
4. We shall have much more to say about myth in a later chapter.
5. Perhaps it is of interest to the reader to hear that in a book entitled King
Solomon's Ring, Konrad Lorenz points out that animal communication involves
one animal's reading a mood of another. In one another and in men, that is what
animals detect. They do not communicate "concepts."
6. It is certainly a misuse oflanguage to claim that words are "sexist" in the sense
that anyone who uses them is by that fact prejudiced. Words signify realities,
some of which are emotionless because they signify non-affective realities, or
because they signify emotions as such realities. To be sexist a word has to signify
an objective reality that is a degraded, or considered as degraded, reality. Insulting
words applied to humans are instances of that sort of thing.
7. As its name implies, the purely arbitrary has its root solely in the will without
reference to an antecedent intelligibility the mind discovers. Hence, although the
mind can recognize the regularities the will has decreed, nothing in the arbitrary
addresses itself to the mind as inherently, as independently intelligible. Perhaps
that is one reason why even mathematicians have difficulty retaining the meanings
of newly encountered or seldom used symbols.
8. See Thomas Bulfinch, The Age of Fable (New York: Harper and Row,
Publishers, 1966) chapter 2, for a fable (Prometheus and Pandora) that might well
have served that purpose.
9. "The Elimination ofMetaphysics Through the Logical Analysis ofLanguage,"

trans. Arthur Pap, in Logical Positivism, ed. A.J. Ayer (Glencoe, Illinois: The Free
Press, 1959), pp. 61-62. In the pages that follow I shall not attempt to give
a precise account ofCamap's own position. My interest lies in the points he has
made that I can accept as consistent with what has been done in the foregoing
chapters. The reader shall judge for himself where he and I part company. Let me
say, however, that I think he was right on many points that since his time have
been abandoned, perhaps because they were not shown to be properties of the
empirical mode of the intelligence.
10. Op. Cit.
11. We wish to remind the reader that the proportion of something to
us, that is, its knowability in relation to us, must not be confused with
its absolute knowability; that is, its intrinsic knowability. Observable
properties are those which we know first without any qualification; yet
substances-elephants and mountain lions, for example-are more
knowable in themselves; there is more to be known about them than
any property.
12. Op. cit.
13. Under such conditions one has no difficulty seeing why many listeners will
all like some speech and yet have very different understandings of what the
speaker said. The bureaucratic jargon we hear more and more frequently often
meets this description.
14. This and every other root given here is taken from the appendix of The
American Heritage Dictionary.
15. We discuss definitions in Part II.
16. The medieval Latin expressions for the English are id a quo nomen
imponitur-that from which the name is taken-and id ad quod nomen
imponitur-that upon which the name is imposed.
17. Naming Day in Eden, by Noah Jonathan Jacobs (New York: The Macmillan
Company, 1958).
18. When God describes himselfto Moses he says, "I am the Lord your God, who
brought you out ofthe land of Egypt, out of the house of bondage." (Exodus, 20,
1-2.) This expression certainly takes into account the human mode of knowing.
The observable effect is the source from which his identity is established.
Chapter 7
The Necessity of Myth
A. Introduction
Chapter 6 dwelt mainly on signs, and so now leaving them behind, let
us return to metaphors, to likenesses, this time for the sake of looking in
a general way at myth and its role in forming the mind. As soon as we see
that knowing unobservables requires the mind to employ metaphors as a
first step ultimately leading to describing or defming unobservables, then
we understand in a general way the cause of the mind's need for literary
images, figures, symbols, etc., all of which we shall include under the
general title of metaphor. We see that we are obliged to treat realities
beyond observation first through the vehicle of the imagination, whose
raw materials are observables, and whose function is not only to retain but
primarily to create. Since unobservables differ from the observables
through which they are known, the creative function of the imagination
is especially necessary for an introduction to what we conceive about
unobservable realities. This is especially true when no systematic
procedure is yet available, or when such an investigation is not
appropriate. Every society-no matter how scientific or advanced its
systematic enterprises-must base its first understanding of the
unobservables important for human life on the constructions of the
imagination. In short, epic poems, dramas, fables, folk stories, and the like
are indispensable, especially, as we have said, for the first formation of
the mind. 1 But our aim here is to consider in a special way the use of myth
as a necessary prelude to systematic investigation; myth and science ought
not to be thought of as opposed but as complementary.
B. Metaphor and myth
In what we said in an earlier chapter, our focus was on the order

82 From Observab/es to Unobservab/es
according to which names are attached and the need to know a word's
first, sensory imposition. But we must also reflect to some extent on the
use of words in a metaphorical sense, a sense that depends on a relation
of similarity. Realities beyond observation we must first treat-on the way
to knowing what they are-through the vehicle of the imagination, whose
raw materials are observables, and whose function is not only to retain but
primarily to create. Since unobservables differ from the observables
through which they are known, the creative function of the imagination
is especially necessary for an introduction to what we conceive about
unobservable realities. This is especially true when no systematic
procedure is yet available, or when such an investigation is not
appropriate. Every society-no matter how scientific or advanced its
systematic enterprises-must base its first understanding of the
unobservables important for human life on the constructions of the
imagination. In short, epic poems, dramas, fables, folk stories, and the like
are indispensable, especially, as we have said, for the first formation of
the mind.
One of the more important uses of metaphor occurs in the construction
of myths, which ordinarily provide images of what are seen as causes of
natural events and things. We wish, therefore, to talk about the role of
myth in human learning. More particularly, we wish to consider the use
of myth as a necessary prelude to systematic investigation; myth and
science ought not to be thought of as opposed but as complementary.
C. The purpose of myth and its empirical basis
Myths are not adequately described when they are said to be

entertaining stories, however true it may be that many of them seem to
have no other function. Centuries ago Augustine of Hippo, following
Varro, divided myths into three classes: the myths of the poets, the myths
of the politicians, and the myths of the philosophers, that is, the
academicians. The first two classes are corruptions of the third, the myths
of the poets originating as adulterations made for the purpose of
entertainment, while the myths of the politicians were the result of the
trait political rulers often have of trying to use religion for political ends.
Only the third class, Augustine said, the myths of academicians, war-
ranted respect; and there is indeed a reason for his claim.
Augustine's myths of the philosophers ought not to be thought of as
philosophical in the modem sense. As the reader will know, in his day, and
l
The Necessity of Myth 83
for centuries thereafter, "philosophy" was a general term that embraced

nearly every reflective enterprise, and anyone who proposed a myth with
a genuinely intellectual aim was a philosopher. So although myths are
stories, those of the philosophers are imaginative representations which
portray gods as causes of natural things and events; they are, as many say,
accounts of"creation." The imaginative devices used by myths with which
we are concerned are principally the metaphor, the simile, and the larger
metaphorical construction, allegory.
Primitive societies generally, whether Greek, Roman, Sumerian, or
other, produced more or less sophisticated myths about origins. Even when
the myths became the focus of religious rites, their first purpose was to
account for natural phenomena. Speaking of the early Greeks, Edith
Hamilton says:
Greek mythology is largely made up of stories about gods and goddesses,

but it must not be read as a kind of Greek Bible, an account of the Greek
religion. According to the most modem idea, a real myth has nothing to do
with religion. It is an explanation of something in nature; how, for instance,
any and everything in the universe came into existence: men, animals, this
or that tree or flower, the sun, the moon, the stars, storms, eruptions,
earthquakes, all that is and all that happens .... Myths are early science, the
result of men's first trying to explain what they saw around them. But there
are many so-called myths which explain nothing at all. These are pure
entertainment, the sort of thing people would tell each other on a long
winter's evening. 2
Another writer about myth has a similar view:
. . . the definition of myth that seems least inadequate because most

embracing is this: Myth narrates a sacred history; it relates an event that
took place in primordial Time, the fabled time of the "beginnings." In other
words, myth tells how, through the deeds ofSupematural Beings, a reality
came into existence, be it the whole of reality, the Cosmos, or only a
fragment of reality-an island, a species of plant, a particular kind ofhuman
behavior, an institution. Myth, then, is always an account of a "creation";
it related how something was produced, began to be. 3
Cyrus H. Gordon, speaking about some tablets discovered at Ugarit,

describes the myths of Canaan as follows:
The Ugaritic myths explain nature so as to satisfy man's craving for the
answers to the universe, and to guarantee the regularity of the processes that
result in fertility: fertility of mankind, animals, and plants. The content of

the myths is conveyed through narrative full of action; the ancients were
not interested in abstractions. Their thinking was concrete and their gods
are portrayed as engaging in lively and significant action. 4
So there would seem to be a consensus that might be stated as follows:

myths are metaphorical "explanations" of nature's origins. But what we
have said in the foregoing chapters requires a caveat about Gordon's
remark that the "ancients were not interested in abstractions." As we shall
see later, the issue is not one of a lack of interest; it is not a question of the
ancients exercising a preference; rather it is a matter of their not being
prepared to do systematic science. The mental apparatus systematic
procedures require had not yet been discovered; it awaited more propitious
times, when circumstances prompted its discovery.
At this point the reader may be wondering why no mention has been
made of other theories about myths. According to one author, 5 there are
five "monolithic theories" of myths (as well as some not so monolithic),
four of which have not been touched on. So for the sake of a certain
completeness, we shall make use of G. S. Kirk to describe the remaining
theories with the aim of showing that the foundation of myth in the
empirical mind is the same irrespective of the general theory.
According to Kirk, the first of the "universal theories" maintains that
"All myths are nature myths [emphasis his], that is, they refer to
meteorological and cosmological phenomena." 6 That there are nature
myths can hardly be disputed, he admits, although efforts have been made
to make the theory fit cases it cannot. Yet, says Kirk:
... the excesses of the nature myth school have meant, if anything, that it
has been too little considered in any serious way. 7
An illustration of such myths is provided by the following:
The primordial pair of gods, still half-envisaged as great world-masses,

were Ouranos and Gaia, sky and earth; and sky lay upon earth and made
love to her without cease. The myth can be seen as a symbolic
representation of the interplay between rain and soil that makes plants come
to life and grow. Aeschylus wrote in his lost Danaids that 'the holy sky
passionately desires to penetrate the earth ... rain falls and impregnates
earth, and she brings forth pasturage for flocks and Demeter's 'life-giving
com'; and according to an admittedly late source the initiates at the
Eleusinian mysteries looked at the sky and called out 'Rain!', then looked
down at the earth and called 'Conceive!'. 8
Whether one takes the lines "The myth can be seen as a symbolic
representation of the interplay between rain and soil that makes plants
come to life and grow," to mean that the mythical gods represent the
hidden causal efficacies of rain and earth themselves, or whether the
mythical gods are taken as causes extrinsic to nature-in either case the
myths represent unobservable causes and are symbolic representations of
them.
Kirk calls the second theory of myths that he describes "aetiological,"
so naming it because
... it implies that all myths offer a cause or explanation of something in the
real world. 9 The first proponent of this theory, Andrew Lang, is described
as holding that. .. many myths are clearly not about nature; he [Lang] was
arguing that even those that are, are more than just pretty allegorical
conceits that are explanatory in some way. 10
This second theory would seem to differ from the first by its explicit
endorsement of myths as representations of realities exterior both to nature
and to human society, realities which are gods that cause entities in nature
and have effects on human society.
The third theory Kirk describes is one he calls a "charter theory," first
proposed by B. Malinowski, who said that myths
. . . should be considered as charters [emphasis his] for customs,

institutions or beliefs. By that he meant something close to 'explanations'
in a loose sense, but devoid of theoretical quality... What the charter
theory implies is that in a traditional society every custom and institution
tends to be validated or confirmed by a myth, which states a precedent for
it but does not seek to explain it in any logical or philosophical sense. 11
By the meaning of the term itself, a "precedent" must be taken to be a "first

case," a kind of model or paradigm that is imitated or copied in posterior
actions, in patterns of actions, or in institutions; and in that way the
precedent is causal. In other words, a charter is causal insofar as it provides
an agent (or agents) with something to be imitated or copied through its
own actions. It is a cause only in that sense. But whatever the role of the
charter, such a myth clearly stands for something that is not present and so
is unobservable; and for that reason it is represented by imaginative devices
constructed under the direction of the mind. But further remarks Kirk has
made are relevant to our point here:
In one way Australian myths can be interpreted as providing strong support

for the charter theory; as the Bemdts put it: 'Myths, then, may be used to
explain or account for certain rites, or to show why various actions are
performed; why a certain tribe practices circumcision, or why it does not
while its neighbors do. 12
Although the charter theory is supposed to differ from the others, it

proposes that myths symbolize causes, which is plain not only from the
role a precedent plays, but also from Kirk's remarks that myths "account
for certain rites, or show why [emphasis mine] various actions are
performed." But that which answers the question why tells us what is
responsible for the reality we are curious about, and that is what a cause is.
Let us repeat: a cause is that which is responsible for the existence of
something else, in any of several ways. Hence a charter theory is just a
more specialized and restricted causal theory of myths insofar as it would
have myths represent only one kind of cause: paradigms, or exemplars, as
the medievals called them.
Kirk names Mircea Eliade as the chief proponent of the fourth theory,
a theory which says that myths are representative of the "creative era."
Eliade thinks of myths as accounts of origins:
... the definition of myth that seems least inadequate because most
embracing is this: Myth narrates a sacred history; it relates an event that
took place in primordial Time, the fabled time ofthe "beginnings." In other
words, myth tells how, through the deeds of Supernatural Beings, a reality
came into existence, be it the whole of reality, the Cosmos, or only a
fragment of reality c an island, a species of plant, a particular kind of
human behavior, an institution. Myth, then, is always an account of a
"creation"; it relates how something was produced, began to be. 13
Although Kirk presents this as a distinct theory, there is little doubt that
it is causal; and the words of Eliade quoted above need no comment from
us to make the point. Moreover, any additional ceremonial or ritualistic
role the myth might play in "reenactments" of "primordial events" or the
"creative era" does not destroy the myth's character as a metaphorical
representation of unobservable causes. One may well expect ceremonies
and rituals to depend upon something causal as that which is being "cele-
brated," as some say.
The fifth "monolithic theory" Kirk describes maintains that "All myths
are closely associated with rituals. "14 This statement by itself does not say
much; for if we reflect we see that it does not propose an opposed theory,
a theory incompatible with the others. Myths could be aetiological, natural,
etc., and still be associated (as Kirk says) with rituals. But in a more
specific form the "ritual theory" maintains that myths "are actually derived
from rituals." 15 A proponent, however, puts the issue more strongly:
True myth may be defined as the reduction to narrative shorthand of ritual

mime performed on public festivals, and in many cases recorded pictorially
on temple walls, vases, seals, bowls, mirrors, chests, shields, tapestries, and
the like. 16
Such a theory makes myths to be symbols (because it is a shorthand) of

symbols, for the very notion of a ritual is that it signifies or symbolizes
something else. A ritual theory would seem implausible if it is supposed to
apply to most of what is accepted as genuine myth; yet even so it does not
deny our main point, that myths are imaginative representations of
something unobservable.
As a last consideration, we should say a little about some theories of
myth that Kirk does not consider "monolithic," namely those theories that
view myths as products of the human psyche, which roughly stated, is to
say that the theories regard myth as expressions of affective states, of
emotional conditions that can have any of a number of different origins.
But despite our objection to making emotions the ultimate goal of cognitive
acts, in this theory we nonetheless encounter the element common to all
myths; for according to it myths remain imaginative, sensory
representations of unobservables, which in this case are supposed to be
internal states directly accessible to the individual who has them but not to
the world at large. Such a "psychic theory" denies of myths any role in
communicating objective intelligible notions of common interest and
concern; yet even so we see that in such a theory myths reflect the
empirical character of the human mind insofar as they are proclaimed to be
sensory expressions of unobservable causal states.
D. A recapitulation
It would seem a mistake to think that myths can be subjected to the

same strict kind of classification that regularities of nature admit. Myths are
human devices that spring from the conscious activity of the intelligence
and imagination; they are not mechanical products of automatic responses.

For that reason we are mistaken if we think all myths can be accounted for
in the same way. On the contrary, since metaphorical representations
precede systematic accounts, we should expect myths to have as many
uses, as many functions as there are kinds of unobservable mysteries to
provoke wonder. Apart from their character as tokens or stories, apart from
the delight they produce in us simply as representations, myths would seem
to be able to play many roles in human life by functioning as principles of
manifestation for unobservables that are determinative of natural entities
and events as well as human actions, institutions, customs, and societies,
not to mention human interior affective states. In short, each of the theories
would seem to fit some myths, precisely because myths do not all address
themselves to the same kind of unobservables. Nonetheless, as we have
said several times, it does seem that the most important function of myths
is to provide a metaphorical representation of causes outside nature, of
gods or divine causal relations. The first three theories Kirk describes all
attest to that point. 17 Thus the role of metaphorical representations called
"myths" in pointing the mind to something unobservable and more or less
mysterious does not seem to be in doubt; so no matter which theory one
wishes to propose to account for all or some of them, myths are
instruments for the mind's corning to know unobservables.
In order that we not be misunderstood, let us note that we are not
advocating that myths be omitted from human education now that our
systematic sciences are developed. On the contrary, in some measure
individual psychological development would seem to benefit from learning
some of the myths, especially the better ones. Pondering on metaphors and
the images of mysterious causes is an exercise the mind ought at some time
to have. And if systematic examinations of the hidden entities forever
escape us, we at least have the beginnings of a contemplative appreciation
of nature when we see a little of what the myths suggest. So, then, as first
accounts of unobservable causes myths provoke wonder or curiosity that
leads to indeterminate and incomplete reflections on natural causes, and
such activities ought not to be underrated, since wonder, supplemented by
satisfaction, provides the drive which moves the mind to its investigations.
Stating the main point in a somewhat different way, we might say that
myths are principles of arguments by analogy that conclude to external
causes. Essentially they suggest that "Just as man is the cause of artefacts
and is himself outside that category, so gods are the causes of things and
events and are themselves outside the realm of nature." Just as a child first
The Necessity ofMyth 89
understands that artefacts are produced by men, without the child's relating
the functions of artefacts to structure and materials, so adults, early in the
generation of a culture, argue by analogy to gods as the causes of things,
without ftrst seeking to explain properties and behavior through the causes
within the things themselves.
It is interesting that the number of deities myths postulate does not seem
to be altogether a haphazard affair. Allowing for corruptions introduced
into the original stories by the entertainers, the evidence suggests that at
least in some mythologies a separate deity appears for every kind of natural
thing or phenomenon that has no observable cause. There is a god to
account for the sea, one to account for the rain, one for thunder, one for
volcanoes, and still others for plants, animals, and the remaining categories
of natural entities. Such a correlation is of course what we should
anticipate.
Aristotle is reported in his later years to have "returned to the myths" in
I\ his reflections on the ftrst cause, and we perhaps ought not to be surprised
because the indetermination of the metaphors that constitute myth allow
the mind to search for likenesses, for similarities that can enlighten the
I I
'~
',
mind on the most remote and inaccessible of all causes, namely God.
Although Aristotle argued very well to the existence of God and
understood his nature as well as one might from the evidence of nature,
most of his philosophical successors did not climb to the same heights.
When one looks at Cicero's De Natura Deorum he sees that those who talk
about God or gods describe him or them through physical attributes. This
work of Cicero's 18 presents the theological views of the three schools of
I r philosophy that were au courant in his day, and they were the Epicureans,
the Stoics, and the Academics. The following quotation illustrates what we
have just said:
For Jet us hear Plato, that divine philosopher, for ... He holds that motion
is of two sorts, one spontaneous, the other derived from without; and that
which moves of itself spontaneously is more divine than that which has
motion imparted to it by some force not its own. The former kind of motion
he deems to reside only in the soul, which he considers to be the only
source and origin of motion. Hence, since aU motion springs from the
world-heat, and since that heat moves spontaneously and not by any
impulse from something else, it fo11ows that heat is soul; which proves that
the world is an animate being.
. . . the world possesses inte11igence ... and if this be so, it fo11ows that the
world must be endowed with wisdom, for, if it were not, man, although a
part ofthe world, being possessed of reason would necessarily be ofhigher

worth than the world as a whole .
. . . But the fourth and highest grade is that of beings born by nature good
and wise, and endowed from the outset with the innate attributes of right
reason and consistency; this must be held to be above the level of man: it
is the attribute of god, that is, of the world, which must needs possess that
perfect and absolute reason of which I spoke. 19
... If again it [the being which embraces all things] be capable of reason
yet has not been wise from the beginning, the world must be in a worse
condition than mankind; for a man can become wise, but if in all the
eternity of past time the world has been foolish, obviously it will never
attain wisdom; and so it will be inferior to man. Which is absurd. Therefore
the world must be deemed to have been wise from the beginning, and
divine. 20
The point of these quotations is simply this: the human mind finds it
extraordinarily difficult to understand the nature of any substantial entity
that is beyond the physical world. The arguments for God's existence
present a maximum difficulty, as was said by Thomas Aquinas, and the
debates of which the passages are representative provide evidence of that
difficulty. Our times possess the revelations of Scripture in which God is
described as existing outside the world and his fundamental attributes are
made known to us. When, however, one has to establish from natural
evidence that God exists and what some of his attributes are, the mind
rarely gets beyond the physical. Perhaps that is why in our own day some
of the environmentalists consider nature herself to be god. Now back to
myths.
E. Myths and truth
Assuming the existence of causes outside nature, it is important to

understand that the representations of the myth makers-the distorted
myths of poets and politicians aside-are not essentially falsehoods; the
judgment is precipitous that says because an account is a myth it is
therefore pure falsehood. 21 Rather, it seems better to say that because
metaphors are the constituents of myths, properly speaking myths are
neither true nor false. For instance, when we say "John is a moose," and
then hear someone object that our statement is false, we may rightly accuse
the objector of not understanding what is going on. Only a claim of
resemblance, not a claim of identity of any sort, is made by a metaphor; so

it can be "true" in a more tenuous sense. 22 Myths and their larger
constructions, allegories, are "true" in the measure that the likenesses they
present actually enable the mind to discover something it had not yet seen.
An illustration of the point is provided by the tenth book of the
Odyssey, where we read that Aeolus gave Ulysses the adverse winds tied
up in a leather bag; and this fiction, as Alexander Pope called it, is hardly
a literal truth. On the contrary, the fiction is an image of something
physically impossible, which is the "most metaphorical" of all metaphors.
But if the poet intended the metaphor to be a stepping stone to a literal
sense, then what is that sense? 23 The critics are not of one mind about the
meaning of the bag of winds, but all of them give it a significance that is
allegorical or symbolic. It is hardly news that the image presents
difficulties of understanding because it only indeterminately represents
what it symbolizes; the image provides a likeness without a difference, and
a likeness can be applied to many things. Even the later action of untying
the bag by some of the sailors has problems of interpretation, in regard to
which Pope informs us that Bossu holds the untying to signify
... that we ought not to intrude into the mysteries of government which the
Prince intends to keep secret: The tempests and confusions rais'd by the
loosing the winds, represent the mischiefs and disorders that arise from
such a vain curiosity in the subject .... But whatever judgment is pass'd
upon this explication, it is certainly an instance of the ill consequences of
avarice, and an unseasonable curiosity. 24
Pope's last sentence shows the confused (in the sense of indeterminate)
character of metaphors. Because Bossu starts from the bag of winds and
moves to a specific kind of evil in political societies, his explanation is
more determinate than his principle allows. On that account Pope feels
obliged to point out that if we do not admit Bossu's explanation by reason
of its not issuing determinately from the metaphor, we must nonetheless
concede that the image does draw our attention to the evil consequences of
avarice and undue curiosity, which is a much more general interpretation.
Myths, then, if they are well constructed, ought to be regarded that way:
the more indeterminate the interpretation of the images, the more certain
the myth is; the more determinate the interpretation, the more open it is to
debate. Furthermore we often see that several determinate interpretations
are compatible with the same images, a state of affairs that does not involve
any difficulty.
There is another point worthy of mention that has to do with the

intellectual status of those who subscribe to the myths. For the most part
the subscribers do not take the imaginative accounts as literally true, even
though they cannot say just how the metaphor differs from the reality.
Rudolf Anthes puts the matter well when apropos of the early Egyptian
heavenly cow he says:
An Egyptian "mythological concept" is a concept by which man tried to

make comprehensible in human terms a figure, an event, group of figures,
or a sequence of events which appear to him to belong to the "divine
world." 25
This means that they are symbols, and
A "symbol" is the manifestation of a human attempt to make an element of

the divine world conceivable in human terms .... The Egyptian sages of
about 3000 B.C. were aware of this fact and did not mistake a symbol for
an actual replica of what it represented. 26
The reason for saying such a thing is:
... since the Egyptians had as much common sense as we have ourselves,
we may conclude with certainty that no one, except perhaps a very
unsophisticated mind, took the composite picture ofthe heavenly cow at its
face valueP
To say the Egyptian and other early peoples were less intelligent than we
is very unreasonable. To say that they viewed their mythical symbols
literally is like saying Columbus sailed west to prove that the earth is not
flat. Before his day the earth's spherical shape had been known for
centuries.
F. The abstract and the concrete
In the usual meaning of the term, "abstract" (an intellectual cut) refers
to a state of affairs one fmds in systematic considerations. The latter are
concerned with general propositions that can be cast in universal form even
when they are not true always but only most of the time. We also abstract
or make a cut when we classify things and properties in categories, for the
mind must slough off differences to obtain broader classes. Abstraction
also occurs when we leave aside irrelevant traits and events in favor of
those that are relevant to the issue being considered. In short, the ordinary
senses of the term refer to systematic considerations and do not describe
what goes on in imaginative works. We may rightly say, then, that abstract
thinking and metaphor are mutually incompatible. Of course even in
literature the mind cannot avoid using general or class terms that leave
aside individual traits; so everything is abstract in that minimal sense. Still,
there is a way in which literary works ought not to be abstract; and that is
what we would like to illustrate.
When we look at the lines below from Homer's Iliad-by way of
Alexander Pope-we see described the internal state of Agamemnon the
night after Achilles refused to return to the army, and the quoted lines
illustrate the point we wish to make about abstraction:
All Night the Chiefs before their Vessels lay,

And lost in Sleep the Labours of the Day:
All but the King; with various Thoughts opprest,
His country's Cares lay row ling in his Breast.
As when by Light'nings Jove's Aetherial Pow'r
Foretells the ratling Hail, or weighty Show'r,
Or sends soft Snows to whiten all the Shore,
Or bids the brazen Throat of War to roar;
By fits one Flash succeeds, as one expires,
And Heave'n flames thick with momentary Fires.
So bursting frequent from Atrides's Breast,
Sighs following Sighs his inward Fears confest.
Now o'er the Fields, dejected, he surveys
From thousand Trojan Fires the mounting Blaze;
Hears in the passing Wind their Music blow,
And marks distinct the Voices of the Foe.
Now looking backwards to the Fleet and Coast,
Anxious he sorrows for th' endanger'd Host.
Here ends his Hairs, in sacrifice to Jove,
And sues to Him that ever lives above:
Inly he groans; while Glory and Despair
Divide his Heart, and wage a doubtful War. 28
As we see, the passage above tells us something about Agamemnon's

internal condition and something about his character. Now in order to
provide a contrast, let me quote from one of Anthony Trollop's novels:
It must not on this account be taken as proved that Dr. Proudie was a man
of great mental powers, or even of much capacity for business, for such
qualities had not been required in him. In the arrangement of those church
reforms with which he was connected, the ideas and original conception of
the work to be done were generally furnished by the liberal statesmen of the
day, and the labour of the details was borne by officials of lower rank. It
was, however, thought expedient that the name of some clergyman should
appear in such matters, and as Dr. Proudie had become known as a
tolerating divine, great use of this sort was made of his name. If he did not
do much active good, he never did any harm; he was amenable to those
who were really in authority and, at the sittings of the various boards to
which he belonged, maintained a kind of dignity which had its value.
He was certainly possessed of sufficient tact to answer the purpose for

which he was required without making himself troublesome, but it must not
therefore be surmised that he doubted his own power, or failed to believe
that he could himself take a high part in high affairs when his own tum
came. He was biding his time and patiently looking forward to the days
when he himself would sit authoritative at some board, and talk and direct,
and rule the roost, while lesser stars sat round and obeyed, as he had so well
accustomed himself to do. 29
Let us say at once that we are not about to enter into a discussion of the
relative merits of epic poems and novels, or any other sort of literature.
Quite apart from our not being qualified for such a task, our aim is much
more modest: we merely wish to make a few comments about the abstract
and the concrete as they reflect the mind's greater or lesser distance from
its origins in observation.
Despite the basic, abstract character of intellectual processes that we
have mentioned, the Iliad though intellectual is a concrete work. That
becomes especially plain when the selection we quoted from the Iliad is
compared to the selection from Trollop. We saw in the passage above that
Homer aims to describe Agamemnon's affective state, and he does so
through the kind of observables from which such internal states are
ordinarily known. For instance, we all experience "rowlings (rollings) in
the breast," and we also experience their emotional causes. Similarly,
"sighs following sighs ...bursting from Atrides' breast" are observables
which convey a clear message to everyone. In short, Homer successfully
focuses the listener's attention upon sensible data which give rise to
inferences the listener or reader can hardly avoid. The poet makes his point
by presenting external and internal observables that produce the conclusion
he seeks, thereby causing his own conceptions to become those of the
listener or reader. Trollop, on the other hand, does not do the same thing;
he merely describes the unobservable, internal traits which characterize Mr.

Proudie without providing the observable data which are the evidence for
them, which means that his passage is analogous to an argument from
authority in a systematic discipline. That is to say, insofar as we must take
the characterization of Mr. Proudie without evidence, on faith as it were,
Trollop's writing is abstract in a sense that implies a less perfect intellectual
state, though it is not on that account poor literature, a judgment I would
not be able to make in any case. Put another way, we are not saying that
such descriptions should not occur any more than we would say that
arguments by authority should not occur. Our only point is that compared
to Homer Trollop is abstract; and in this case, being abstract is a less
perfect state of affairs. 30
We may rightly say, then, that we ought not to speak of the presence or
absence of "abstract thinking" in the works of ancient peoples without
making distinctions. First, no mind can function without employing general
terms and therefore abstractions, and in this sense every man is capable of
abstract thinking. Such an ability is native to and inseparable from the mind
itself. Second, when we employ observables to describe unobservables of
any sort, especially those that are extrinsic to nature, our first descriptions
must be metaphors; and so the use of concrete images or concrete language
under such circumstances is not a fault. Third, we cannot expect primitive
people to give evidence of the sort of abstract thinking which characterizes
science, since they are primitive precisely because they have not developed
the tools required for systematic enterprises.
G. A summary of the principal points
Myth is necessarily a first step in the cultural development of a people

because the mind's empirical nature requires it to begin to know what
unobservable realities are through likenesses, that is, through metaphors.
This state of affairs cannot be construed as indicative of lesser native
intelligence because "abstract procedures" and the systematic sciences
cannot develop without a first investigation that is done in terms of
likenesses. The empirical mode of the intelligence is evident in its artistic
works, as we should expect.
At this point we would like to add a note on the kind of
"demythologizing" that some scripture scholars have attempted to do.
Without entering into this business at length, let us simply note that the
word "myth" has a very different sense when it is used by these people.
Because they do not accept the New Testament as authentic accounts of

actual events, "myth" for them is a fiction, and that sense of the word is a
distortion of all that we have said above.
Notes
I. All that we have done in our considerations on observables implies that the use
of the imagination solely for the purpose of stimulating emotions without regard
to their subordination to and control by the intelligence is a kind of disorder. An
emotional state by itself can hardly be the primary aim of education or
investigation. The current tendency in the arts to make the stimulation of emotion
the artistic goal is clearly a kind of anti-intellectualism not in the best interests of
human life and culture.
2. Mythology, a Mentor Book (New York: The New American Library, 1942),
p. 19.
3. Mircea Eliade, Myth and Reality, Harper Torchbook (New York: Harper &
Row, Publishers, 1963), p. 5.
4. "Canaanite Mythology," in Mythologies ofthe Ancient World, ed. Samuel Noah
Kramer, Anchor Books (Garden City, New York: Doubleday & Company, Inc.,
1963), p. 183.
5. See G.S. Kirk, The Nature ofGreek Myths (New York: Penguin Books, 1974),
c. 3.
6. Loc. cit.
7. Loc. cit.
8. Loc. cit., p. 46.
9. p. 5.
10. Ibid.
I I. Ibid. p. 59.
12. Ibid., p. 63
13. Myth and Reality, HarperTorchbook (New York: Harper & Row, Publishers),
1963, p. 5.
14. Op. cit., p. 66.
15. Ibid.
16. Robert Graves, The Greek Myths (Baltimore: Penguin Books, 1955), Vol. I,
p. 10.
17. We ought to note that Fustel de Coulange in The Ancient City maintains that
mythical figures first began as household gods. Family ancestors first came to be
honored and then were gradually transformed into family gods, who ultimately
came to be honored beyond the family of their origin. And though the interest of
Fustel de Coulange is in such gods, he does grant that gods of nature were also a
part of the culture. In a way, family gods are even more closely linked to
observation than gods of nature, for they originate from the conviction that
providential causes come to the aid of men in their daily activities.

18. De Natura Deorum (Cambridge, Massachusetts: Harvard University Press,
1933).
19. Op. cit., p. 155.
20. Ibid., pp. 157-159.
21. In our time, especially in the media, myth is becoming a synonym for
falsehood. And it seems to me that such a use of the word is but one more sign
,,
',.1
ofthe deterioration ofthe contemporary intelligence. Does that intelligence truly
'.,II understand the character of literary images and devices?
22. We ought to add an obvious point: metaphors can become dead, and new ones
ought then to take their places.
23. Homer's is admittedly a distant metaphor and difficult to interpret.
24. The Poems ofAlexander Pope, ed. Maynard Mack (London: Metheun & Co.,
Ltd., 1967), Vol. IX, p. 341, commentary on line 50.
25. "Mythology in Ancient Egypt," in Mythologies of the Ancient World, ed.
Samuel Noah Kramer, Anchor Books (Garden City, New York: Doubleday &
Company, Inc., 1963), p. 183.
26. Ibid.
27. Ibid., p. 21.
28. Op. cit., Vol. VIII, pp. 1B2, lines 1B22. It need not be said that the poetry is
perhaps more Pope than Homer, but that is not relevant to our issue.
29. Barchester Towers, (New York: The New American Library of World
Literature, Inc., 1963), pp. 28-9.
30. As one of my colleagues in literature has said, injustice to Trollop, one ought
to say that Trollop's description of Proudie does become "concretized" in the
novel insofar as Proudie's traits of character are seen in the actions he performs.
On this account Trollop's novel is superior to those in which extensive
descriptions ofthe characters of a novel are given without the character traits ever
appearing in concrete actions. Such works are, of course, inferior because they
contain irrelevancies.
Chapter 8
Theory-Laden Observation
A. A representative view
Throughout the foregoing, observation has been distinguished from

the intellectual activities of arguing, defining, and forming propositions
because these are activities that do not occur in any sensory apparatus.
Also we have been discussing observables and observation with a view to
showing how they function as an origin of our intellectual activities,
showing too that our concepts are dependent upon observation without
being Humean washed-out copies. In a sense, then, this essay is on what
has come to be called "foundationalism"; yet our position is, we think,
different from that of most authors insofar as they incline towards
invoking either logical criteria or theories on the nature of science to make
their cases. Instead of beginning with observables themselves, which are
the mind's ultimate foundations, they tend to consider the character of
basic statements as distinguished from those that are derived. To begin
epistemological considerations with the mental is tantamount to starting
from the obscure and derived rather than the clearer and underived. Yet
idealism and the a priori tend to insert themselves in many places, and so
not surprisingly some philosophers have come to hold that observation is
theory-laden and that the intelligence imposes itself upon the senses by
determining, at least to some extent, the content of sensory impressions.
Theory-laden observation becomes, then, the topic for the present chapter.
Rather naturally we tend to think oflmmanuel Kant and his claim that
the intelligence contributes to the content of its necessary propositions
when we hear people claim that observation is theory-laden. To maintain
that observation is theory-laden in the sense that observations, particularly
those necessary for science, contain within them something of the theory
or theories they are intended to confirm or disconfmn amounts to
maintaining that there are no observations which are "uncontaminated" by
the intelligence, and the observations themselves are (at least in some
measure) a priori. Now if the intelligence is responsible even in part for
the content of observations, then it has to be independent of observables
for the origin of its concepts. Now in what follows we shall take as
representative of the view that observation is theory-laden a work of
Norwood Russell Hanson, 1 whose approach is more particular and more
empirical than most, and whose claims provide an opportunity to consider
a number of important points and difficulties. Also we shall briefly
consider some views of others.
B. Hanson's position
An empiricist of the sort that we are making ourselves out to be has

to maintain that no matter how greatly two people differ in their
theoretical accounts of nature, their basic observations are the same in
kind. Hanson, however, has no time for this view and claims that
... saying that Kepler and Tycho see the same thing at dawn because their
eyes are similarly affected is an elementary mistake. There is a difference
between a physical state and a visual experience. Suppose, however, that
it is argued as above-that they see the same thing because they have the
same sense-datum experience ... .If this is argued, further difficulties
obtrude. 2
As a first example of observations that give rise to difficulties, Hanson

refers to what people see in a line drawing (below), about which he says:
Fig. 1
Do we all see the same thing? Some will see a perspex cube viewed from
below. Others will see it from above. Still others see it as a kind of
polygonally-cut gem. Some see only criss-crossed lines in a plane. It may
be seen as a block of ice, and aquarium, a wire frame for a kite_ or any
of a number of other things. 3
Theory-Laden Observation 101
Though Hanson says we do not all see the same thing, he will not allow
us to explain our differences by saying that though the seeing is the same
the interpretations are different because
This sounds as ifi do two things, not one, when I see boxes and bicycles
[which means that] the concept of seeing which is natural in this
connection does not designate two diaphanous components, one optical
and the other interpretative. Fig. I is simply seen now as a box from
below, now as a cube from above; one does not first soak up an optical
pattern and then clamp an interpretation on it. 4
About another line drawing (not reproduced here) Hanson says that
although he sees a bear climbing a tree, someone else will not. He adds
that paintings are not only colors, plots are not merely details in stories,
tunes are not just notes because organization gives the elements of each
of these a shape or pattern, without which what we see would be
unintelligible:
Organization is not itself seen as are the lines and colors of a drawing. It
is not itself a line, shape, or a color. It is not an element in the visual field,
but rather the way in which elements are appreciated. Again, the plot is
not another detail in the story. Nor is the tune just one more note. Yet
without plots and tunes details and notes would not hang together.
Similarly the organization of Fig. III (not shown) is nothing that registers
on the retina along with other details. Yet it gives the lines and shapes a
pattern. Were this lacking we would be left with nothing but an
unintelligible configuration of lines. 5
As the quotation shows, Hanson concedes that organization is not seen as

lines and colors are, and he concedes that organization is not an element
of the visual field, and though it is not visible, it is intelligible. But
Hanson's concession is more than one would expect, and it effectively
takes him out of his original position, because the issue bears on what is
seen, or more generally, sensed. Few would deny that much that is
observed is not intelligible without an unobserved order; but the
intelligible is understood not sensed; understanding takes place in the
mind, not in the senses or imagination. Nor has Hanson even attempted
to show how that which is intelligible actually intrudes something into the
content of the observation itself. Furthermore, he does admit that two
observers with different theories can see the same thing; for he says that
when Tycho Brahe and Johannes Kepler look at the sun (one of whom
thinks of it as moving, the other as static)
. [which] perhaps constitutes their seeing the same thing. 6 both Tycho
and Kepler have a common visual experience of some sort ..
So he appears to concede what we want; yet that is not always true, he

says; for some people see Fig. I as a block of ice, whereas others see it as
an aquarium, etc. On that account we are asked to accept the following:
There is a sense, then, in which seeing is a 'theory-laden' undertaking.

Observation of x is shaped by prior knowledge of x. 7
Speaking more determinately about this last point, Hanson describes

seeing as follows:
There is a corresponding gap between visual pictures and what we know.

Seeing bridges this, for while seeing is at least a 'visual copying' of
objects, it is also more than that. It is a certain sort of seeing of objects:
seeing that if x were done, y would follow. 8
When he says that seeing is at least a visual copying and then goes on
to add that it is more than that, he tells us he is equivocating. "Seeing" has
an ordinary meaning according to which it has to do with "visual copying"
and refers solely to the sensation of sight; but although Hanson qualifies
his view by saying that "There is a sense . . . in which seeing is
theory-laden," he would make "seeing" signify something beyond what
it actually is and what the ordinary man understands by it. Given the
equivocation on "seeing" to mean at one time only the external sensory
act ofseeing and at another to mean the operations ofimagination and/or
memory as well, one might readily admit on the basis of such an
ambiguity that observation is "theory-laden." Moreover, if observation is
taken to include, in addition to sensation, the intellectual act of inferring
that y follows from x then for even stronger reasons observation is indeed
"theory-laden."
Other remarks he makes leave little room for doubting the
equivocation:
Our visual consciousness is dominated by pictures; scientific knowledge,

however, is primarily linguistic. Seeing is, as I should almost like to say,
an amalgam of the two c pictures and language. At the least, the concept
of seeing embraces the concept ofvisual sensation and ofknowledge. 9

It would seem almost certain that Hanson felt uncomfortable with what
he said; for he himself points out the equivocation. The seeing that is
theory-laden is supposed to be an "amalgam," a union of physical seeing
and something else; but plainly a union of two distinct things is not to be
identified with either of the united parts. Also Hanson appears to
recognize that he cannot make his point because he attempts to take away
the multiplicity when he says:
We see Fig. I as from underneath, as from above, or as a diagram of a rat

maze or gem-cutting project. However construed, the construing is there
in the seeing. The thread and its arrangement is the fabric, the sound and
its composition is the music, the color and its disposition is the painting.
There are not two operations involved in my seeing Fig. I as an ice-cube;
I simply see it as an ice-cube. Analogously, the physicist sees an X-ray
tube, not by first soaking up light and then clamping on interpretations,
but just as you see this page before you. 10
Without doubt Hanson thinks the intelligence puts something into the
content of sensation itself, and the reason appears to be his
misunderstanding of what is required for multiple operations. He seems
to think that if the operations were several, then they must be successive.
Simultaneously occurring, ordered operations do not appear to him to be
what takes place. Moreover, that he is inconsistent can hardly be denied
in the light of his saying that seeing "embraces the concept of visual
sensation and of knowledge," while maintaining at the same time that
there are not "two diaphanous components ... there are not two opera-
tions." But let us consider further what he has said.
First we should note that Hanson has not distinguished the several
cognitive capacities the human species has. That they are several would
hardly seem open to doubt, and that they operate together with a kind of
unity also seems not open to doubt. In fact, the close coordination of
activities is the reason why they are run together and why the distinction
of capacities is often overlooked. On the other hand, Hanson does
recognize the multiplicity, even though he wants to introduce the effect
of coordination into the act of seeing. Or more properly, he wants to
introduce the knowledge of the pattern, the order-which is something
known by the intelligence with the aid of other cognitive capacities-into
the operation of the senses.
But the failure to distinguish cognitive capacities is not peculiar to
Hanson; for the British Empiricists as a group do the same thing. John
Locke, David Hume, and George Berkeley, for instance, identify the mind
with imagination, and usually they are not taken to task for their mistake.
We are hardly surprised, then, to see an error such as Hanson's passed
over. In the next chapter we shall, in association with a discussion of
innate ideas and the a priori, discuss how one goes about determining the
distinction and number of operational capacities that a species possesses.
Now, however, we shall limit our remarks to some particular points
Hanson has tried to make.
When Hanson says it is a mistake to claim that Brahe and Kepler see
the same thing on the grounds that their eyes are similarly affected, he
does not give a reason why. He says only that there is a difference
between them, and apparently that is supposed to establish his claim. Now
let us concede at once that there is a difference between physical and
psychic activity, but that concession does not entail our having to grant
that people do not see the same thing when they are affected by the same
object. Such an argument amounts to maintaining that because two things
differ in species they do not belong to the same genus or category, and the
fallacy is patent. Sensation is passive in a very determinate way, as we
were at pains in an earlier chapter to show; and insofar as it is passive in
the face of a stimulus, it has something in common with purely physical
activities, such as being heated, undergoing magnetization, etc. So just as
we may infer that several bodies proximate to a Bunsen burner will all be
heated by it and that the heating is the same in kind (though not all bodies
will absorb heat at the same rate), we may also infer that two people
exposed to the same visual stimulus will undergo the same kind of effect.
In short, the primary effect of a visual stimulus, for instance, is the same
in each person, though secondary effects such as emotional responses may
well be different. Thus when the agents and passive recipients are the
same, the effects brought about are the same in kind, though they are not
numerically -individually-the same. But now to another matter.
On considering Hanson's claim that in the line drawing introduced
earlier one might "see" a transparent cube either from above or from
below, we now understand that "seeing a cube" involves more than the
sense of sight. In the drawing, sight attains a two dimensional continuum;
that is what we actually see on the page. But a cube is a three dimensional
continuum; and to "see" a cube in the two dimensional drawing one must
simultaneously employ his imagination, which is where such "seeing"
actually takes place. Furthermore, since the imaginations of different
people can do different things with the same initial observables, we ought
not to be astonished when they do not "see" the same thing.
Hanson also said that in the line drawing one man may see an ice cube,
another an aquarium, another a frame for a kite, etc. Ice we take to be a
substance, while an aquarium and a kite are artefactual things made of
glass and other materials that also are substances. The latter, however, are
not (in the strict sense) observable; only their properties are. We do not
see, feel, etc., water, glass, metals, and other substances; rather, we sense
their qualities of being wet, cool, mobile, hard, dry, etc., no one of which
is a stuff or substance. So because "ice cube" and "aquarium" and all other
such things include a substance or substances in their conceptions, we
must say that in the proper sense of the term-the one at issue here-they
are not observable; they do not themselves affect the senses; only their
properties do. As we said earlier, substances are understood by the
intelligence as a consequence of data provided by the senses, but the latter
are not themselves stimulated by the substance. If they were, we would
not err, for example, by confusing real gold with fool's gold. In short, we
do not see ice, an aquarium, etc., except insofar as "observable" or
"visible" is taken in the secondary sense of the term which includes in its
signification the intelligible stuff or substance in which the properties are
found. 11 If, then, we take "observable" in its secondary sense that includes
what only the intelligence attains-substance, for example-then we can
understand why Hanson would think that the intelligence does contribute
to observation. But as we have said, here more than sensation is involved.
Still another point of importance, one that we treated briefly earlier,
arises in connection with words that signify observables in relation to
unobservables, a point that we treated briefly in chapter 3 and which we
shall repeat here. It has to do with words that signify an observable along
with a relation of the observable to an unobservable. For instance, the
word "wrinkled" said of the brow is neutral in the sense that it signifies
only an observable configuration of the forehead; it has no unobservable
element. On the other hand, "smirk" is not neutral, for in addition to a
configuration of the lips, it signifies an unobservable internal state.
According to Webster's (2nd ed.), "to smirk" is "to smile in a conceited or
affected manner," and conceit is clearly an unobservable state or
disposition in the one in whom the smile occurs.
Many words are like that: "leer," "scowl," "chortle," "sarcasm,"
"groan," "scoff," and "scream" are only a few. "Groan," for instance, is
defined by Webster's as "a low moaning sound ... " and if the definition
stopped there, we would have only an observable element in the

signification. But Webster's goes on to add "usually a deep mourning
sound uttered in pain or great distress," and with those words an
unobservable element is introduced. Similar results are obtained when we
pursue the other illustrations above. Thus, it seems unnecessary to say that
when the observable and unobservable elements are run together people
can "observe" different things. To repeat: whenever an observable is
related as a sign, as an effect, as a cause, or in any way whatsoever to one
or more unobservables, then different people can "see," "hear," "feel,"
etc., different things by reason of the different unobservables that they
may associate with the observable element. This, however, does not imply
that sensation or observation itself is theory-laden; all it says is that the
mind is involved along with the senses.
Force perhaps requires some special mention because it has been put
forward as an instance of an observable. 12 Yet analysis shows that not to
be so. By "force" ordinary language does mean, despite the disclaimers of
some, a cause that is in something else; for the forces we "feel" are agents
or movers outside us. When we feel a force, we actually detect a motion
some cause brings about in us. Perhaps we are obliged to take a leaning
position against our desire, or perhaps we undergo a qualitative change
such as temperature. And since we rather easily relate the observables we
experience to their cause, force taken in its ordinary understanding is like
other examples discussed above.
There is another point to be made about certain words which are
somewhat different from those above but which also occasion someone's
saying that observation is theory-laden. The words in question primarily
signify properties that modify a substance and are used to identify it. For
instance, such words as "Negro," "Caucasian," "American," "musician,"
"carpenter," "boulder," "pebble," etc., appear to be like "man," "elephant,"
"chipmunk," "water," and others that signify substantive entities. But
reflection shows that the likeness is only apparent. "Negro" and
"Caucasian" signify "black man" and "white man" respectively, and that
which is principal in the signification is the color. If, however, we were
to categorize black and white we would say each is a quality, whereas
man is a substance (certainly he is not a quality). Similarly, a "boulder"
is a "large stone" and a "pebble" a "small stone," and each of these is a
similar sort of composite. Large and small are relative measures in the
class relation, whereas stone is a stuff. Hence the significations of all such
words are complex insofar as each includes a property together with a
substantive thing or stuff; and once again different people might "see" or
"observe" different things. But now back to Hanson.
Some of the examples that Hanson adduced ran properties together
with the realities of which they are properties. Music, for example,
although it is not a substance, is more than just sound. Music is ordered
sound, and order is an unobservable relation (as Hanson admits) perceived
by the intelligence; we do not hear the relations. More generally, any
organization of colors, lines, etc., is a set of relations that is unobservable,
even though the relations may be founded on observables. On that account
we can say that a visual, auditory, or other observation involves an
"interpretation" whenever the observation is accompanied by a
simultaneous recognition of an unobservable element by the intelligence.
And though the recognition is not an interpretation in the sense of an
explanation, it does require the activity of the mind as a cognitive capacity
distinct from the senses. So when Hanson says that (interpretative)
"seeing" is not an interpretation, only the meaning of "interpretation" as
explanation would make that statement true.
That the mind is involved in many judgments about unobservable
entities which have observable properties is perhaps the most commonly
occurring reason for saying that observation is theory-laden. When we
drive our car down the street just after filling the gasoline tank and fmd
that it comes to a stop, we quickly infer that something is wrong. We
observe the cessation of sound and motion, and then without hesitation
infer the conclusion. The inference is trouble-free and instantaneous, and
so we are inclined not to notice it; nonetheless it is an intellectual and not
a sensory act. Similarly, the neophyte prospector who thinks he holds gold
in his hand has inferred the substance from the properties he detects, and
his mistake is accounted for by his "interpretation," that is, by the
inference his intelligence has made from inadequate observations. The
same can be said of such claims as "I see a cat on a roof," or "I see an
aquarium," etc. In each of these cases the mind concludes to a substantive
entity from observable properties; and the entity is not observable in the
primary sense of the term but only in its extended, secondary sense,
according to which "observation" is indeed "theory-laden." Nonetheless
in those cases the "theory" does not contribute to the content of the
sensation itself, rather, the "theory" consists in an inference from the
observable.
Before we continue, let us note that obviously observable and
observation-term are not the same. It is one thing to say that an
observation-term is theory-laden, which can indeed be true, and quite

another to say that the act of observing is itself theory-laden. Since a
theory is something in the mind and requires a mental existence, the
observable as apprehended by the sense would have to have a content
contributed by the intelligence (a point we have made more than once),
unless we wish to claim that the mind extends itself to the reality and
modifies the external world. In short, observation taken in the primary
sense of the term cannot be regarded as theory-laden.
The ways in which we correlate an observable with something that is
not observed are instances of what H. H. Price calls "secondary
recognition" in contrast to "primary recognition," that is, observation.
Although Price thinks one is not entitled to say that secondary recognition
is inferential, nevertheless he does admit that it is derivative; and as soon
as the mind knows something derivatively, the possibilities of people
"seeing" different things arises; for some kind of movement must occur
(whether inference in the fullest sense or not) by which the mind gets
from the underived to the derived. Moreover, since many traits (when
taken alone or with too few other traits) allow, by reason of their
contingent connections, a number of different secondary recognitions, the
possibility of several people "seeing" different realities cannot be avoided.
To summarize, then, let us note that observation is theory-laden only
in the secondary sense of the term, taking "theory-laden" to refer to a
mental content suggested or signified by a word or term attached to an
observable, yet the mental content is not in the sense itself. There are,
then, some starting points we all have in common, though we may differ
even in the first "interpretations" we have of them. We might add, too,
that the tendency some have to put the intelligence within the sensory
activities points once again to the subjective difficulties that lead
philosophers like Kant to prefer the intelligence to the senses, so much so
that the content of sensory impressions has to be separated from reality
itself. The pre-Socratics did something similar, except that they denied the
reality of sensory properties and their changes.
C. Other voices
We have chosen to look at Hanson in some detail because he provided

us an opportunity to cover nearly all the points we wished to make. Others
have endorsed the view of observation as theory-laden, but in the main
they have done one or the other of the things we have already described.
Thomas Kuhn, for instance, says that Aristotle and Galileo in looking at
a pendulum do not see the same thing 13 because Aristotle sees
"constrained motion," while Galileo sees "free fall." The reader will
recognize, however, that this is only another instance of an observable
named in relation to an explanatory cause; and so he has introduced an
element that is not part of the observation itself.
P. K. Feyerabend takes another position. 14 He quotes Galileo when the
latter deals with an argument saying that if the earth moved, then a stone
dropped from a tower would not appear to fall on a perpendicular line.
Galileo grants the appearances, namely, that the stone does move without
any observable deviation from the perpendicular; but he asks that we use
the power of reason to confirm or disconfirm the appearances. The role
of reason in Galileo's mind is supplied by the following illustration:
One may learn how easily anyone may be deceived by simple appearance,
or let us say by the impressions of one's senses. This event is the
appearance to those who travel along a street by night of being followed
along the eaves of the roofs. There it looks to them just as would a cat
really running along the tiles and putting them behind it; and appearance
which, if reason did not intervene, would only too obviously deceive the
senses. 15
Before we consider Feyerabend's own remarks, let us note that Galileo

does distinguish sense from reason, and he also puts the error in the sense
itself, although reason can intervene to prevent the error in the sensory
apparatus. According to Galileo the corrective capacity of the intelligence
extends to the interior of the senses so as to affect its impressions. But as
we have noted in connection with Hanson, this poses serious problems.
If the mind is to correct the impressions, then in effect the intelligence
determines the character of the sensory activity; and one is hard pressed
to see how the observation is theory-laden. Instead, we should say the
sensation has been modified and is therefore a different observation. Or
perhaps to put the point another way, the intelligence-because it modifies
the character of the sensation-has become an observable.
Furthermore, the intelligence would have to be in a position to judge
the sensory evidence on the basis of what it, the intelligence, already
knows, which means its concepts would have to be a priori. And if that
were so, we could not know things as they are in themselves, which flies
against every spontaneous conviction we experience. 16
Another objection is that if the impressions can arise from the stimulus
and also be corrected by the mind, then the sensory capacities are the
locus of two different passivities, one regarding the stimulus, the other
regarding the mind. But such passivities would be of such radically
different orientations that we could hardly consider them to belong to the
same operational power. In other words, we would have two kinds of
"senses," one passive with respect to the activity of the observable
stimulus, the other passive with respect to the activity of the intelligence.
And just as the ability to be affected by gravity is different from the
ability to be affected by a magnetic field and the ability to be molded, etc.,
so the two abilities of the sense would also be distinct. But that the senses
should have two such passivities is not a tenable view; so now let us turn
to Feyerabend's own remarks.
About the illustration we quoted above he says:
To start with, we must become clear about the nature of the total
phenomenon: appearance plus statement. There are not two acts-one,
noticing a phenomenon; the other, expressing it with the help of the
appropriate statement-but only one [emphasis his], viz. saying in a certain
observational situation, "the moon is following me," or, "the stone is
falling straight down." We may, of course, abstractly subdivide this
process into parts, and we may also try to create a situation where
statement and phenomenon seem to be psychologically apart and waiting
to be related (this is rather difficult to achieve and is perhaps entirely
impossible). But under normal circumstances such a division does not
occur; describing a familiar situation is, for the speaker, an event in which
statement and phenomenon are firmly glued together. 17
There are not, Feyerabend maintains, two acts but only one when one says
in a certain observational situation that "the moon is following me," etc.
But how his claim is anything other than gratuitous is hard to see,
especially when a few paragraphs later he says (after granting what he
thinks is a doubtful assumption to the effect that sensations which enter
the body of science are independent of their linguistic expression):
Making the additional simplifying assumption, we can now distinguish

between sensations and those "mental operations which follow so closely
upon the senses," [quotation from Francis Bacon] and which are so firmly
connected with their reactions that a separation is difficult to achieve.
Considering the origin and the effect of such operations, I shall call them
natural interpretations [emphasis his ]_18
"Natural interpretations" are "mental operations which follow closely

upon the senses," but as we have seen, that does not warrant saying in an
absolute sense that they constitute one act with observation. (Nor does it
follow that the mental operation modifies the sensory reaction.) But let us
consider their union more closely.
A statue is indeed one thing, a unity of shape and material, but can we
say that because the shape and the material constitute one thing, therefore
the shape is not distinct from the wood or stone? Feyerabend grants a
distinction between the sensory and the mental when he says "appearance
plus statement (statements signify the mental)" and "mental operations
which follow closely upon the senses." So his point seems to be that the
union of the two is of such a character that by it the mental modifies the
sensory, which is like saying that the shape modifies the physical
properties of the material, or that an order, an organization, modifies the
properties of the parts which are ordered, a view that will not stand the
light of day. Certainly "gluing them together" does not allow us to make
such a claim. Sensations and the mental are indeed united by an order, the
sensations subordinate to the mental, so they might seem to be "glued
together." Being united, however, is not the same as being one without
qualification. In a way, a television technician who uses an oscilloscope
to diagnose a problem in a circuit performs one act when he diagnoses.
Yet absolutely speaking his act of seeing the wave-form on the
oscilloscope and the intellectual act of judging what the wave-form tells
him are distinct. The understanding of one of a comedian's jokes and
laughter are coordinated together, but who would say that the
understanding is the same as the laughter? In sum, Feyerabend's view is
like one of Hanson's that we talked about. He concludes that observation
is theory-laden because he runs the act of interpretation together with the
act of sensation, confusing an ordered unity with that which is absolutely
one. Thus we see again that the problem of distinguishing intellectual
from sensory knowledge runs into confusion that depends in part at least
on the difficulty the mind has in accounting for the origins of its
intellectual knowledge from sensory data. In modem philosophy the two
get run together, either by claiming as Hume did that concepts are washed
out images of sensory impressions, or by assigning as Kant did the
unobservable elements of intellectual knowledge to the intelligence itself,
or by having the mind contribute to the content of sensory impressions.
We ought to note before closing that other authors have discussed the
issue we have treated here, but we would have to say too much to present
their views in our context here. For instance, one might see W.V.O.
Quine, Theories and Things (Cambridge, Massachusetts: The Bellknap
Press of Harvard University, 1981), Chapter 2, entitled, "Empirical
Content." The reader would see that his discussion is in terms of
sentences, and that would entail much analysis before it could be made to
fit our essay. To analyze a position on observables, on properties of
realities, that would start from sentences would be lengthy, to say the
least.
One could also see W.H. Newton-Smith, The Rationality of Science
(Boston Massachusetts: Routledge & Kegan Paul), 1981. For similar
reasons this author too must be left aside, and so with this our
consideration of theory-laden observation is closed, and we shall now tum
to other matters. 19
Notes
1. Patterns ofDiscovery (Cambridge University Press, 1958), c. 1 No one else that

we know of provides as thorough and convenient a vehicle for examining the
theory-laden position.
2. Op. cit., p. 8.
3. Ibid.
4. Ibid., p. 9.
5. Ibid., p. 13.
6. Ibid., p. 7.
7. Ibid., p. 19.
8. Ibid., p. 29.
9. Ibid., p. 25.
10. Ibid., p. 23.
11. Perhaps we ought to recall at this point our earlier remark to the effect that
common substances such as water, salt, etc., are said to be observable in a
secondary sense, insofar as "observable" means any stuff or thing that has
observable (in the primary sense) properties.
12. See, for example, W. H. Newton-Smith, The Rationality ofScience (Boston:
Routledge & Kegan Paul, 1981), pp. 24-25; and Herbert Spiegelberg, op. cit.,
whose view we have already discussed, and for whom "force" signifies an internal
state.
13. The Structure ofScientific Revolutions (Chicago: University of Chicago Press,
1960).
14. Against Method (London: NLB, 1975), c. 6.
15. Text quoted by Feyerabend, ibid., p. 71.
16. We shall consider innate ideas and the a priori in a later chapter. Here we
content ourselves with noting that regularly men think themselves to know, to
some extent, things "out there" as they are "out there."
17. Op. cit., p. 72.
18. Ibid., p. 73.
Chapter 9
The A Priori and Innate Ideas
A. Introduction
That the mind constructs-at least in part-what it knows is a notion

that has been with us since the appearance of modem idealism. Its
companion doctrine, that the mind derives its concepts from God rather
than things, received its impetus from Descartes; and though Descartes'
view is not current, variations of it, or variations of the Leibnizian innate
ideas, are still on the scene. Of course in some measure scientific
hypotheses are constructs of the intelligence, but they do not justify an
idealist epistemology. Now, however, the time has come to show that the
human intelligence cannot be empirical in its nature and at the same time
be capable of a priori knowledge of any sort; or what is similar, an
empirical intelligence cannot possess innate ideas. Our discussion will
focus on an argument that covers all idealism, and particular theories can
be evaluated in its light.
B. The active and the passive: a general consideration
Some of the physical properties that serve to describe natural,

inanimate substances are active, while others are passive. Consider, for
instance, elasticity, conductivity, breakability, ductility, viscosity, etc.
Each of these properties involves a relative movement of one part of the
stuff in relation to the other, a movement that results from the action of an
external agent. 1 Elasticity is the ability of a stuff to return to its original
configuration after having been subjected to deformation by some agent;
conductivity obviously implies the transference of heat or an electrical
field from one part of a stuff to another because of a source exterior to the
conductor; breakability is the disposition of a substance to have its parts
separate as the consequence of a dividing agency; ductility is the ability
to bend and retain the shape after the action of an external force; viscosity
is a property of stuffs that flow, one part moving in relation to the another,
the movement resulting from gravity or some other agency exerting an
influence on it. And so we could continue. The point, however, is evident
enough: these physical properties, as well as others we did not mention,
imply an ability to undergo the action of some agent cause.
But certain spontaneous movements of natural entities brought about
by fundamental natural forces are determinative of things in a way others
are not. For instance, gravity, whatever its nature, is an active source of
motion in everything that has mass. It also is the cause that determines the
formation, shape, movements, etc. of galaxies and their components, as
well as the orbits and movements of planets. Speaking generally, gravity
may be said to determine the relative position of one body in relation to
another within the universe. And we see that the earth's constituting
materials exhibit an order that resulted from gravity at a time when the
earth was in a molten stage. Our point in all this is simply that gravity is
the major determinant of the character of the physical universe insofar as
the latter is constituted by an arrangement or position of bodies or stuffs.
Ultimately the earth's order is an arrangement of elements and
compounds, which are substances that are complete in species. 2
Another spontaneous natural movement is the one that is brought
about by the electrostatic or electromagnetic force that binds charged
particles; and because all elements are composed of such particles, their
electrostatic forces can be thought of as the "glue" which holds things
together. Nuclear forces, of course, play an analogous role. But our point
is a general one: to each of these active causes there corresponds an
appropriate passive property. Gravity has no effect on something that has
no mass; electrostatic forces do not affect particles or bodies that lack a
charge; a magnetic force cannot affect something that has no ability to be
magnetized. In short, natural activities of the sorts we have just mentioned
require their own passive principles in the bodies they act upon, and
conversely, the passive principles require a corresponding active source
in order to be affected. Gravity cannot affect a charged body by reason of
its charge, and a charged body cannot affect some other body by reason
of the other's mass alone.
If we now consider living things, we find that they possess active
abilities or capacities, and these capacities have operations which are
directed to definite kinds of objects in a clear and fixed way. Now
operational powers or potentialities or capacities are active; they are
The A Priori and Innate Ideas 117
abilities to do or produce something, while the physical properties of

inanimate things are passive. The genus of operational or active capacities
can be subdivided, as the reader will know, into (1) those that are passive
insofar as they require stimulation and specification by the
property-objects to which they are directed (for example the sense of
sight, which is stimulated by color-light, and the sense of hearing, which
is stimulated by sound); and into (2) those that are active insofar as they
produce something (for example the power of growth, which produces an
increased magnitude, or reproduction, which produces a substance).
Non-living substances undergo movements or modifications; they do not
operate, act on, or modify themselves as do living substances, which
initiate and terminate their own movements-the inanimate must always
be affected by something extrinsic. 3
We must repeat what we said before, namely that every operation is
intrinsically directed towards an object, whether the operational power is
active in the sense that it produces some physical effect, or whether the
power is a passive operational capacity, such as a sense, that is directed
to an object it does not modify or produce that which stimulates it. There
is no seeing without something seen, there is no hearing without
something heard, there is no loving without something loved. A similar
relation occurs in the active operational powers; for there is no
reproduction without something reproduced, there is no (self)
maintenance without something maintained; there is no growth without
an increased magnitude produced in that which grows. In short, every
operational capacity, whether it be related passively to its object insofar
as the latter must stimulate or specify its operations, or whether it be
active in the sense that it produces its towards which object-every one of
these operational capacities takes its character from the object toward
which it is oriented. The Aristotelians expressed this state of affairs by
saying that operations and potentialities of whatever sort get their specific
natures from the objects toward which they are connaturally directed. The
relation is important enough to bear emphasis.
Were we to attempt to defme some art, painting for instance, we would
characterize the art by indicating in some appropriate way the sort of thing
the art produces. Two dimensional representations using color and line are
the object-effect of painting; compositions of sounds are the object-effect
of music, etc. Medicine, electrical engineering, architecture, and other
such arts would be distinguished in the same sort of way. And were a
biologist in doubt about whether an organism had a sense of sight or a
sense of hearing because the appropriate organs were not observable, he

could test to see if the sensory powers were present by observing whether
the organism in question responded in the typical way to light-color and
to sound. In short, because such operational powers are inherently
oriented towards determinate kinds of objects, we say that they are
specified by the objects; and our proposition holds whether the capacities
are active or passive.
But all this has implications for the intelligence; for if our minds are
empirical and derive their conceptions from objects through the
stimulation the objects exercise on the senses, then the relation of the
intelligence to the entities it can know is in this regard essentially passive.
That is to say that although the intelligence is active insofar as it does
something, namely think, it is passive insofar as its conceptions are
specified by the realities that affect the senses that determine its concepts.
And we must stress that the proportion of the intelligence to such an
agency, the proportion ofthe intelligence to the sensible objects presented
to it, is inherent in its own constitution. Intrinsically, in its own nature, the
intelligence is passively directed to the kind of actions which the senses
can exercise on it.
But if, now, the intelligence were to derive its conceptions directly
from God, as Descartes would have it, then the passivity of the
intelligence would have to be radically different in its character. A
passivity or potentiality that is proportioned to the activity of God as its
natural4 source of concepts cannot be the same in its constitution as a
potentiality that is oriented directly to the actions of sensible objects that
affect us, any more than mass as a passive property can be the same as
charge as a passive property. In short, were we to derive concepts both
from God directly and from physical realities through the senses, then we
would have two radically different kinds of intelligence. Moreover, we
would each be two persons, not one.
The same can be said if one holds that certain concepts are derived
from the intelligence itself, for this would make it an agent in regard to
itself. Under such an assumption the intelligence would be intrinsically
different in character from one that is empirical. An intelligence that
knows directly from itself, by reason of its constitution, an intelligence
which is constituted able to derive concepts from itself or is constituted
as already possessing innate ideas derived from a non-empirical source is
not a passive power oriented to the activity of an extrinsic sensible object.
Hence were Kant or Leibniz correct, in either case we would know some
The A Priori and Innate Ideas 119
things by one mind and other things by a second mind; and again we
would be two persons, not just one. But our last remark requires
elaboration.
Some years ago the Russians were featured in certain popular
magazines for having grafted the head of one dog onto the body of
another, the grafted head functioning in some measure and deriving its
sustenance from the body of its host. Something similar occurs, of course,
when a calf is born with two heads. Furthermore, if in regard to the latter
we ask whether there are two calves or one (assuming both heads are
functioning), we have to reply that despite the obvious deficiencies, there
are two calves. The first criterion of individuality is the numerical
distinction of principal operations; so if there are two visual activities,
there are two organs, and if there are two acts ofhearing, again there are
two organs. More directly, if there are two distinct sets of principal
cognitive activities, then there are two dogs or calves. Similarly, if we had
two similar but numerically distinct intellectual actions we would be two
persons." To repeat, wherever the principal actions are numerically
multiplied, so are the entities, and were we in possession of two radically
distinct kinds of intelligence, we would be two radically distinct kinds of
person. It would seem, then, that the human mind is empirical and
unendowed with either a priori concepts or innate ideas, for none of us
has two or more intelligences or is more than one person.
Notes
I. We treat this issue at greater length in Substance and Modern Science, cc.
11-12.
2. Just as an arm is not complete in species because it is only a part of a
substance, it would seem that elementary particles, too, in a somewhat different
way but for even stronger reason, are not complete in species. That is why we
speak here primarily about the effect of gravity on elements and compounds.
Our remarks do not deny the reality of pulsars, black holes, and other esoteric
entities astrophysicists discuss, entities which are not composed of matter as
we know it here on earth; nor do our remarks deny the effect of gravity on
particles existing separately.
3. These notions, too, are discussed in S&MS, c. 23.
4. Descartes is not talking about a supernatural infusion of knowledge which
presupposes a natural mode of knowing as already established; that is a
different matter and not at all the issue here.
Chapter 10
Regularities, Experience and Experiment
A. Introduction: a note on systematic considerations.
As soon as one speaks of a systematic procedure, he suggests

conclusions drawn from premisses. We may express this same idea in a
more technical language and say that a scientific procedure draws
conclusions from axioms. The issue is stated in the following lines:
The problem of demonstrability of a science was solved by Euclid in so far as

he had reduced the science to a system of axioms. But now arose the
epistemological question how to justify the truth of those first assumptions.
If the certainty of the axioms was transferred to the derived theorems by
means of the system oflogical concatenations, the problem of the truth ofthis
involved construction was transferred, conversely, to the axioms. It is
precisely the assertion of the truth of the axioms which epitomizes the
problem of scientific knowledge, once the connection between axioms and
theorems has been carried through. In other words: the implicational character
of mathematical demonstrability was recognized, i.e., the undeniable fact that
only the implication "if a, then b" is accessible to logical proof. The problem
ofthe categorical assertion "a is true b is true", which is no longer tied to the
"if', calls for an independent solution. The truth of the axioms, in fact,
represents the intrinsic problem of every science. 1
That is the problem: how do we establish the truth of the premisses, which
themselves are not deduced from other propositions. That is the principal
issue before us in Part II, and we shall begin our discussion with a
discussion of regularities, which play a foundational role in every
systematic consideration that pretends to explain the world of nature.
Moreover, the regularities play different roles in science and philosophy,
and the differences are very important when it comes to the question of
premisses or starting points. And so to them we now tum.
B. Concepts
Our earlier chapters have considered how the mind's concepts of

unobservable realities originate in sensory experience. When we say
concepts, we are speaking of the simple conceptions of things that are
combined in propositions, as when we say "Man is an animal," by
composing the simple concept "man" with the simple concept "animal."
Now everyone knows that the formation of these simple concepts depends
upon the mind's ability to recognize similarities in individuals, its ability
to form classes on the basis of these similarities; for exercising that ability
is the first thing the mind must do. In other words, the recognition of
likenesses or similarities in things and their properties is prior to the
mind's movement from observables to unobservables. Now when we put
simple classes together we form propositions, and if they belong to some
systematic discipline, they will be "regularities" used as premisses and
conclusions. So the issue is how premisses and conclusions are related
directly or indirectly to their observational foundations. As we said, they
are related to observation through the simple concepts of which they are
composed. But our question goes beyond that because the union itself of
simple notions depends in one way or another on sensory data.
How premisses and conclusions are related to observation is an
important issue, and the relation differs in the natural sciences and
philosophy. For that reason we shall discuss first what we have already
called "regularities," for they ground the general propositions-some of
which are necessary and some of which are not-that constitute systematic
disciplines. The current question, then, bears on the nature of the
propositions that for;m the premisses and conclusions in science and
philosophy. So after considering regularities in general, we shall look first
at the relation of hypotheses to observation and observables because
hypotheses account for the regularities the sciences seek to explain, and
then we shall talk about analytic propositions and their relation to
observation and observables because analytic propositions are the ideal
starting points for philosophical explanations. To be sure, some
philosophers are likely to say that absolutely speaking the order we
propose is backwards, and we ourselves shall argue that philosophy
precedes the natural sciences because it systematically examines ordinary
regularities that are presupposed to every other discipline. But because
philosophy does not today have a determinate character and because it is
Regularities, Experience, and Experiment 123
not regarded by many as systematic, we have chosen to reverse the natural

order and talk first about scientific hypotheses and their relation to
observation for the reason that their procedures are more commonly
known than those that ought to characterize philosophy and its principles.
In what immediately follows, however, we shall talk about regularities in
a general way.
C. Preliminary reflections on behavior
No matter what the science, whenever we engage in a systematic

investigation of nature, we instinctively center our attention on the
behavior of the natural entities we are inquiring about, and we regard
behavior as our basic interest. Physicists, for instance, investigate many
kinds of behavior: mechanical, electrical, thermodynamic, the oscillatory
behavior of bodies that conduct sound, the movements of electrons,
protons, and neutrons, etc. Chemists investigate the reactive behavior of
compounds and elements, ethologists the instinctive behavior of animals,
physiologists the behavior of tissues and organs, botanists the behavior of
plants. The social sciences investigate human behavior of one sort or
another: sociology the behavior of man as a part of some society,
economics, wealth-producing behavior, ethics, moral behavior, and so on.
In short, whatever the systematic discipline, its predominant concern is
with the changes or activities-all of which may be signified by the one
term "behavior"-of the things or stuffs that constitute its subject matter. 2
Thus the scientifically relevant characteristics of entities are first their
behavior, and second the properties that are related to the behavior. Stated
negatively, to the extent a property has no relation to behavior, to that
extent it is scientifically irrelevant, which is well illustrated in biology.
Biologists use survival-the continued existence of the species -to
account not only for activities, properties, physiological processes, etc.,
but appendages as well, all of which are looked at from the point of view
of their being adaptive; that is, from the point of view of how they
facilitate survival. Survival, however, is the result of the nutritive,
defensive, reproductive, and other behavioral activities of individuals.
Thus when survival accounts for a property or an attribute or a part of an
organism, it does so by showing how the property supports the behavior
that preserves the populations of the species.
Stating the point more generally, everyone admits that in order to
know what something is we must first know what it does; and that is the
basic reason why behavior is central in most systematic disciplines. What

a thing does is the "opening" through which we read within it; and our
reading is productive in proportion to the "size" of the opening, that is to
say, in proportion to what the behavior can reveal through its own
observable character or its effect on an observable or observables. In sum,
most important for systematic considerations of nature is the behavior
proper to the subject-category with which the discipline is concerned; it
provides the discipline's point of focus. But now to our principal topics.
D. Regularities and experience
As we noted earlier, the sciences bear on regularities, and regularities

consist of correlations either between one kind ofbehavior and another,
between a kind of behavior and a property or properties, or between one
property and another. Standing in contrast to regularities are freak or
chance (in one sense of the term) events, which are not regular. And
though it is isolated, a chance event is in a sense explicable after its
occurrence. The character of the chance event as such precludes its being
foreseen, which means it is not explainable in the same way as a
regularity. Regularities allow us to make predictions about the future
because new cases will behave like those we have already examined,
which means that systematic disciplines attempt to account for the
behavior and properties that are common to the members of a class. But
they do not set out to discover and explain "laws" governing chance or
freak events.
Although regularities are correlations between real properties (we
include behavior under the more generic term "properties"), or between
real properties and the things to which they belong, they are signified to
the mind by propositions called "laws"; and in its principal meaning in the
natural sciences the term tells us of a correlation between a behavioral
characteristic and the thing to which the characteristic belongs, or between
one behavioral characteristic and another. But when "law" signifies only
a connection between or among "static" properties, the term is extended.
The statement "All men are two-legged" illustrates this extended use,
while "Chemical reactants combine in defmite proportion" illustrates the
principal sense, as do the statements, "A gas diffuses from a region of
greater density to a region oflesser density," and, "The volume of a gas
will increase or decrease in proportion to its temperature." Now Pierre
Duhem considers "All men are mortal" to be one of the most certain of
common sense laws; 3 and although the statement tells about a change that
occurs in the human (and other) species, it is not "behavior" in the usual
sense of the word, despite its obvious reference to a change that happens
to men.
The illustrations above all indicate the use of "law" to signify an
empirically obtained regularity that is obscure and requires an argument
to show why the law is so; in other words, a law of this sort becomes a
conclusion of an explanatory argument. But there is another use of "law"
according to which it signifies a regularity (usually behavioral) that is a
first principle and on that account an explanatory premiss, examples of
which are the law of universal gravitation and the laws of
thermodynamics. To be sure, most of the "law-premisses" or
"law-principles" are hypotheses, but that does not affect our point; for
empirical regularities that are not assumptions and serve as argumentative
starting points are also laws in this sense of the term.
In the ordinary speech from which its scientific use is derived "law"
signifies a legal directive or proscription that regulates human behavior
and is a directive issued by civil authority. Still, a civil law does have
something of the character of a proposition insofar as it communicates
what the citizens ought to do, or insofar as it represents the de facto state
of affairs in some society. Formally, however, a law is a directive that
obliges citizens to initiate or inhibit certain kinds of behavior under
specific circumstances. Hence the ordinary sense of the term must be
distinguished from the scientific, according to which "law" signifies a
regularity which we understand but which is not a directive we issue.
There are, however some caveats.
Although laws are expressions of regularities that can carry universal
quantifiers, we must be careful not to say that every proposition with a
universal quantifier is a law; for appearances can be deceiving. As will be
plain to many readers, if we were to say "All of the men in this room have
pneumonia," or "All of John's brothers are physicians," or again, "All of
the people in test group A showed symptoms of the disease," not one of
the statements would be genuinely universal, despite the occurrence of the
word "all" in each ofthem. On the contrary, every one of the illustrations
is a singular proposition, for in each of them the individuals about which
something is said are capable in principle ofbeing designated. And as the
reader will know, when a proposition talks about individuals that can be
designated, it is singular. The indication of more than one individual does
not alone warrant saying the proposition is general; it must also say
something about a class; that it, the proposition must signify primarily that
something belongs to the instances to which it can be applied, either as
pertaining to their defmition, or as following from or explained by the
definition. When we say "Elephants are able to remember," the
proposition is general and capable of universal quantification because
"able to remember" is understood to accompany the nature of elephant
generally, and we understand remembering to belong to the individuals
as a consequence of their having that nature. But when we say something
about a group that can be designated, we understand the attribute to
belong first and foremost to individuals. Put another way, at first we know
the attribute only as belonging to some determinate individuals, and
another step is required to go beyond them to say something about the
class commonly. Thus, test groups and the like provide singular
propositions because the propositions talk about something that can at
least in principle be designated, and by themselves they do not allow us
to move to the class generally, which means the quantifier in those cases
is not truly universal. It simply means "the whole of." Again: though such
a proposition signifies the whole of the group being described, it does not
signify or express a regularity. 4
When one recognizes that systematic considerations bear on
regularities and hence on classes, he sees at once that the object of a
theoretical scientific investigation does not to seek to understand the
singular instance in its singularity. The biologist who dissects a shark does
so with a view to discovering the common anatomical features of the
species and perhaps of vertebrates more commonly, but he is not
interested in the peculiarities that belong solely to the individual specimen
on his lab table. Similarly, a chemist is not interested in the individual
samples as such which he puts in a test tube in order to observe a reaction.
On the contrary, he is concerned with the samples only in the measure that
they illustrate the character of the reacting substances generally.
Applied or practical disciplines are a different matter, however, since
their concern is with doing or making something, and doing or making
always bears on the singular. Consequently, we ought to keep in mind that
physics and chemistry are not the same as mechanical and chemical
engineering. Engineers and other practical scientists may begin with
general, theoretical considerations, but they design things that will be
produced as individual entities. Perhaps the "most singular" of the
mechanical disciplines is civil engineering, for the civil engineer has to
design a bridge for this location on this river, etc. His individual
construction has to take into account all the individual conditions of the
location in which the construction will occur. In sum, there is a "polar"
difference between those disciplines whose concerns are solely investi-
gative or theoretical and those whose concerns are fully practical.
E. Experience: what it is
"Experience" is another word that has many meanings. Often it is a

synonym for "observation" and stands for what we see, hear, smell, etc.,
or for what we observe going on inside us. At other times it signifies an
emotional or affective reaction we undergo as a result of observing what
happens either to us or to other people, as when we say "To live in a
concentration camp is a terrible experience." We could perhaps find still
other meanings that are close to those just given, but the sense of the term
that is of interest here is one employed by Aristotle 5 when he distinguishes
men from other animals on the ground that the former are capable of
"experience," whereas the latter are not, except in a diminished way. This
kind of human "experience," he says, results from many remembered
observations in which the mind sees a common correlation. 6 Formally
taken experience is not the regularity, for the regularity is in reality not
in the mind .. Nor is it a law, which is either a proposition or that which
the mind conceives in the proposition. Once again: formally speaking,
experience is a multitude of remembered sensory impressions gathered
together on the basis of something they show to be common-a point upon
which Stromberg rightly insists-and experience in this sense is the basis
for inductive argumentations leading to laws. 7
F. Experience and problems
Problems about natural phenomena provoke wonder or curiosity; they

arouse the mind's desire to learn about obscure natural occurrences, about
the laws the mind has empirically discovered. Of course we are often
astonished, overjoyed, angered, overwhelmed, and mystified by chance
events; but we are not moved to ponder them in the way we ponder
regularities. 8 Chance events do not yield to the same kind of investigative
processes as regularities. And since learning begins with the desire to
overcome something obscure or mysterious about which we are ignorant,
it follows that the first consequence of experience is to provoke wonder
or curiosity and hence the desire to know. The first function proper to
experience, then, is to motivate the mind by moving it to its proper

activity. Such stimulation is distinct in kind from those accidental
rewards, those "reinforcements," those candy bars that are used to
stimulate small children and which are only temporary, unenduring
stimuli one must outgrow. Arousing curiosity is, therefore, an especially
important function of experience, given the difficulties that must be
overcome in acquiring scientific understanding. Thus it would seem wise
to begin scientific education with an exposure to and reflection on the
regularities that are foundational for the intelligence.
To return to the main topic, let us repeat that experience provides us
first of all with regularities which we formulate in laws, and from them
problems arise as questions that seek to know why things occur as they
do. As an obvious instance, once investigators discovered that chemical
reactants combined in definite proportions, they asked themselves why
that is so. Astronomers asked themselves why planets move about the sun
in elliptical orbits and why galaxies have spiral shapes; biologists wanted
to know why Arctic mammals are large and why elands are able to
survive without drinking. Plainly each of these questions came from an
empirically obtained proposition, and each of them shows that experience
moves the mind to scientific pursuits.
Because problems are so important as a natural stimulus for the
investigative process, human intelligence has often been described as a
"problem solving ability:" To be sure, the defmition has been predicated
of other animals too, a practice which would appear to remove
intelligence from the list of uniquely human capacities. Few people will
deny, however, that man stands out in the animal kingdom by reason of
his mental accomplishments; hence if one defmes the mind as a problem-
solving ability, he must keep a clear head about the equivocal character
of the phrase; for obviously no non-human species solves problems in
astronomy or quantum mechanics. The problem-solving capacities of
other animals are practical; they correspond to the needs imposed upon
the organisms by their respective patterns of instinctive behavior. On that
account we might with justice say that the drive most fundamental to man
in his uniquely human character is not hunger, sex, or drives for self
preservation-which other animals have-but rather, the desire to solve
problems that are purely theoretical and of interest for their own sake.
Even the ordinary man has some theoretical inclinations, however well or
ill he may be able to satisfy them and however secondary such interests
may be in his catalogue of aims and however imperfectly he may fulfill
Regularities, Experience, and Experimen4 129
those that he has. In the past some whole societies have forgone any large
measure of technological development solely because they preferred
non-practical cultural pursuits as their social ends. Thus in man "problem
solving ability" refers not exclusively but primarily to theoretical
activities that are legitimately goals in themselves; in that way man is
most fundamentally distinguished from other species. But one final and
important note: the obscurity of natural entities is due principally to the
obscurity of natural causes 9 together with the unobservability of natural
substances. Were everything in nature observable, there would be no
obscurities, no problems, and no arguments. No need would exist to
illumine one proposition through others that implied it.
G. Experience as supplying premisses
But experience does more than supply us with problems; it also shows
us regularities which, when expressed in propositions, function as
premisses in systematic accounts. It seems fair to say that stated in one
way the second law of thermodynamics is a generalization from
observation: "Heat passes from a body at higher temperature to a body at
a lower temperature." Whatever heat is ultimately conceived to be,
whether the fluid, "caloric," whether random molecular motion, etc., we
certainly must admit that when a cold body comes in contact with a hot
one, the cold body becomes warmer and the warm one gets cooler. That
is the first correlation we discover in relation to "heat transfer." Moreover,
if one asks how we know that the more sophisticated formulation of the
second law given by Clausius-it is impossible for any self-acting
machine to convey heat continuously from one body to another at a higher
temperature-is true, 10 he would have to start with the experience
described just above. Other illustrations of empirically obtained premisses
could be given, and later we shall point some out when we discuss another
topic. But for the moment, we shall make do with those we have
presented.
In summary, it appears that experience supplies some principles for
systematic procedures (we do not claim that all principles are so
obtained); so it would seem unnecessary to labor the point further.
Moreover, the mind's dependence on such regularities is plainly a
consequence of its dependence on observation, and that is central to our
considerations.
H. Experiment
It need not be said that a great deal has been written about the
characteristics of experiment. The control an experimenter exercises over
his observations, his use of instruments in making them, the active
relationship he has to nature when he varies objective conditions for the
sake of determining whether corresponding variations occur in the
property being investigated-all of these characteristics and others as well
have been described and discussed in detail by a number of authors.
Needless to say, a thorough knowledge of the character of
experimentation requires one to understand them all; but a remark by
Claude Bernard allows us to put aside most of them because it draws
attention to the feature of experiment that is most important here. We
refer to his statement that experiment teaches. 11
Because an experiment teaches, because it enlightens the mind in
regard to something initially obscure, it differs from that which is only
observation. To be sure, experiments are made for the sake of
observations, but they do more than allow us to observe. As we said
earlier, observations acquire a scientific or systematic utility when
memory collects them into what Aristotle called experiences; and that is
the key: experiment provides the mind with deliberately acquired
experience, thereby making man more fully the master of his own
learning. And because .experiment brings about deliberately acquired
experience, we have no trouble seeing why scientists are not satisfied until
an experiment or set of experiments has been successfully conducted by
several laboratories. Experiment, then, has the same teaching function as
experience in general, and it can lead to regularities that (1) give rise to
problems, or (2) that provide generalizations which can function as
premisses, or (3) that confirm or disconfirm hypotheses.
Some people are inclined to assign other roles to experiment, roles that
depend on the physical activities that are a part of nature's being pushed
to show herself in the laboratory. Much has been written about the
modification of experimental objects that occurs in certain experiments.
Sir Arthur Eddington suggested that some of the entities we "discover"
may well be the results of the experimental activities and not at all things
of nature. Certainly laboratory activities can modify realities so
extensively that the realities are outside their normal state. The last
elements of the periodic table are cases in point; for they are spoken of as
"artificial," and we do not encounter them outside the laboratory. They
Regularities, Experience, and Experiment, 131
illustrate our point even though they are not "by products" of experiments
intended for other purposes.
To the extent that instruments may modify the objects which they help
us to know, to that extent they impede nature's teaching function with
regard to us. But no matter the role of instruments, no matter the
modifications brought about by experimental activities, no matter the
unnatural state of some of the end results, the essential nature of an
experiment is to instruct the intelligence; that is why we perform them. So
to repeat our main point: the goal of experiment is the acquisition of
experience, which means that experiment may appropriately be defmed
as deliberately acquired experience; that is what it is and why it teaches.
Notes
1. Hans Reichenbach, The Philosophy ofSpace and Time tr. Maria Reichenbach
and John Freund (New York: Dover Publications, Inc., 1957).
2. Even when there is no behavior in the proper sense of the word, we sometimes
speak as if there were. For example, mathematicians occasionally talk about the
behavior of sets or functions when in point of fact they are speaking of relations.
In that case the word "behavior" is predicated as a metaphor. Also, in the course
of defining what color is, one could very well ask himself what a color "does."
It reflects light, of course.
3.. The Aim and Strncture of Physical Theory, trans. Philip P. Wiener (New
York: Atheneum, 1962, p. 144.
4. It need not be said that the so-called "accidental universal" is not properly a
universal at all, as its name indicates. It amounts to a "counting" of the
individuals.
5. Metaphysics, Bk. I (alpha major), c. 1.
6. The medievals had much to say about experience in this sense, and those
interested ought to see a fine article by James S. Stromberg, "An Essay on
Experimentum," in the Canadian journal, Laval Theologique et Philosophique,
Vol. XXIII, 1967, No. 1; and Vol. XXIV, 1968, No.I. Needless to say,
Stromberg treats the issue at length and in detail.
7. It is perhaps apparent that a man can have a memory full of observations but
lack experience in the sense we are discussing because he may fail to see the
correlation, the regularity they illustrate. Twins could in principle have had
precisely the same sensory observations yet only one of them have had
experience in the sense at issue here.
8. The word "chance" has a number of meanings, and we hope no reader will take
to mean a statistical probability. We discuss this topic in our Nature's Causes,
where we point out that the relevant sense in natural events is that some
unforeseen effect comes about from an incidental union of two agents whose
union was not foreordained.

9.The reader may see our Nature's Causes (New York: Peter Lang, 1995). The
causes of natural things and processes that are employed by the sciences are the
object ofthis work, an important aim of which is to show that the four causes first
described by Aristotle are actually those that are employed in scientific
explanations.
10. This wording is from David Halliday and Robert Resnick, Physics (New
York: John Wiley & Sons, Inc., 1962, p. 539.
11. An Introduction to the Study ofExperimental Medicine, trans. Henry Copley
Greene, A.M. (New York: Dover Publications, Inc., 1957), p. 5.
Chapter 11
Hypotheses and Reality
A. The relevant features of hypotheses
We are now in a position to discuss the flrst of our two main issues,
which is the relation of hypotheses to observables; and in a particular
way, the issue bears on the relation of theoretical entities to observables,
because theoretical entities seem to be constructs of the intelligence that
in the minds of some have little or no relation to observables. That is to
say, they seem to originate from the intelligence itself, and that will be the
main focus of our discussion here.
Philosophers have written much about the reality of theoretical
entities, about particles and forces, and their main positions go under the
titles of realism and antirealism. Many aspects of the issue have been
considered, and no one could consider them all even in a reasonably long
book. Here, however, we wish to talk about hypotheses, and though in
some ways they are well understood, it nevertheless happens that some
philosophers regard starting points of "metaphysical" philosophical
theories to be hypotheses when they are not. In his day too Newton had
much to say about what appeared to be opinions on hypotheses that would
not stand the light of day. 1 So it seems desirable to make very plain
exactly what we think hypotheses are, and later in another chapter to show
why this is so. Our task, then, will be ( 1) to show the function of
hypotheses as explanatory premisses; (2) to show in what measure
postulated realities, whether forces or elementary particles, are real; (3)
to show how they are related to observation and observables.
Let us begin our discussion with a commonplace illustration. Suppose
that Socrates returns home one day to fmd that his lawn mower is missing
from his garage. Because he has an inquiring mind, he wishes to
understand why it is missing; but we must quickly add that any attempt
he makes to account for its absence will require that he assume some
proposition containing a plausible cause as a premiss from which to

deduce the conClusion, "My lawn mower is missing from the garage."
Obviously more than one alternative will occur to him, the first perhaps
being "My lawn mower was stolen." Another possibility is, "My lawn
mower was borrowed by my neighbor," while still a third is, "My lawn
mower was given to the rummage sale by may wife." Each of these
propositions contains a possible cause and can function as a premiss in an
argument that answers the question why, which means that the
intelligibility of the premiss comes from the causal relation it declares.
To be sure, choosing among the alternatives presents Socrates with a
difficulty, since each of them is plausible and each satisfactorily accounts
for the absence of the lawn mower; and as everyone knows, no direct
verification establishing any of the premisses as the truth is possible.
Indeed, the only proposition that is empirically verifiable is the
conclusion, the proposition with which we began, and as is well known,
affirming the consequent permits no necessary inference that will
establish the premiss. Thus on purely canonical or purely formal grounds
(in an Aristotelian sense as well as the modem), no choice can be made
among the premisses.
Nonetheless there is a way to exercise a choice, and it requires
Socrates to make certain "predictions" that accord with each of his
assumptions. First, if he assumes that the lawn mower was stolen, then he
can predict that marks of forced entry will appear on the door or window
of the garage, which provides him with a proposition he can verify by
observation. Second, if he assumes his neighbor borrowed the lawn
mower, then he can expect to see that his neighbor's grass is cut; and
again he has a proposition subject to empirical verification. Third, if he
assumes that his wife gave the lawn mower to the rummage sale, he can
expect to find a receipt in the tax file; and once more his prediction is
open to empirical confirmation or disconfirmation. But if Socrates now
sets out to determine which ofhis predictions is confirmable, he will again
be faced with affirming the consequent. This time, however, he hopes that
one premiss will allow the deduction of two conclusions, while the
remaining two account only for the initial observation, the absence of the
lawn mower. And since two are better than one, Socrates opts for the
premiss which gives him two observably verifiable conclusions, which is
an entirely reasonable way to make a choice.
Because the explaining premiss was chosen on the basis of a procedure
that is formally invalid, because no necessary inference about the truth of
Hypotheses and Reality 135
the premiss follows from affirming the conclusion, Socrates cannot be

certain that his account is true. Yet if he is willing to settle for probability,
then he can accept the procedure and take his stand on the premiss he was
able to confirm twice. Formal invalidity removes necessity from the
inferential process but does not render the argument valueless. That is to
say, "valid" means that the conclusion follows necessarily, but it does not
mean that there can be no probable relation between premiss and
conclusion, and so the absence of validity does not totally exclude the use
of probable, non-necessary patterns. Defect there is in such a procedure;
yet, as we said, the argument is not useless. In short, affirming the
consequent allows us to accept the premiss from which the conclusions
follow with greater or lesser probability, but it never will permit us to say
we are certain; that is the common logical doctrine.
Having now presented our illustration, let us recapitulate the points of
interest. First, we should note that the assumption of a premiss or
premisses is necessary because the phenomenon to be explained, the
absence of the lawn mower, does not of itself imply the precise nature of
the cause. Certainly we can make the very general inference, "My lawn
mower is gone from the garage, therefore some agency moved it," and
such a conclusion is certain because we know very well that the lawn
mower cannot remove itself. Put another way, from the generic nature of
the effect we can argue with necessity to the generic nature of the cause.
However, a number of specifically different agencies might have been
responsible, and on that account no necessary inference from the effect to
a specific cause is possible. We are obliged, therefore, to "guess at it" and
then choose among our guesses, which means that the invalid formal
procedure is a consequence of the inherent obscurity, the inherent
inability of the effect to manifest one and only one cause as its producer.
And let us add that Aristotle went to some lengths in his Prior Analytics
to show that the formal patterns of syllogistic arguments must be
proportioned to the kind of evidence available.
A second point to note is that although the possible premisses are all
assumptions, every one of them tells about something of which we have
direct experience. We all know about thieves, neighbors who borrow, and
wives who give things to rummage sales; hence the premisses are not in
principle beyond the range of empirical verification. Instead, they are
assumptions only because they cannot be directly verified and their
application in this instance is uncertain. And so with our illustration
before us and our points delineated, we may now go on to take up
hypotheses in the natural sciences.
B. Hypotheses and theoretical entities 2
One need not be an historian of science to know that chemistry

developed later and more slowly than physics; but by Dalton's time a
number of regularities were known, regularities which for the first time
made an atomic theory plausible. Consider, for instance, the law of
defmite proportions, which says reactants combine not according to
varying proportions but only according to proportions that are fixed, or
constant. The law of multiple proportions is similar in that reactants which
combine to form more than one compound do so according to quantities
that are multiples of those that combine to form the simpler compound.
These regularities, together with others, suggested to Dalton that an
assumption which took reactants to be composed of particles called
"atoms" would account for the observed behavior. Dalton's position was
similar in important respects to that of Socrates in the illustration we
employed above. There, the reader will recall, the assumption was
necessary because the cause could not be inferred from the effect, and the
same state of affairs confronted Dalton. If we put ourselves in his shoes
and ask, "Why do reactants combine in fixed proportions?" we see that we
cannot deduce the cause from what we have observed in the laboratory.
But we do know that the phenomena must have a cause; that is, something
real must be responsible for the behavior of the stuff. Fictional
causes-purely conceptual instruments-cannot bring about the real
changes experience reveals. Moreover, we know before we start that the
cause is somehow intrinsic to the substances which react; we know that
somehow the behavior of reactants is rooted in the substances that
undergo the reactions and is not the consequence of some wholly extrinsic
influence. In sum, we know both that there is a real cause and we know
that it is to be found within the reacting substances. Also we can infer the
general nature of the cause and say that the discrete behavior of reactants
implies that the latter are composed of particles; such a necessary
inference appears to be warranted, even without further elaboration. The
precise character of the particles, however, cannot be so deduced, and that
is the true obscurity. So if we are correct in what we have said above, then
clearly atomic hypotheses have to do with what the particles, the atoms,
are. The question is not whether there is a real cause, nor whether the real
cause is somehow in the stuff, nor is it about the most general nature of
the cause; for particles of some sort there are. It is, as we said, a question
of the nature of the particles. What, then, are the defming attributes of
those particles? That would seem to be the issue.
Dalton assumed that among other things the atoms or particles were
solid and indivisible, from which it followed that bombardment should
not show elementary substances to be porous, nor should it be possible to
break something off from their atoms. Experiment, however, proved
otherwise, and so the particles that entered into chemical reactions were
soon known not to be elementary; that is, they were known to be
composed of other particles. What that meant, of course, was that the
definition of the chemical particles had to be modified to show their
composition. Let us repeat: the experimental evidence did not show there
were no such particles; it showed only that their character had been
wrongly conceived, requiring that their descriptions or defmitions
undergo modification, a process that still continues, leading as it has to the
realm of charmed quarks.
To consider, now, hypotheses from the point of view of modus tollens,
we know that they can be disconfumed with certainty because such an
argumentative process warrants a necessary inference. One may argue that
if( chemical) atoms are indivisible (uncomposed), then no particles can be
broken off from them; but experiment shows that particles are broken off;
hence (chemical) atoms are not indivisible (uncomposed). The destruction
of the consequent does not destroy the entire conception of chemical
particle but only a part of the concept, which plainly accounts in large
measure for the steady improvement of the atomic theory over the years.
With each new development only a part of the earlier definition or
description is modified; a substantial remainder is left. The very general
notion that natural substances are particulate is left undisturbed, and the
steady development of the theory shows in more and more detail how that
general statement is true. In short, the theory approaches, as a kind of
limit, a more and more perfect understanding of the first, most general
proposition bearing on the particulate constitution of natural substances.
It approaches a precise understanding of what the elementary particles
are; yet we shall never be certain that we have attained the limit.
And so it would seem that neither a straight yes nor a straight no can
be given to the question, "Are elementary particles real?" Each time a new
one is discovered, it is discovered because some observable effect
(directly or indirectly, with or without intermediary instruments) has been
isolated; and no effect is capable of bringing itself into existence.
Necessarily, then, something real and particulate is the cause, but exactly
what the nature of the new particle-cause is remains uncertain. At this
point we are constrained by the limits of our defective process, and we
cannot assert that we know the particles to be constituted exactly as we
conceived them. Thus in a limited way we know theoretical particles to
be real, and in another way we are uncertain about them.
The character of hypotheses as we have described them seems to be
well expressed by a physicist who discusses the earth's composition,
having to start from considerations of its diameter, the precession of its
axis of rotation, its moment of inertia, its volume, mass, and average
density:
As we had guessed, the moment of inertia of the earth is less than it would
be if its density were completely uniform. This indicates that the mass is
concentrated toward the center, but does not allow us yet to find the
detailed distribution; it is useful in eliminating many distributions that
might be imagined without giving a single unambiguous answer for the
actual distribution. Nevertheless, on the basis of the results already
established, we can begin the construction of provisional models for the
density distribution inside the earth; models which will almost certainly
have to be revised as evidence accumulates. 3
This illustration seems to make our point, although we admit that it does
not bear on theoretical entities. Nevertheless, the difference between
understanding the composition of the earth in a general way in contrast to
understanding it in detail illustrates the point of interest. The physicist, on
the basis of what he has established very generally about angular rotation,
knows that the mass of the earth is unequally distributed; he has no doubts
about that. The precise character of that distribution is, however, obscure.
If we now look back at our illustration involving Socrates and his lawn
mower, we see an important difference between the assumptions which
account for chemical reactions as well as other such phenomena and the
assumptions which accounted for the absent lawn mower. Atoms and their
constituents enter into propositions which are in principle incapable of
being verified inductively, that is, by observation, because the effects to
be explained are not sufficiently intelligible for us to see in them the
precise nature of their cause. On the other hand, although the assumptions
Socrates made to account for his absent lawn mower were not verifiable
in that instance, nevertheless considered generally they are empirically
verifiable and are therefore independent of the argument in which they
occur. Furthermore, the causes they contain are known to be real and their
character understood, whereas the causes in current scientific hypotheses
are inherently obscure. To go beyond our example: if one calls the
assumptions of the social sciences "hypotheses," he is talking about a
different kind of animal; such hypotheses are not of the same nature as
those of physics and chemistry.
C. Forces: a case of metaphor?
To be sure, the discussion of theoretical entities is not confined to

particle-constituents ofnatural substances. Gravitational, electromagnetic,
nuclear, and other forces are also objects of consideration; and their
ontological status has likewise been debated, again with arguments on
both sides. But despite the many detailed and particular issues one might
discuss in connection with this topic, it seems to me that a fundamental
point can be made that throws light on the direction in which the
discussion ought to go. Let me begin with some remarks on gravity as a
force.
Newton used his three laws of motion together with his law of
universal gravitation to account for the motions of the planets insofar as
those motions were expressed in Kepler's laws. When, for example, we
consider that the second law tells us that a radius from the sun to a planet
will sweep out equal areas in equal times, we spontaneously tend to ask
the question why. And since the question seeks an explanation, there is no
doubt that the law of universal gravitation is explanatory insofar as it
responds to the question asked and tells us why the regularity is so. And
though we have used the second of Kepler's laws as an illustration, the
first and the third are similar. Newton's theory of gravitation accounts for
the orbits of the planets being ellipses, and it accounts for the square of
the periods of revolution being related to the cube of the mean distance
from the sun. On that account, then, are we not entitled to say that gravity
is a cause?
It is well known that many think Newton's theory of mechanics does
not explain the natural motions ofbodies but merely describes them. They
claim the theory presents the data in an organized way but does not offer
a cause to account for natural motions. Hypotheses non jingo, Newton
himself said in answer to the question, "What is gravity?" Thus it seems
the theory is not explanatory but only descriptive. But can we have it both
ways? Can the theory answer the question why and still not tell us the
cause? Or we might put the question another way and ask, just what does
the theory do?
In view of the large amount of literature that has appeared over the
years on the nature of scientific knowledge, it might seem presumptuous
on our part to offer an answer to the question just raised in the small
amount of space we shall utilize here. Yet we wish once again to remark
that our aim is to consider the issues only insofar as they are related to the
general thesis of this essay, namely, the way in which the intelligence
learns about unobservables through observables. On that account, then,
the present considerations will be limited; and we shall not dispute the
many positions that have been taken on the large number of topics
associated with the general issue but instead confine ourselves to a few
points that are rather directly related to the discussions that have gone on
in earlier chapters.
In line with what we said earlier, we ought to note again that the
question whether something is must not be identified or confused with the
question what something is or with the question why it is. Each of these
is distinct, and the first can be answered even when the others cannot.
Moreover, the reader will recall that the inference from an observable to
an unobservable cause is precisely of that nature: it is an inference to the
actual existence ofthe cause, not to its particular nature. On that ground,
then, I would like to argue once more that certain distinctions must be
made apropos the reality of theoretical entities; for there is a sense in
which we are certain that theoretical entities exist, that they are real, and
there is another sense in which we cannot know that they are anything
more than constructions of the intelligence.
If we consider planetary motion, for instance, we know from the start
that the motion is not responsible for itself. It does not determine itself to
be elliptical, it does not slow itself down in one part of the orbit and speed
itself up in another, etc. In short, we are certain that the motion and its
characteristics do have a real, an existing cause. Again, real effects cannot
issue from fictional or other non-existent causes, and so we do know that
there is a referent for "cause of the motion." To sum up: we are certain
that the motion of the planets, the falling of bodies, the movement of
charges, etc., all have some real cause, to which is attached the general
name "force."
The force responsible for planetary and other, similar motions, has
received the name "gravity." But once it has been named, what more can
we say about it? What additional attribute or attributes other than
existence can we assign to it? Such was indeed the quandary Newton
found himself to be in; he could determine no defining or descriptive
attribute of the cause and so declared that he did not fashion hypotheses.
Thus "gravity" remained for him and appears to remain for us a
metaphor. We do not have an analogy, the reader will recall, until we can
extend the naming word to the unobservable cause on the basis of
something belonging to the latter that is distinct from the characteristics
of the observable effect. But when gravity and other forces are conceived
as fields, then they are conceived in the mode of substance, a point we
discuss elsewhere. 4 If, then, gravity or any other force is postulated to be
a field and thus a substantial entity, the notion of the force would seem to
lose its metaphorical character, at least in a significant measure.
At this point it seems advisable to repeat, perhaps with some emphasis,
a point we made earlier when we said that the obscurity of the effect is
precisely what demands that we introduce hypotheses in order to explain
gravitational motion. When we consider movement that is change of
position, we see that as a reality it is minimal. That is, the "ontological
content" of such motion is next to nothing. Motion, either translational or
rotational, does not introduce modifications into the physical properties
of natural substances; it does not produce a qualitative alteration of
properties, as does heat. This this inherent ontological obscurity, owing
to the impoverishment of motion as something real, is the proximate
source of the mind's inability to infer the nature of the cause from the
effect; and the deficiency is not so much one of the mind as it is of the
reality known. After all, an argument cannot yield in the conclusion more
than is contained in its principles. In sum, then, the mind's inability to
infer the character of the cause, its inability to get beyond a metaphor in
its conception of gravitational force except to say that it is a field is the
consequence of the radical unknowability of the effect upon which the
inference to the cause depends.
As evidence that what we are saying here is actually the course the
mind must follow, consider the development of the notion of force in the
mind of Kepler. According to Max Jammer, Kepler's early conception of
force made the latter to be a soul, a conception he later modified as
follows:
In his letter to David Fabricius of November I 0, I 608, Kepler clearly

envisages the forces of attraction exerted by the earth on a stone as
magnetic lines, or chains, as he says, thereby approaching Gilbert's
conception of gravity as a magnetic emanation. 5
Jammer then quotes Kepler as follows:
How is it possible that a sphere, thrown vertically upward-while the earth

rotates meanwhile-does not return to the same place? The answer is that
not only the earth, but together with the earth, the magnetic invisible
chains rotate by which the stone is attached to the underlying and
neighboring parts of the earth and by which it is retained to the earth by
the shortest, that is, the vertical line.
And so from soul to chains to ... ? Truly, we cannot escape the use of
metaphors. But if our knowledge of the nature of gravity does not pass
beyond being metaphorical, in what way can gravitational theory be
explanatory? Is it not, as is claimed, merely descriptive?
In reply, let us note that given our analysis above, we can say that the
theory is explanatory insofar as it tells us how the cause operates; that is
to say, when we know that the planets, as well as all other bodies, "attract"
each other with a "force" that is proportional to the product of their
masses and inversely proportional to the square of the distance separating
them, then we have described the mode in which the cause -whatever its
nature-operates. But of course that is not an explanation in the full sense
of the term; it is only partial and obliges us to say, therefore, that we have
a description of the behavior of the cause through its effects but that we
do not know the inherent attributes of the cause and so do not have an
explanation in the full sense of the word. Put another way, we know that
there is a referent, but we do not know its character, its nature; hence we
have an explanation in a diminished, qualified sense. So it seems to us that
the realist and the anti-realist are each partly right, which makes our
answer somewhat similar to but also somewhat different from the one we
gave in connection with particles. Our knowledge of the latter is not
altogether metaphorical, since we know that they are parts of substances;
still, we cannot say without qualification that the particles as conceived
are real, since we cannot determine that our definitions describe them as
they actually are. It would seem, then, that from the realist point of view
our knowledge of those forces which are understood in no greater
measure than gravity is inferior to our knowledge of elementary particles,
at least in some measure.
It might seem that at times theory has missed the boat on both
accounts, failing even to determine that a cause exists. An example which
might appear to make the point is the ether, which was postulated as a
medium for the propagation oflight waves. Yet we may argue that further
consideration shows that not to be the whole truth. Certainly experiment

failed to show that an ether existed; yet no one denied that the propagation
of light required a real, existing cause. The difficulty came, of course,
when the theorists hypothesized that the propagation was the oscillation
of a medium-substance and then later had to say that it was an emission
of particles (having, to be sure, an associated wave function). But I would
argue that even this pertains to what the cause is, not whether it is. A
medium of some sort is indeed implied if one assumes the propagation of
light to be a wave motion similar to that of sound; the notion of wave up
to that point carries such an implication. Since, however, light manifests
not only wave characteristics but particle characteristics as well, can we
determinately assign anything even as a category? except perhaps the
general notion of a field, the precise character of which is unknown? Have
we then, surpassed a metaphor? To repeat: all of these considerations
appear to relate to the question which asks what light propagation is, not
whether there is a propagation of some sort. One might rightly emphasize,
I think, that too often we neglect to consider that the formulation of
scientific hypotheses presupposes the existence ofthe causefor events and
phenomena as known.
D. Hypotheses and observables
At this point we may note that the formulation of hypotheses which

say what certain things are depends on the imagination and its ability to
construct images that do not have known counterparts in the real world.
That is to say, the model, the postulated entities are constructs formed by
the imagination from the observables already present within it and derived
from ordinary experience; and so in a limited, qualified sense they are a
priori. But nothing in the hypothesis stems solely from the intelligence,
or from the cognitive powers independently of the real world. This, then,
seems to be the way that the "background knowledge" plays a role in the
formation of theories. In other words, the background consists of the
common concepts derived from observation that provide the raw
materials of imagination and conceptual construction. The background
:;onsists of those regularities that are common to aH ~atural entities and so
are not modifiable by the particularizations of hypotheses. No matter that
the imagination fmally becomes inadequate in the sense that the known
effects make demands on the theoretical entities which the imagination
cannot fulfill, demands which the imagination cannot meet by
constructing an image adequate to them all. The causes are so tenuous,

their effects so disparate, that the nature of the cause escapes adequate
formulation, even by a well-endowed imagination and an inventive
intelligence. Still, that does not affect our main point; whatever the degree
of success, the attempt is to say what the cause in question is.
Finally, the unity holding many hypotheses together in one theory
comes from the attributes which define the theoretical entities or entity,
for they provide most of the hypotheses. The many are ordered insofar as
all are directed to manifesting what some one real cause is. In the case of
elementary particles, the unity and order come from the fact that the
particles are conceived to be composing parts of one stuff or thing.
We have now fmished our discussion of the features ofhypotheses that
are relevant to the central issue of this work, but it remains for us to make
a general remark about realism as distinct from instrumentalism and
various forms of antirealism. The reader will know that our position on
hypotheses is basically realist because we have maintained that
hypotheses are not totally false and their constant improvement
presupposes that something of the reality they present is true. Also we can
say without qualification that the fictional insofar as it is fictional cannot
account for what is real. One is obliged to recognize, it seems, that
hypotheses in the natural sciences become ever more probable as they
account for more and more (instrumentally) observable regularities
precisely because they attain reality in an ever-increasing measure, despite
our inability to confirm them in a perfect, apodictic manner.
Notes
1. See E.A. Burtt, The Metaphysical Foundations of Modern Science (Garden

City, New York: Doubleday & Company, Inc., 1934), chapter VII, p. 215ff.
2. A main point around which our exposition revolves has been made by William
Wallace, Essay VIII, "Are Elementary Particles Real?", in From a Realist Point
of View, (Lanham, New York: University Press of America, 1983), p. 171. See
also Essays VI and IX, which deal at greater length with some of the points we
just touch upon.
3. 0. M. Phillips, The Heart of the Earth (San Francisco: Freeman, Cooper &
Company, 1968), p. 75.
4. See Substance and Modern Science, pp. 59B60.
5. Concepts of Force, a Harper Torchbook (New York: Harper and Brothers,
1962), p. 81.
Chapter 12
The Mathematical Character of Physics
A. Introduction
In order to say in an adequate way why natural philosophy differs

from the natural sciences, we must begin by showing first why physics
became mathematical. Although in this context the topic may appear in
some respect so to be a strange one, it will nonetheless help us throw light
on the role philosophy plays in understanding natural entities and natural
phenomena. That the question merits our attention is best seen when one
considers that the history of modem philosophy differs greatly from the
history of modem science, which is perhaps well illustrated by Burtt's,
The Metaphysical Foundations of Modern Science, to which we have
already referred. "Metaphysical" will indeed appear to many to be a
strange word to use in discussing such foundations, especially when the
main issue appears to be the mathematization of natural science. Yet to
the extent that reality becomes understood as basically extension and
number, the word "metaphysical" is perhaps appropriate. But as the reader
will see, we have no intention of doing a history of modem science, and
the point we shall ultimately make is a simple one that does not require
extensive considerations on Newton's or other philosophers opinions on
the nature of space and motion. All such issues can be left aside. So now
to our task.
B. Some historical preliminaries
The inhabitants of Academe are likely to know that the origins of

science and philosophy are the same, both having started with Thales, the
Greek theoretician who attempted to account for natural entities by
hypothesizing, Aristotle tells us, that they were produced out of water
through variations in t4e latter's density. Speaking roughly, one may say
Page 146 Observables and Unobservables
that because Thales' speculations embraced the whole of nature, and

because he attempted to understand things through first principles, some
Academicians look upon him as the founder of philosophy. On the other
hand, because his speculations explained natural things in terms of a
material constituent (he proposed the first theory of matter) other
Academicians look upon him as the first scientist. And what can be said
of Thales can also be said of the other pre-Socratics, for as a group they
are regarded by historians as both scientists and philosophers. The Greeks
themselves used the names "science" and "philosophy" synonymously, a
convention that was still observed at the time ofKepler and Newton. Even
as late as the nineteenth century Claude Bernard spoke about the
characteristics of "experimental philosophy." This is hardly news,
however, so let us come to the point: if for centuries "science" and
"philosophy" were synonymous terms, how at a later date did they come
to stand for different disciplines? What set them apart? More
specifically, what was the principal reason for their separation? Though
often discussed, this question seems worthy of further pursuit; for it
appears to us that a precise, formal understanding of what brought about
the divorce between science and philosophy is necessary if one is to
determine their separate functions as systematic disciplines and see why
a philosophy of nature is necessary for understanding the physical world,.
C. Science and philosophy: their distinction
The kind of answer often given to the question, "What distinguishes

science from philosophy?" is well known and can be accurately stated in
a single line:
... the modern quantitative and descriptive approach to terrestrial motion

is very different from the qualitative approach of Aristotle.'
The quotation expresses the substance of the modern view, according to

which the introduction of quantitative procedures into the study of motion
initiated the separation of science from philosophy; and in the minds of
many that left philosophy without any claim to speak seriously in a
systematic way about the world of nature. To be sure, the question is not
fully answered by the lines above, nonetheless they present a substantive
part of the position of those who would maintain that the systematic
investigation of the physical world is solely the business of the natural
The Mathematical Character ofPhysic! 147
sciences.
We must grant at once that philosophy is not quantitative; it does not
measure, and its arguments are not mathematical. But lest one think that
the measurements and the mathematics which are employed by physics
and chemistry are the solely the consequence of the genius and a new
"insight" of men such as Galileo, Kepler, and Newton-lest one think that
is the case I wish to argue that the use of mathematics is per se to the
science of physics and can be accounted for very simply by considering
the cause oflocal motion, change of position as the physicist says. But we
must explain ourselves.
Opinions on what is called the "Copernican revolution" have, as
everyone knows, been many, and they are summarily represented as to
type by Alexander Koyre in a passage it is useful to quote:
This revolution has been described and explained-much more explained

than described-in quite a number of ways. Some people stress the role of
experience and experiment in the new science, the fight against bookish
learning, the new belief of modem man in himself, in his ability to
discover truth by his own powers, by exercising his senses and his
intelligence, so forcefully expressed by Bacon and Descartes, in
contradistinction to the formerly prevailing belief in the supreme and
overwhelming value of tradition and consecrated authority.
Some others stress the practical attitude of modem man, who runs away
from the vita contemplativa, in which the medieval and antique mind
allegedly saw the very acme of human life, to the vita activa: who
therefore is no longer able to content himself with pure speculation and
theory; and who wants a knowledge that can be put to use: a scientia
activa, operativa, as Bacon called it, or, as Descartes has said, a science
that would make man master and possessor of nature.
The new science, we are told sometimes, is the science of the craftsman
and the engineer, of the working, enterprising, and calculating tradesman,
in fact, the science of the rising bourgeois classes of modem society. 2
Considering them commonly, such accounts of the scientific

revolution are accidental in the sense that they regard the theoretical mind
as being shaped by intellectual and affective attitudes that are extrinsic to
the content of the issues themselves. If one divides the accounts into
types, he can say that some of them assume practical advantages to be the
determining historical factors initiating the scientific revolution, implying
thereby that the contemplative function of the mind either was not
considered a real option, or that whatever there is of the
contemplative-the theoretical understanding of nature for its own sake
-in human heads is subordinated to practical benefits. To be sure, one can
hardly deny that the technology accompanying the scientific revolution
accelerated the development of the modem commercial era and gave men
an extensive dominion over nature; but the question is: did the
development of an interest in trade and manufacturing bring about the
quantification and mathematization of the study of nature? Now such
interests may well have disposed scholars of the times to look at the world
mathematically, yet we shall argue that it is not the fundamental reason
why physics became mathematical..
But Koyre offers another explanation, one that he endorses himself:
I am convinced that the rise and growth of experimental science is not the
source but, on the contrary, the result of the new theoretical, that is, the
new metaphysical approach to nature that forms the content of the
scientific revolution of the seventeenth century, a content which we have
to understand before we can attempt an explanation (whatever this may
be) of its historical occurrence.
I shaH therefore characterize this revolution by two closely connected and

even complementary features: (a) the destruction of the cosmos, and
therefore the disappearance from science-at least in principle, if not
always in fact-of ail considerations based on this concept, and (b) the
geometrization of space, that is the substitution of the homogeneous and
abstract-however now considered as real--<limension space of the
Euclidean geometry for the concrete and differentiated place-continuum
of pre-Galilean physics and astronomy.
As a matter of fact, this characterization is very nearly equivalent to the

mathematization (geometrization) of science. 3
In substance Koyre does maintain that the scientific revolution came

about as the result of a new metaphysics, a metaphysics that regarded the
universe as fundamentally mathematical in nature. One way to make
mathematics the science of nature is, as we mentioned above, to regard
reality as equivalent to extension, to magnitude, if you will, and perhaps
also to look upon number as somehow the basic constitution of natural
entities. Koyre argues that with modem science Pythagoras has returned,
and henceforth God is to be viewed as a mathematician. The temptation
The Mathematical Character of Physics 149
to want to understand the world through mathematical principles is easily

understood, for mathematics is the science most fully developed and the
one which the human mind most fully masters. But we shall leave that
aside and note only that an important consequence of the Newtonian
revolution is that it tied the mind down to measurements obliging it
thereby to be empirical, a state of affairs from which philosophy had
divorced itself and as a result had lost its anchors in reality. And so in
contrast to modem philosophy, we shall attempt to show that the origin
of the mathematical character of Newtonian physics has its roots in the
observable world of nature instead of a new "metaphysics." To make our
point, however, we must first look at the theory of motion which the new
physics overthrew and which had prevailed for many centuries.
D. Elements of the Aristotelian theory of the cosmos
Aristotle's theory of the cosmos was founded first on a theory of four

physical elements: fire, air, water, and earth-rock, or mineral as we
would say. Each of the four elements was thought to have a natural
motion by which it tended to an up or down location with respect to the
others. Fire tended toward the extremity of the universe, moving in the
direction we call up, and so was considered to be absolutely light. Earth,
or mineral, tended toward the center of the universe, moving in the
direction we call down, and so was absolutely heavy. Air and water were
in between. Air was heavy in relation to fire but light in relation to both
water and earth and so occupied a position between them. Water was
heavy in relation to air and fire but light in relation to earth; so it occupied
a corresponding position between air and earth. The motions that
characterized these elements were held to follow on the constitutions of
elements themselves, much as the motion of a photon is held to follow on
its constitution, or as diffusion follows on the constitution of gas, except
that the motion of heavy and light bodies, like the diffusion of a gas,
could be impeded, which is not the case with the motion of a photon. In
short, the fundamental motions-toward the extremity of the universe and
toward its center-were specifically, that is, qualitatively, distinct; the
differences were not merely of degree and were absolute insofar as the
universe was considered to have an absolute boundary and an absolute
center. Finally we must add that their natural motions, as we said, were
held to be present as soon as the elements came into existence; such
motions did not require the immediate application of an agent to initiate
them, just as a photon, once it exists, does not require the action of an
extrinsic agent to initiate its motion.
Turning now to the heavens, in addition to the earth, there were the
celestial bodies which were held to be radically distinct in kind from
bodies on the earth. The celestial bodies were thought to be incorruptible
and everlasting and on that account not constituted of the same sort of
material as things of the earth. Aristotle considered such a supposition to
be necessary because according to the astronomical records available to
him no one had observed either a star or planet come into existence or
pass out of existence, which made reasonable the hypotheses of a different
kind of material constitution and the consequent incorruptibility of
celestial bodies. 4
But with the advent of the new data that grounded a new astronomy,
that of Copernicus, Galileo, Kepler, and Newton, the understanding of the
cosmos underwent a very radical change. First in importance was the
implication that celestial bodies are not composed of a material radically
different from the substances found on the earth. Once the face of the
moon was seen (with the aid of the telescope) to contain many irregular-
ities, once the earth was understood not to be the center of the universe,
once the planets were held to revolve about one star among millions, then
it became more reasonable to assume that terrestrial and celestial bodies
are composed of the same kind of materials. From then on celestial bodies
could not be thought incapable ofbeing produced or destroyed by natural
agencies, however infrequently such events might happen. And if the
celestial bodies are made of the same kind of materials as terrestrial
bodies, then there is no reason to think of them as being moved by
agencies different from those that move bodies on the earth. Furthermore,
if the earth revolves about the sun, then the elements do not move toward
the center of the universe or toward its boundary by reason ofbeing heavy
or light. In short, once the dike had been pierced-to use a cliche-the
trickling water soon became a flood, sweeping before it the entire edifice
of Aristotelian science. Furthermore, space now came to be considered an
existing container, the extent of which was infinite, or at the least
unknown, so that there could not be a question oflocating moving bodies
in relation to the boundaries of space, of the universe. Such conditions
clearly required a new understanding of the cause of local motion.
The Mathematical Character ofPhysics
151
E. The new causal theory
There seems little doubt that the most serious weakness in the
Aristotelian theory of local motion was the account it gave of projectiles,
which are instances of what Aristotle called constrained motion, a motion
he thought could be continued only by reason of a complicated movement
of the air through which it traveled. From early times this part of his
theory met opposition, but not until the medieval period did its chief
opponent, the impetus theory, become accepted. The important advantage
of the impetus theory was that it did not require a continuously acting
agent to sustain the constrained motion; instead it held that contact
between an agent and a moved body was necessary only at the beginning
of the motion. Thomas Aquinas stated the relevant principle as follows:
It is necessary that mover and moved exist simultaneously with respect to

the beginning of the motion but not with respect to the entire motion, as
is evident in projectiles. 5
Such a principle requires motion to be regarded not as transient but as

enduring, like shape in a hard body, unless some extrinsic cause acts to
oppose it, which means, to repeat the point, that motion does not need the
continued action of an agent to keep it in existence. And let us note that
the cause in question is an active cause, a mover. Thus on those grounds
one seems justified in saying that the notion of impetus contributed
importantly to preparing the way for Newton's theory of mechanics,
which becomes more evident when we look at the first law of motion, a
law which contains in itself the entire justification for what follows.
Newton formulated his frrst law as follows: "Every body continues in
its state of rest, or of uniform motion in a right line, unless it is compelled
to change that state by forces impressed upon it." In one way this physical
principle is very clear in what it says: neither motion nor rest follows
necessarily on the constitution of a body. A body in a state of rest will
remain in a state of rest unless an extrinsic force affects it, and a body that
is moving will continue to move in a straight line at a constant speed
unless an extrinsic force acts upon it. In short, the first law denies that
motion according to place follows on the constitution ofphysical bodies,
and it therefore specifically denies what had been central to Aristotle's
theory of the elements. (Not until photons came on the scene did we
return to elements upon which translational motion followed in the mode
of a necessary property.)
Physicists are quick to point out that an absence of extrinsic forces can
be understood in two ways: either no forces whatsoever are acting on a
body, or the forces acting on it are balanced, which means that their
vector sum is zero. Hence the first law also says that no unbalanced force
is affecting the body, which means that the only effect for which we can
detect an agent cause-the unbalanced force-is an acceleration, which
by defmition is a change of speed or direction. Now no determinate
direction in relation to the universe or to space as a whole can be assigned
to a motion; so to repeat: according to Newton the only observable effect
for which a cause can be assigned is an acceleration. Thus unlike
Aristotle he could not attempt to describe a particular motion by referring
to the extremes of the universe.
In the absence of known limits of the universe an acceleration can be
established only by reference to a set of coordinates, and whether the
change be one of speed or one of direction, acceleration is a variation in
a motion that is a more or less. That differences in speed are differences
of more or less is obvious, for speed is an indication of how fast or slowly
a body is traveling. But in the new physics a change of direction is also to
be regarded as a difference of more or less; for when directions can no
longer be defined absolutely in relation to parts of the universe they must
be described as angles measured from coordinates (that are not established
by the boundaries of the universe), which means that direction too
becomes a difference of more or less. That is to say, directions are distin-
guished quantitatively rather than qualitatively. Even movements that are
oppositely directed and represented as positive and negative come to be
regarded as having no absolute qualitative differentiation or direction in
relation to the universe. In short, Newton's first law tells us that the only
difference in the motion of a body that can be causally explained is a
difference ofmore or less. On that account a science of moving bodies can
no longer be founded on the natural, qualitatively distinct kinds of local
motion that characterized Aristotelian physics.
And so now we see that if they are to be known precisely, differences
of degree must be measured. Consequently the new physics was obliged
to become quantitative, for according to its fundamental philosophical
position only variations in speed or direction, only differences of degree
had causes that could be detected. And once the official data of physics
became numbers standing for measurements, then all correlations of data
had to be represented by mathematical propositions; for every relation
The Mathematical Character ofPhysics 153
between numbers is a mathematical relation. This amounts to saying that

Newton's view on the physical character and cause of motion made
mathematics as necessary for physics as measurements. Hence the
mathematization of physics would not seem to be the result of a new, a
priori metaphysical point of view, it would not seem to require reality to
be fundamentally geometrical in character, nor would it seem to be the
result of a growing interest in and disposition for technological advances.
On the contrary, it was the direct result of a new theory of local motion
that is fully compatible with an Aristotelian position on the nature of
physical reality, with an Aristotelian "metaphysics," so to speak.
But making physics mathematical does not mean that motion is a
mathematical entity. Motion remains a physical phenomenon unchanged
in its nature; for not every difference of degree, of intensity, of more or
less, is, unlike the continuum, a difference that is intrinsically mathe-
matical in nature. To be sure, as soon as the official data or starting points
become measurements, then every observable that is immediately relevant
to the explanatory process of the science is correlated insofar as it is
measurable with a number representing a measurement. And because
ordinary observations of differences of degree such as fast and slow, high
and low, right and left, hot and cold, heavy and light, hard and soft, etc.,
are too imprecise for scientific considerations, differences of degree in
qualities or characteristics other than magnitude (extension, dimension)
or multitude are correlated with that sensible trait in which differences of
degree are most clearly and deterrninately recognized, namely length. A
thermometer, for example, is a device that measures how hot or cold
something is, but the thermometer does not tell us directly about that
which we sense when we say something is hot. Instead, a thermometer
allows us to correlate the quality hot with a column of mercury, which in
tum is correlated with graduations on a scale (itself a length) placed
alongside the column of mercury. (Meters that measure in degrees of arc
do the same sort of thing.) Neither of these, however, is what we sense
when we feel something is hot. Nevertheless, numbers are the official data
of physics when it deals with heat phenomena because, as was said,
numbers are directly related to something more easily observed than
differences in the intensity of heat. Briefly, what the scientist directly
observes when he measures temperature is not the hot but the juxta-
position of two lengths. 6 1
Similarly, although motion is measured by length, motion is not
length. To use a simple example, in measuring velocity one measures the
distance (a length) over which a body travels and then he compares it to

the arc (a length) measured by the hands of a clock. That which is directly
measured is not that which the physical consideration seeks to explain;
rather, the lengths are used as means by which the mind can come to an
understanding of the (minimal) reality of motion. And so it would seem
that the conclusion to be drawn is that physics became a mathematical
science because (I) new observations led to the view that the celestial
bodies are made of stuffs similar to those on earth and therefore one
should not seek the causes of their motions outside the physical universe
itself; because (2) it seeks causal accounts of differences of more or less
in motion; because (3) differences of more or less in movements can be
accurately known only by correlating them with differences in something
which can be directly measured and more accurately observed than the
differences they help us to know. In short, the prompting reasons and the
confirming data are all physical in character.
The account that has just been given of mechanics flts some other parts
of physics. Electrical forces produce accelerations and in that respect are
treated in the way that mechanical forces are: no acceleration, no
detectable causal agency. Thermodynamics considers heat phenomena,
which, as its second law reveals, are dependent on differences of more or
less in relation to temperature, and only in the presence of such
differences can work be done. We must also mention that chemistry did
not become mathematical in character until much later, precisely because
the discrete character of reactions did not become apparent until much
experimental work had been done. But a law such as the law of defmite
proportions demands a theory that accounts for reactions through parts
that function as units; hence proportions and their mathematical
representations became formal to the science of chemistry.
And so to restate our point: physics became mathematical, not because
of social circumstances or changes in metaphysical viewpoints but
because empirical demands affected its theory of physical motion. And
chemistry became mathematical only much later precisely because it took
more time to discover the discrete character of the combining amounts of
elements in reactions, which alone could justify an atomic theory and its
consequent mathematics. Here, too, the mathematization was the result of
a change in the causal theory of "mixed" substances, and like physics,
chemistry became quantitative and mathematical because of new data
together with the kind of causal theory the data permitted. But what, we
may now ask, does this mathematization of physics imply about

The Mathematical Character ofPhysics 155
philosophy? That is the next question.
Notes
1. The quotation we give is from Gerald Holton and Duane Roller, Foundations
of Modern Physical Science (Reading, Massachusetts: Addison-Wesley
Publishing Company, Inc., 1958, p. 22. The quotation is, we think, representative.
2. Newtonian Studies (Cambridge, Mass., 1965), pp. 5B6.
3. Ibid.
4. Aristotle was well aware that his account of celestial movement was founded
on an hypothesis, for he says: "The mere evidence of the senses is enough to
convince us of this, at least with human certainty. For in the whole range oftime
past, so far as or inherited records reach, no change appears to have taken place
either in the whole scheme of the outermost heaven or in any of its parts." (De
Caelo et Mundo, Bk. I, chap. 3, 270b, 11 16.) Thomas Aquinas comments on this
passage as follows: "This is not necessary but probable. To the extent that
something is long-lasting, more time is required in order that this change be
detected; just as the change in a man will not be detected throughout a two or
three year span during which the change in a dog or some other animal of shorter
life will be observed. Therefore, someone can say that although he heaven is
naturally corruptible, it is so long-lasting that the total time span of which we have
memory does not suffice for detecting its change. (In I de Caelo, Iect. 7, n. 6.) The
translation is ours.
5. Questiones Disputatae de Potentia, (Rome: Marietti, 1949), q. 3, a. 11, ad 5.
6. Plainly an analogue device such as a mercury thermometer is closer to
experience in this regard than a digital counter. One might even ask if the latter
would be intelligible if it were not known to do the sort of thing that a
thermometer does.
Chapter 13
General Considerations on Philosophy
A. Introduction: the separation of philosophy from science1
The history of natural science and the history of philosophy, both of

which begin with Thales ofMiletus, show that originally philosophy and
science were not distinct entetprises; their histories are the same up to the
late medieval period. They are no longer, however; and in our day they
are taken to be separate entetprises. The separation has been completed
for some time, but what does it show philosophy to be? If we ask the
question "What is philosophy?" can we, on the basis of ordinary
experience, give an answer that tells the listener how to identify the
discipline and the sort of thing it does? As everyone knows who has tried
to answer that question for the uninitiated, the task is not easy, and many
are familiar with the word only because some football coach talks about
his "philosophy." One reason for this state of affairs is that the parts of
philosophy do not at first sight appear to have much in common; they
seem to be too disparate to belong to a common category, which is not
true of either the natural or the social sciences. We readily understand that
chemistry, biology, and physics all bear on natural entities and so are
called "natural sciences." Similarly, when we hear history, economics,
sociology, etc., described as social sciences, we know that they all treat
the social actions of men under one aspect or another. But epistemology,
philosophy of mind, ethics, metaphysics, aesthetics, logic, et alia, all of
them parts of philosophy, do not have the same effect. The ordinary
listener cannot identify some of the parts, much less a common category
according to which all these "subjects" are classified. An added obscurity
arises from the philosophical disciplines not being utilitarian; that is,
certain fundamental philosophical disciplines are not applied disciplines,
as are ethics to human actions and logic to the reflective activities of the
mind. Furthermore, the great variety of philosophical opinion contributes
in a substantial way to the layman's confusion; and instinctively he tends

to shun philosophy when he discovers it has little unity of thought.
Everyone understands that on some topics natural scientists do not have
one mind, and everyone also understands that the social sciences are the
scene ofgreater disagreement. Nevertheless the layman does not deny that
the sciences, even those in which the disagreements are greatest, have
legitimate areas of investigation, unlike philosophy, which is unique in its
disparate character. There is little the ordinary man can see that tells him
philosophers have a common enterprise; and when he learns about some
of the disagreements, he turns away, perhaps saying that philosophy is
beyond him and then making an act of faith that philosophers know what
they are doing. Are we not to expect, then, that philosophy will appear
irrelevant to the important activities of human life?
But the problems go beyond those that appear to the ordinary man.
Recently Richard Rorty has claimed that throughout history philosophers'
attempts to solve problems in a systematic way have failed. The attempts
are, he says, mere pretensions; and philosophers cannot claim to do
anything more than "improve human conversation." The philosopher's
role, he tells us, is that of an intellectual gadfly; it is to stimulate
discussion of one's own and other cultures to see whether they are
consistent, whether they "hang together." 2
Though not so striking, evidence of another sort is available to indicate
a malaise in philosophy. Rather commonly, logicians separate explanation
from argument, claiming that the two are not the same, though they
generally concede that explanation employs argumentative patterns. Now
such a separation signals a tendency to think of explanation as proper to
the sciences and to view the function of philosophy as something
different. Yet even when philosophy is allowed an explanatory role, its
kind of explanation is so described as to have little in common with
explanation in the sciences. 3
Also, some at one time held that philosophy and logic-he latter being
viewed as a calculus-are identical in their concern; both focus on the
formal structures of the mind. That such a view could get serious attention
tells us by itself that philosophy was not seen as having the same sort of
general relationship to the world of realities as the sciences do. Logic was
taken to be like mathematics, not bearing on realities and not knowing
whether or not what it said was true. Such a divorce from reality is
thoroughgoing, and though philosophical practice indicates that the view
is hardly common, nonetheless it tells us something about the trouble that
General Considerations on Philosophy 159
philosophers have in understanding their role in the life of the human

intelligence. Moreover, empirical propositions, except those of ethics, still
seem to be regarded as the property of the sciences, a state of affairs that
requires the divorce of philosophy from the physical world.
Now our aim in what follows is to argue that philosophy is not rightly
conceived if it is taken to have a different domain from the scientific
disciplines; and the points we shall make are four: (I) philosophy is
founded on empirical regularities; (2) its explanations are in important
ways similar to those of the sciences (except it does not use mathematics);
(3) the nature of argument is taken from explanation; (4) explanation is
a paradigm that precludes logic from being adequately conceived if it is
thought to be a calculus. And let us add that the common experience and
data upon which philosophy depends is presupposed to everything the
sciences do. As a simple example, to come to know atoms and molecules,
the stuff that is divided must first be known as the whole which is being
divided, and as such it is known through ordinary observations and
experience.
B. Philosophy and regularities
Earlier we were at pains to show that systematic considerations start

from generalizations which are laws or other universally stated
propositions derived from experience and founded on regularities in
behavior or properties. In systematic disciplines, experience provides the
mind's contact with realities; and if the intelligence is to be objectively
grounded, if it is to account for the world 11 out there, 11 then no exceptions
for the regulating role of experience can be made for any discipline that
is acquired solely through natural endeavors. Thus, if philosophy has a
legitimate role to play in the mind's coming to understand the real world,
then it too must start from regularities. Otherwise were we to deny
experience its role in philosophy we would have no choice but to say that
philosophical knowledge is a priori, that philosophy is derived either
directly from the mind itself or directly from God, which no one really
believes. In short, philosophy is like other disciplines and must start from
and be measured by the regularities that observation and experience make
known. No other alternative is open to it.
Physicists, chemists, and biologists begin their scientific activities by
investigating real entities, seeking to solve problems which regularities
bring to light. Their investigations lead to explanations; and after their
explanations are proposed, attention can be directed to the activity of

explaining itself, which is a very reflective business. Stated another way,
scientists first do physics or chemistry, etc., and then later someone
reflects on what they have done. Similarly, philosophers who start with
regularities investigate realities directly, and only later can they tum to
thinking about the philosophical activity itself. Therein, however, lies the
difficulty; for that has not been the philosophical practice. Richard Rorty,
of whom we spoke earlier, describes philosophical activity as it has been
conceived in modem times when he says:
Philosophers usually think of their discipline as one which discusses

perennial, eternal problem-problems which arise as soon as one reflects.
Some of these concern the difference between human beings and other beings,
and are crystallized in questions concerning the relation between the mind and
the body. Other problems concern the legitimation of claims to know, and are
crystallized in questions concerning the "foundations" of knowledge. To
discover these foundations is to discover something about the mind, and
conversely. Philosophy as a discipline thus sees itself as the attempt to
underwrite or debunk claims to knowledge made by science, morality, art, or
religion. It purports to do this on the basis of its special understanding of the
nature of knowledge and of mind. Philosophy can be foundational in respect
to the rest of culture because culture is the assemblage of claims to
knowledge, and philosophy adjudicates such claims. It can do so because it
understands the foundations of knowledge, and it finds these foundations in
a study of man-as-knower, of the "mental process" or the "activity of
representation" which makes knowledge possible. To know is to represent
accurately what is outside the mind; so to understand the possibility and
nature of knowledge is to understand the way in which the mind is able to
construct such representations. Philosophy's central concern is to be a general
theory of representation, a theory which will divide culture up into the areas
which represent reality well, those which represent it less well, and those
which do not represent it at all (despite their pretense of doing so). 4
According to Rorty, modem philosophy has been concerned with a

"general theory of representation," that is with epistemological and
psychological problems, with a theory of knowledge, and as a result of
that concern philosophy has turned from the world "out there" to the
"world within." Later he adds another point worth noting:
Besides raising "the science of man" from an empirical to an a priori level,

Kant did three other things which helped philosophy-as-epistemology to
become self-conscious and self-confident. First, by identifying the central
issue of epistemology as the relation between two equally real but

irreducibly distinct sorts ofrepresentations-"formal" ones (concepts) and
"material" ones (intuitions)-he made it possible to see important
continuities between the new epistemological problematic and problems
(the problems of reason and of universals) which had bothered the
ancients and the medievals. He thereby made it possible to write "histories
of philosophy" of the modern sort. 5
Rorty, it would appear, is correct. When philosophy abandons realities as

starting points and turns to the activities of the mind, it begins to construct
a philosophical "system" in the modem sense. When it attempts to begin
with reflective activities instead of those that are directed outward, it
becomes a priori, idealistic, constructivistic, and has lost its leading
strings. In thinking directly about the world and then afterwards reflecting
on what has been thought, the mind follows the procedure demanded by
its very character. Mental activities resemble physical activities insofar as
they cannot be understood apart from their relations to their objects. As
we said earlier, heating as an action is not intelligible apart from the rise
in temperature it produces; cutting is not intelligible apart from the
separation of parts for which it is responsible; and we could add to the
examples. In the same way, understanding or knowing cannot properly be
grasped without reference to objects; that is to say, without some
understanding of realities outside the mind, epistemology is impossible.
So it seems that omitting empirical regularities as starting points has led
over the years to a variety of "theories" that are in part, at least,
responsible for the current malaise and loss of identity that philosophy
seems to be experiencing. 6 On the other hand, if philosophy starts from
empirical regularities, then it will not think of itself as occupying a
domain distinct from that of the other sciences; rather, it will understand
that it occupies some part of the same domain. And if philosophy has a
"foundational" as well as a critical role it can only be because the
regularities of ordinary experience and the observations on which they
depend are fundamental to all the disciplines. But another word on
regularities.
Regularities depend on the fact that natural entities-every one of
them-belong each to its own species, and this is especially true of
organisms. When we name a moose, a robin, an eagle, etc., we name the
kind of organism each of them is. And when we defme, we defme that
which is common, the nature, the species, and not the individuals. When
we say that man is an animal capable of inferring conclusions from
premisses, we are talking about the kind of thing man is; we are talking
about a nature that is found in every individual. If it is not, then the
individual is not of the human species. Our point is that we cannot do
without these species-names, for we have no systematic understanding of
the specimens or individual members without knowing first of all the
nature that is common to them all; so we must have words that signify the
species. Furthermore, that the intelligence can understand realities
according either to their existence in the world "out there" or according
to the existence they have within the intelligence-their intentional
existence-is well understood by philosophers, if not so well understood
by many others. Now the difference in the kind of existence that can
accompany what a word signifies can bring about difficulties, which a
certain use of the word "man" illustrates. Consider the statement "Man
evolved from a lesser primate." If one now wishes to use the statement as
a premiss and say that "John is a man," and then say therefore "John
evolved from a lesser primate," he draws an utterly false conclusion, for
we may not say "John evolved from a lesser primate." On the contrary, he
was born ofhuman parents. Hence neither "person" or "human being" can
be substituted for "man," since they will render some propositions false.
For instance, no Christian may say that Christ was a "person like us in all
things but sin," because they all understand him to be a divine, not a
human person, and so not like us in that way. Nor as we just said can we
substitute "human being" for "man" because "human being" too signifies
a person. 7 In sum, man is a species and the word "person" cannot be
substituted for "man" whenever the species as such is being discussed.
C. Hypotheses and explanation in the sciences-some illustrations
We must now look at hypotheses for a second time for the sake of
seeing better how they function as premisses in explanations so that we
may contrast them to premisses that are explanatory in philosophy. Earlier
we noted that atomic theory provides hypotheses as principles of
explanations, as explanatory premisses, which means that the laws
obtained by induction from which the problems arose are now seen as
deductive conclusions of arguments. The investigators knew that in its
essentials the explanation would read: chemical reactants combine in
definite proportions (in multiple proportions, conserve their mass)
because p. In other words, they already had conclusions; what they
needed was a premiss. Dalton's assumption that chemical reactants are
composed of atoms provided the reason why reactants combine in definite

proportions as well as why they combine in multiple proportions, and why
they conserve their mass. To repeat: the ability to combine that elements
had shown in their empirically established reactions was explained by
assuming that a certain unobservable characteristic belongs to the
reactants, namely, their being composed of atoms. Expressed in terms that
describe the propositions of the argument, an observationally established
predicate was said to belong to the subject in the law-proposition because
another unobservable predicate was assumed to belong to the same
subject in the premiss-proposition.
But that was not all; there was also the need to defme atom, without
which the explanation was no explanation. Indeed, the defmition was the
heart of the assumption; the descriptive characteristics of atoms provided
the essential content of the atomic theory. Hence the argument went:
chemical reactants are composed of atoms, therefore chemical reactants
combine in defmite proportions. Implicitly contained in the explanation
was the set of defming characteristics of atom, one or all of the
characteristics substituting for a single term in a given argument.
Another illustration of a scientific explanation is provided by the
account geologists give of the (observable) cleavage manifested by some
minerals. Cleavage, they tell us, is the propensity some minerals have to
break under moderate stress along a plane. To account for this behavior,
geologists assumed that the crystals of the minerals had weak atomic
bonds and that the breakage occurred along the line of those weak bonds.
So here, too, we have an argument that says p implies q, that is, "Some
minerals tend to cleave (q) because some minerals have weak atomic
bonds" (p); and once again we have two predicates assigned to the same
subject, one predicate telling us about the observable behavior, the other
predicate telling us about the assumed unobservable cause of that
behavior.
Konrad Lorenz explained the herding of small animals by saying it is
a protective device which makes attack more difficult for predators. When
small animals herd, predators fmd it difficult to single out one individual.8
Stated in its essentials, we once again have an argument in which one
predicate assigned to a subject explains another: small animals herd
together because small animals are better protected by that action.
Furthermore the premiss itself is accounted for by an argument: small
animals are better protected by herding because small animals are harder
for a predator to single out when they gather in herds. The last premiss
depends on an induction from experience, together with the intelligible

relation between subject and predicate. The point, however, is that the
same three-term pattern is present here that we saw above.
Nor do the social sciences differ in the general nature of their
explanations. In one of the chapters of Democracy in America, 9 Alexis de
Tocqueville asks why Americans are restless in the midst of their
prosperity; and he replies by saying they are restless because they have an
excessive desire for physical gratification. Tocqueville, we must say, is
very complete in his account, and he gives a number of other arguments
that are related to the one we quoted; but his basic explanation is the one
we gave above. So here, too, the reader can see that we have the same
three-term pattern.
D. Syllogistic arguments
Because the illustrations given above represent what is done in both

the natural and the social sciences, they allow us to say that explanation
involves two predicates said of one subject (or if the conclusion is
negative, one predicate denied of a subject on the ground that another
belongs to it); and so the explanation is a syllogism. By defmition, three
terms constitute a syllogism when two are affirmed and/or denied in
combination with a third. The central point here requires emphasis, for
explanations are that at which all other investigative activities ultimately
aim. And because explanations are arguments constituted by three terms,
they are syllogistic in character. Therefore when we account for a
property, we first look for its cause within the thing or stuff to which the
property belongs, 10 but if we fail to find the cause within, we then look for
an extrinsic cause to account for what the thing or stuff has undergone or
had happen to it. For instance, if we were to say "Uranus has an aberration
in its orbit because it is affected by a gravitational field centered at
coordinates x, y, z," we would assign a property-effect to Uranus by
reason of an extrinsic cause. In sum, then, we account for a characteristic
property through something else that we either know or hypothesize to be
present in the thing or stuff or to affect it, thereby generating a syllogism.
The syllogistic form of the arguments above is readily seen when
ordinary expansion procedures are used to provide the missing (major)
premisses:
Every active substance composed of atoms combines in definite proportions;

every reactant is an active substance composed of atoms; therefore

every reactant combines in definite proportions.
The same procedure can be used with all the other examples; for instance
the argument on herding goes as follows:
All animals that are better protected from predators by herding do herd;
small animals are better protected from predators by herding; therefore
small animals herd.
An obviously unarticulated general premiss presupposed to all such

arguments in biology is the statement that animals (as a rule) do that
which is advantageous (better, adaptive) for the preservation of the
species. Hence the argument above tells us animals herd because herding
is advantageous for them, which is another three-term argument: whatever
is advantageous for small animals is done by them; herding is ad-
vantageous for small animals; therefore herding is done by them. This
expression of the explanation is less euphonious than Lorenz's actual
words, but it is plainly syllogistic and exhibits the connections the mind
actually makes.
(Let us pause for a moment to note that arguments are rarely stated as
we have stated them above. Usually they are adorned with a rhetorical
language that makes them easier to read but harder to ferret out. In other
words, finding an argument in modem prose can often be very difficult
because of its burial in language. Needless to say, this state of affairs is
much more characteristic of philosophy and the social sciences than it is
of the natural sciences.)
And so we may repeat our point: wanting to know why some natural
species behaves as it does, why it has the properties it has, is the terminal
question of our investigations. First we must know that something is so,
and we must also know what some thing, property, behavior, or
behavioral process is; but after that we want to know why two things
occur or exist together, which means that we seek a cause-effect relation
as the principal scientific pursuit. Nor do we exaggerate when we say that
being able to answer the question why is most of all (but not exclusively)
what distinguishes human intelligence from the "intelligences" of other
animals. Explanation is argument par excellence, and it is the standard to
which all other kinds of argument are compared. Furthermore, since
explanations are syllogistic, the syllogism is the mind's main instrument
for showing causal relations. On the other hand, because it is only an

instrument, whenever the causal relations are obvious, we do not bother
to put the argument in syllogistic form.
Some have said that we do not argue syllogistically in the natural
sciences. But that claim will not stand the light of day. One reason for the
claim is that we tend to relate atoms directly to macroscopic properties,
forgetting that atoms are by defmition parts of substances and must be
understood as such. In other words, atoms (and molecules, etc.) enter into
arguments through propositions which say that natural substances are
composed of atoms, or that natural substances are composed of molecules,
or that atoms are composed of particles, etc. The relation of the
composing part to the composed whole must not be forgotten, for when
it is, the argument plainly cannot be written in three terms.
E. Explanation in philosophy
It is now time to discuss whether philosophy explains, and if so,

whether its explanations are like those of the sciences. As an answer to
both we shall present an illustration that depends on empirical regularities,
an illustration that, because it is philosophical and deals with
unobservables, requires a number of paragraphs to become clear.
When we observe individual animals of various species, we see that
they behave according to patterns that are uniform from one individual to
another. Except in regard to incidentals, beavers in Minnesota behave like
one another and like those in Ontario, as well as those in Maine.
Similarly, crows are alike everywhere, as are black bears, otters, mud
turtles, northern pike, and every other species we might name. It is true
that within their behavior patterns the members of more advanced species
show greater variability and complexity than the members of less
advanced species; yet observation shows that the basic kinds of actions of
each class are uniformly the same. Thus what we may call the
"occupational behavior" of animals is uniform in kind and in exercise and
is governed by a pattern for the whole of what a species does. By
"uniform in exercise" we mean that unless they are sick animals do not
shirk their species-behavior. Such, then, is the regularity that observation
reveals and from which we infer that animals have unobservable inborn,
instinctive patterns of behavior, and this general inference founds the
biological science of ethology. Briefly: animal behavior is regular,
therefore animal behavior is instinctive; that is the (unadorned) inference.
General Considerations on Philosoph~ 167
Man, however, stands in contrast to other animals. External

observation does not reveal in him a corresponding uniformity of
occupational behavior. On the contrary, it reveals variations in kind, not
only from one individual to another, but from one society to another and
from one historical period to another. Experience also shows that human
occupations have to be learned, sometimes over very extensive periods.
From these regularities we easily (almost unconsciously) infer that men
do not possess a set of inborn behavioral patterns that is sufficient to
govern the whole of human occupational behavior. Of course there are
some instincts in men; but by themselves they are not able to provide an
occupational behavior sufficient to attain the good life, as are the instincts
of animals in relation to what is the "good life" for them. Nor do we need
to say that just as the occupational behavior of other animals includes the
production and raising of young, so, too, the occupational behavior of
men includes the begetting and rearing of children. These activities are
occupational in the sense in which the term is used here, and the same can
be said of activities of amusement. ("Occupation" in its vocational or
professional sense is but a part of the whole of human occupation in the
broader sense of the term in the present context.) In short, observation
shows that man's occupational behavior is varied and must be acquired,
a state of affairs that sets him apart from other species. Furthermore, man
can shun his activities without being sick, and in this, too, he differs from
other animals. To summarize: man's occupational behavior is varied;
therefore man's occupational behavior is not instinctive. Another
argument follows: man's behavior is not determined by instinct; therefore
man's behavior is self-determined. Plainly this disjunctive argument
supposes that one or the other of the two options belongs to animals.
If we now ask what word is appropriate to signify this indetermination
in the human species, this lack of congenitally determined behavioral
patterns, this self-determination ofbehavior, we may reply by saying that
man is appropriately said to exhibit "free" behavior and is appropriately
said to be a "free agent." By defmition, then, a free agent is one who lacks
inborn patterns that determine both the kind of behavior and its exercise
and who therefore must direct his behavior himself. Put another way, a
free agent is one that is self-determining both in regard to the kind of
occupational behavior he performs, and in regard to the exercise of his
behavior.
We wish to repeat that observation-experience-justifies this
defmition, which no further elaborations can contradict. To be sure, we
do have internal experience of being able to choose to do one kind of

thing rather than another and of being able to put off doing anything at all;
but external experience is more fundamental than internal experience,
especially when we go to define unobservables. In other words, what we
observe externally we confirm through our awareness of selecting the
kind and of commanding the performance of our actions. A clear
advantage of such an empirical procedure is that not only does one "get
things straight," but as a consequence he avoids many needless
discussions; for observation in no way permits us to say that a free agent
is one whose actions are uncaused, or that it is one whose actions are
totally independent of external human influences as well as the limitations
of the environment. etc.
Having argued first that man is free, we must at this point ask why
man's actions are free, why his behavior is varied. Of course we cannot
reply by saying that he is free because he lacks an inborn set ofbehavioral
patterns, for that is to do no more than describe the state of affairs the
name signifies. Such a reply merely gives the defmition which determines
the signification of the word "free." Rather, the sense of the question is:
why is man able to behave in a variety of ways and to determine for
himself (within the limits of environmental circumstances and
opportunities) the kind of occupation he will acquire and exercise? The
answer, of course, is not hard to provide; for we all realize that man acts
as he does because he has an intelligence that is able to understand a
universal domain ofbehavior. Starting from external observation, we may
infer that since man actually performs many kinds of activities, it follows
that he possesses an internal principle or cause of activity universal in its
nature that we readily recognize as intelligence. Stated directly, man is a
free agent because he is an intellectual agent. 11 (Activities involving
operational capacities other than the intelligence are free only in the
measure that they are able to be directed by the intelligence.)
But there is another empirical regularity that goes along with those we
have already seen, one that tells us men are civil or political animals.
Experience shows that neither individuals nor families nor tribes are
adequate for providing the many physical and cultural goods that
contribute to a full human life. In other words, if we ask why is man a
political animal, we reply by saying because he cannot supply either by
himself (or through a lesser society) what is needed for a good human life.
Then if we ask why he cannot supply those goods by himself, we must
reply because his intelligence allows him to conceive a quasi-infinity of
--------------------
General Considerations on Philosop~ 169
goods which he cannot attain except in a political community. Thus, man

is political because he is intellectual; and we now have explained two of
his properties through intelligence, employing in the process two
syllogistic arguments.
At this point we may put our two propositions about man together and
note that human freedom is fundamentally related to man's social nature.
Stated another way, in order to attain a good life, individual men must
contribute to society through their occupational behavior, which means
that freedom is at root an innate indetermination which obliges us to
decide how we shall direct our activities so as to contribute to the good of
the whole, within the limits that our circumstances allow. In other words,
freedom is fundamentally a causal liberty that is directed to the social,
common good. Freedom is first of all a liberty to produce, not to obtain;
and certainly it is not primarily related to the pursuit of pleasure. In
outline, then, that is an instance of what we think is a philosophical
explanation; and though it starts from commonplaces that are a part of
everyone's experience, the conclusions to be drawn from these obvious
starting points are, as the reader will know, at variance with actual
political practice and custom.
We must now note that our illustration shows a three-term explanation:
man is free because he is intelligent (rational), u and man is political
because he is rational. Two terms are said of a third in each syllogism,
making the explanations the same in kind as those of the sciences. We
also see through this example that philosophy seeks causal accounts. Like
the sciences, it attempts to hunt down the cause-effect relation between a
thing and its properties, or between two properties. Furthermore our
examples of philosophical explanations started from empirical
regularities, not from hypotheses. That is not to say philosophers never
employ hypotheses, but it is to say that hypotheses are not the ordinary
starting points for philosophical explanations, and certainly not for the
fundamental issues. At the risk of being repetitious, philosophy
characteristically argues and explains by relating one empirical regularity
to another; that is its proper procedure. So in a sense (part of) philosophy
is more empirical than the natural sciences. But metaphysics, because it
bears on unobservable realities, is not so directly related to experience.
Though they must ultimately be grounded on observable regularities, the
problems and explanations of metaphysics arise from and depend on
propositions that are inferred from observation and so are inherently more
obscure to the mind. Any discipline that pretends to bear on the
unobservable per se as its proper subject matter must be related to

observation through other realities that are the source of its intellectual
light. As Camap rightly said, the absence of such a relation is responsible
for nonsense.
But now the question naturally arises as to how philosophy differs
from the sciences, for in view of the claim that it starts from and
investigates regularities, one might well wonder. For instance, are the
regularities the same or different from those of other disciplines? If the
same, then are there differences in the way the regularities are used? How,
for instance, does a non-mathematical part of biology-ethology,
say-differ from philosophy? Or psychology? The question requires an
answer, and it will be considered in another chapter.
Returning to our earlier point, it would seem that there is good reason
why philosophy should be experiencing a malaise. Rather than search out
natural regularities, philosophers tend to construct-in the words of a
colleague-"mechanisms" which they seek to impose on reality to render
it intelligible, thereby becoming a priori and rationalistic.
E. Logic and language
Although it is a bit of a digression, we shall now take up the last of the

four points we listed at the end of section A above, namely, that logic is
not adequately conceived if it is thought to be principally a calculus, and
the acceptance of a symbolic apparatus to replace analyses done from
language indicates the separation of the contemporary philosophic mind
from its empirical starting points. Neither explanations as the paradigm
kind of argument nor the differences in kinds of evidence can be seen
through symbols. 13 In general, modem logic attempts to consider
arguments and their propositional components without taking into account
what at one time were called "semantic links." Basic logical relations are
defmed using propositional variables that are universal in their
signification, which means that the variables can stand for any proposition
whatsoever, thereby supposedly making the propositional relations purely
formal. A consequence of this point of view is that compound
propositions as well as inferences are defmed in terms of truth tables,
which are taken to establish the fundamental logical relations. Consider,
for instance, the statements made in one of the more popular and
respected texts in symbolic logic:
General Considerations on Philosophy, 171
It should be obvious that MPP [modus ponendo ponens] is a reliable

principle of reasoning. It can never lead us, at least, from true premisses
to a false conclusion. For it is a basic feature of our use of 'if... then ... '
that if a conditional is true and if also its antecedent is true then its
consequent must be true too, and MPP precisely allows us to affirm as a
conclusion the consequent of a conditional, given as premisses the
conditional itself and its antecedent. 14
Lemmon is very clear in what he says. The fundamental affirmative

mode of argumentation constitutes a "reliable principle of reasoning"; it
is taken as primitive, as unexplainable through anything prior, a point
that is conf"mned by Lemmon's subsequent description of the inferential
relation in mpp as a "basic feature." Thus, the inferential relation is
defmed through a truth table, and no more fundamental principle of such
a table is admitted. But this is precisely what requires to be challenged.
When implication is defmed through a truth table to mean that p
implies q is true if and only if p cannot be true and q false, we may then
ask the question, why is this so? Why does a valid argument allow modus
ponens and modus tollens but nothing else? It is not enough to say that
they are so by defmition, that we (stipulatively) defme validity that way;
for it is easily seen that the issue is not evident on its face. 15 But answering
the question requires us to return to the syllogism, for it provides the
origin of the rules of modus ponens and modus tollens.
When we take the most obvious case, Every B is A and Every C is B,
therefore Every C is A, we see without difficulty that if the premisses are
true the conclusion must also be true. We see that certain things are said
from which something else necessarily follows. Now given the necessary
relation, it is not possible for the premisses to be true and the conclusion
false. But though we readily see that such is the case, we still have not
fully answered why the premisses cannot be false and the conclusion true;
and in reply we must refer to the relations of predicability. If A is
predicated of all B (if it is predicated universally), and ifB is predicated
of all C, then C may be predicated of all A. In short, the relations of
predicability are what necessitate the conclusion's being true whenever the
premisses are true. If one considers A, B, and C extensionally, he has a
similar state of affairs; for if A is a class (collection, etc.) that includes B,
and ifB is a class that includes C, then necessarily A includes C. In both
cases the necessity is seen as soon as the nature of the relationship of the
terms to one another is understood; and here is where the analysis ofthe
inferential relation stops. We may add that an extensional consideration
of classes cannot be prior to or more primitive than an intensional

consideration, for only on the basis of "intensions" do we do know why
the members of a natural class are gathered together; the nature, the
characteristic (the intension) is that by which the members are identified.
Furthermore, without the intensional considerations we cannot see why
two or more classes are related.
The formal relations among terms are only the beginning. After we see
that one class can be said of another, we then turn to examine the
character of the identifying class attribute or characteristic to see how one
accounts for another. Only through that sort of activity can we actually
examine the kind of evidence being offered; only then do we see that we
have a cause-effect relation, or a relation of effect to cause, or perhaps a
sign relation. Hence, we may repeat that the formal relations on which
predicability is founded are but instruments for attaining realities and
ultimately observables. If we limit our logical concerns to formal
relations, not only do we fail to see why our concepts are connected, we
also empty them of their contact with reality. We even fail to see their
character as instruments; and so we make them into an abstract system
studied for itself, which of course is consistent with the view that logic
and mathematics are continuous. But all this may be summarized in one
statement: a syllogism is an ordered application of a principle to
instances and nothing more; a principle, of course, that directly or
indirectly is derived from observation. The first figure, whether
affirmative or negative, is the obvious, direct application of a principle.
The second figure applies only a negative principle, while figure three is
a more obscure and less direct application of either an affirmative or
negative principle.
When we say that philosophy investigates regularities, we must also
say that philosophy's principles are the most important of its regularities,
for the principles contain all that is accounted for by them. And since the
principles are obtained from sensible, observable properties, we can see
that observation is the ultimate source of all that we know. From
principles obtained from observables we come to know everything else,
either because of a cause-effect relation or because of a likeness, a
similarity-relation. We might also note here that everything we have said
here about the properties of the mind-its use of signs, symbols, the
importance of etymologies, the nature of imitative art, the nature of
language-all of this has depended on three principles: (1) the defmition
of an observable as that which determines the character of a sensation; (2)

the statement that unobservables become known through a modification

of a sensible trait; and (3) the mind's use of similarities to manifest the
nature of unobservables. Everything we have said has been developed
from those there principles; and one can readily see how the modification
of a sensible trait tips the mind off to the entire business of accounting for
unobservables. More generally, everything philosophy rightly establishes
must be traceable to the very first observations we make and upon which
we impose the names that supply our starting points.
The last point to be discussed has to do with relevancy of the
explaining terms in the premisses to the explained term in the conclusion.
As described by Aristotle, a sophistical argument is a kind of error that
deceives not the untrained but the educated mind. It deceives because its
error is difficult to detect; and one of the most fundamental reasons for
these more than ordinary obscurities is to be found in the subtle shifts of
meanings that occur in words that signify realities, which, though distinct,
are very similar or closely related to one another. Symbols, however, do
not allow the mind to see sophistries because symbols abstract from the
meanings of words and propositions and as a result from relations of
relevancy. Consequently if we are actually to fmd the causal relations in
things, we must first argue linguistically; otherwise we shall fail in our
purpose.
F. A summary
In the foregoing we have tried to show that philosophy's malaise, its

uneasiness about its proper role and identity, are the consequence of its
separation from reality which it brought about in order to occupy itself
with the domain of the intelligence and problems ofknowledge. Our case
has been constructed around several points. First, philosophy must be
founded on regularities; second, explanation in philosophy does not differ
in its basic nature from explanation in the sciences; third, the nature of
argument is taken from explanation, which is a search for a cause and is
syllogistic in character; fourth, logic is not adequately conceived if it is
held to be primarily a calculus because symbols cannot reveal the
relevant, causal relations through analysis. Logical instruments are
ultimately for the sake of attaining causal relations in regularities, and
they cannot be separated from the use oflanguage, which alone allows us
to signify the "qualitative" distinctions in things. Furthermore, symbolic
systems do not focus on analysis for the sake of seeing the cause-effect
relation; instead they construct systems the validity of which is known

through the substitution of truth values for the components, an activity
that is consistent with the view that the proper concern of philosophy is
a theory of knowledge. On the other hand, if what we have advocated in
this chapter is accurate, then philosophy cannot be conceived as belonging
to a domain distinct from that of the sciences, and its "foundational" role
has to be differently understood.
Notes
1. We shall speak here of natural philosophy.

2. Philosophy and the Mirror ofNature. We recommend that the reader see also
the "Introduction," Consequences ofPragmatism (Minneapolis, Minnesota: The
University of Minnesota Press, 1983). Though we have represented Rorty's
position in perhaps an overly simplified way, our description nonetheless carries
enough of what he says for the reader to see in the pages that follow that our view
of philosophy is radically different from his, as well as from the views of many
others.
3. See Robert Nozick, Philosophical Explanations (Cambridge, Mass.: Harvard
University Press, 1981).
4 .Philosophy and the Mirror of Nature (Minneapolis: University of Minnesota
Press, 1983), p. 3.
5. Ibid., p. 138.
6. We have not made mention of the further developments philosophy underwent,
namely, from seeing itself as concerned with a theory of knowledge to seeing
itself as properly considering purely formal systems, purely formal intellectual
structures, and from there to formal structures oflanguage. These developments
do naught but illustrate further philosophy's separation from empirical regularities
as providing the first grist for its mill.
7. This equivalence is given by Webster's II and III, The Oxford English
Dictionary, and The American Heritage Dictionary ofthe English Language. We
cannot, then, rightly make that substitution. A word that is sometimes used
because of its ambiguity is the pronoun "one," which when it is applied to us
signifies a person. "Person" in its very first meaning signifies an individual of the
species man.
8. On Aggression (New York: Harcourt Brace Jovanovich, Inc., 1966), p. 136.
9. (New York: Random House. Inc., 1954), Vol. II, Bk. II, c. 13.
10. Syllogisms appear deceptively simple-minded when we state them formally,
partly because a reader coming on them set out in that neat fashion often does not
actually comprehend the character of the causal relation which they signify. He
really understands only the verbal expressions, whereas a scientist who grasps a
causal premiss sees its application in many instances, every one of which involves
a syllogistic inference.
11. Plainly our explanation is a set of bare bones. Questions having to do with
pressures on the exercise of liberty that come from emotions, etc., are left aside;
they are not necessary for the purposes we have in mind here, namely, the
illustrating of philosophical explanation, since they bear on the application ofthe
notion to human acts, not on the justification of the notion itself.
12. It goes without saying that the character of the intelligence must be known in
detail to appreciate the efficacy of the explanation fully. A mere statement ofthe
explanatory premiss cannot do that sort of job.
13. An extensive comparison of calculational, truth functional logic of modern
times to the kind of logic first worked out by Aristotle is made by Henry B.
Veatch, Two Logics, (Evanston, Illinois: Northwestern University Press, 1969).
Veatch considers our present issue, namely the unsuitability of relational or
symbolic logic for philosophy, in a thorough and competent way. We can touch
only on a central point or two, whereas he takes up all the major topics associated
with the character oflogic. We recommend the reader see his work for his more
thorough treatment.
14. E. J. Lemmon, Beginning Logic (Indianapolis: Hackett Publishing Company,
1978), p. 11.
15. As my colleague James Stromberg notes, this way of defining a valid
argument begs the question, for one cannot presuppose the truth of the conclusion
because that has to be established by a formal syllogistic pattern.
Chapter 14
Philosophical Principles: Preliminaries
A. Introduction
When we compare the adult to the infant mind, we see that the
intelligence, like the body, undergoes development over the course of
time, our descriptions of the mind employ the language of the body.
When we say, for example, that this or that student is developing well,
that he is improving his physics or mathematics, that he is strong in
chemistry, we extend the language applied to observable entities to the
unobservable intelligence. And just as the physical is subject to certain
general developmental principles, so too the intelligence, which moves us
to ask the question we now must consider, namely, what are the general
principles that determine the course of the mind's progress, its
development from a state of ignorance to a mature grasp of a systematic
discipline or disciplines? Let us note at once that the issue we shall
discuss does not belong to the "psychology of learning," as that phrase is
commonly understood. We shall have nothing to do with emotional
preconditions, "motivation," ambient circumstances, etc. Instead, our
concern is with the procedures that are innate to the intelligence and that
are not subject to variation by human manipulation. Stated another way,
our concern is with the mode of knowing that is built into the mind and
necessarily manifests itself in every functioning human being,
independently of his cultural or environmental circumstances.
B. The mind's movement from the sensible
Because we know the sensible properties of physical objects before we

know unobservable properties, causes, and substances, and because we
must first know exterior observables before we know internal activities,
it follows that one way the intelligence develops is by moving from
concepts bearing on observables to concepts bearing on unobservables.

The basic orientation to physical objects which characterizes the mind's
mode of knowing is responsible for the medievals having said that the
proper object of the intelligence is sensible reality, which obviously does
not imply that the mind cannot know realities beyond the observable. The
import of that proposition is simply to say that a commensurate proportion
exists between sensible objects and the human intelligence- they fit one
another. So once again, the orientation of the mind to observables
determines the first way in which the intelligence develops, namely, by
moving from understanding observable realities to understanding those
that are unobservable and ordinarily of superior ontological status.
C. From the confused to the distinct
Not only is the mind innately oriented to moving from observables to

unobservables, it also is innately endowed with a determinate way of
developing in its acquisition of systematic knowledge. When we start
from a state of ignorance to investigate cats or sharks or automobiles or
houses, etc., we begin by apprehending such objects as wholes. We first
grasp objects as wholes, but we do so in an indeterminate way. We do not
see or understand distinctly the nature and role of the parts of which the
wholes are constituted. For instance, we daily encounter squirrels and blue
jays and other animals, as well as rose bushes and maple trees and other
plants, not to mention the universe taken as one thing, and all of these are
wholes having parts. Yet how many of us can describe accurately and in
detail the parts that constitute the organisms we mentioned? How much
do we know about limbs, muscle systems, skeletons, roots, leaves, etc.?
If, however, our understanding of these organisms is to develop, we shall
have to come to know the nature of the organisms' parts more clearly.
The point of our remarks is uncomplicated: our first understanding of
the objects we investigate is a confused, indistinct (that is the sense of
"confused" here), indeterminate knowledge of them as wholes, after
which we move on to examine the parts with the aim of possessing a more
distinct knowledge of the whole itself. To be sure, the parts can be known
according to varying levels of distinctness, as a result of which the wholes
too become known according to varying levels of distinctness. It would
seem, then, that the principle is evident: the mind develops by moving
from a confused to a distinct knowledge of wholes by coming to know the
parts of the whole more distinctly.
Philosophical Principles: Preliminaries 179
There is still another way in which the mind develops by going from the
confused to the distinct, which we shall discuss by making use once again
ofPiaget's interrogations of children. In Chapter V, he describes interviews
which show that at an early age children ordinarily identify conscious
activity with activity generally:
From the results obtained, four groups may reasonably be distinguished,

corresponding grosso modo to four successive stages. For children of the
first stage, everything that is in any way active is conscious, even if it be
stationary. In the second stage consciousness is only attributed to things
that can move. The sun and a bicycle are conscious, a table and a stone are
not. During the third stage an essential distinction is made between
movement that is due to the object itself and movement that is introduced
by an outside agent. Bodies that can move of their own accord, like the sun,
the wind, etc., are henceforth alone held to be conscious, while objects that
receive their movement from without, like bicycles, etc., are devoid of
consciousness. Finally, in the fourth stage, consciousness is restricted to the
animal world. 1
Piaget interrogated the children by asking them whether certain objects

felt or knew or experienced pain, etc., which of course implies
consciousness. The replies he received indicated that in the earliest stage
the children could not distinguish conscious activities from those that are
not conscious; and only with maturity did they come to recognize that a
certain distinction is necessary: some actions are conscious, others are not.
What this means is that at first children take "knowing" and "feeling" and
"hurting," etc., as synonyms for movement in general: no separate
signification for such names exists in their minds. Let me repeat: children
first understand realities according to their most general traits, and only
later do they make distinctions that require the general class to be separated
into subdivisions.
The same confusion shows up in the child's knowledge of other
characteristics, and Chapter VI ofPiaget's book describes the way children
regard the living and the non-living:
During the first stage everything is regarded as living which has activity or
a function or a use of any sort. During the second stage, life is defined by
movement, all movement being regarded as in a certain degree
spontaneous. During the third stage, the child distinguishes spontaneous
movement from movement imposed by an outside agent and life is
identified with the former. Finally, in the fourth stage, life is restricted
either to animals or to animals and plants. 2
The child's failure to grasp the distinction between living and non-living is
quite different from the attribution of life to the inanimate that an adult
might make for the purpose of defending a panpsychist position, a
confusion that is an error rather than an imperfection. We all understand
imperfection to be part of the normal developing process native to the
intelligence, but the adult who erroneously assigns to an entity an attribute
which the mind has come to understand with some distinction is not in a
state of imperfection but of error, a quite different state of affairs. Error is
a confusion of quite another sort, and as we said, pansychism is an
illustration of it.
Although the data in the illustrations above are new, the principle is not,
for Aristotle illustrated the developmental character of the mind when he
pointed out that little children call all men "father" (or all women
"mother"). His remark does not mean that children do not recognize their
parents on a sensory level, for they obviously do, and at an early age.
Rather, his statement tells about the child's intellectual apprehension. A
neighbor's boy once asked me, inquiring about my wife, where my mother
was, which showed that to him "mother" was a synonym for "adult
woman." He had not yet come to understand that his mother is the
particular woman who was a cause of his existence.
There are other ways, too, in which this sort of development shows
itself. If, for instance, a hunter looks across an expanse of prairie to the foot
of a hill and sees something on the slope, he will know that it is a
four-legged animal, though he may not know whether it is an antelope, an
elk, a deer, or something else. The hunter may be certain about the animal's
generic character yet be ignorant of its specific kind. Similarly, a physicist
may see an unexpected track in his cloud chamber and not know what
caused it, except that he is certain the track was caused by something real,
and here, too, the statements he makes are limited to generic attributes. He
is unable to assign those that are specific, which shows the character of this
sort of development, namely: we can know the general with certainty
without knowing the specific, and that is to know in an indeterminate,
confused way. Let us repeat: the mind develops by going from the general,
which is indeterminate or confused, to the specific, which is more
determinate or distinct. (Unfortunately the bureaucratic mind too often
goes in the other direction.)
It is, of course, characteristic of the generic that it is "empty" relative to
the specific. In order to gather beavers, chipmunks, earthworms, etc. into

the one category animal, we must leave out of our concept of animal every
distinguishing difference that separates one species from the others. The
generic includes only the common, only that which every species exhibits,
and it leaves out whatever is unique, peculiar, or distinguishing in each of
them. As a consequence, when we know a species only according to its
generic attributes, we know the species indeterminately or confusedly,
insofar as our concept does not make us capable of distinguishing one kind
from another. 3
The process of defming also reveals the sort of "acquisitional" priority
which the general has over the particular. If we defme a triangle as a
three-sided figure, we do not understand the defmition fully until we know
what a figure is; and if we define animal as a sentient organism, we do not
understand the defmition if we do not know what an organism is. Of course
one might have some sort of understanding of the more specific without
understanding the generic, but to the extent that is so, to the same extent his
knowledge is confused or indeterminate. In short, we can know the generic
determinately without knowing the specific, but we cannot know the
specific deterrninately without knowing the generic. And to know the
species not in its specific character but only according to what is generic
in it, is to know the specific confusedly. And that is the state of the children
Piaget interrogated.
But not only do our individual minds develop, those who come after us
obviously take advantage of what we have learned, and they can build on
it. That is to say, the sciences themselves come to a fulfillment and
flowering as the consequence of the maturing that goes on in individual
heads. Over the centuries, the "collective intelligence" responsible for what
we know about nature, for example, has itself developed, a state of affairs
which is reflected in the phrase, "the state of the art." Not surprisingly,
then, the general progress of the sciences is an obvious consequence of the
developmental principles which govern the individual mind in overcoming
its initial state of ignorance.
D. A summary
Our earlier chapters have attempted to show some of the important

implications of the mind's dependence on observables. In this chapter we
have talked about the way the mind develops by going from the more
general to the specific. More determinately, the mind develops by moving
from the confused to the distinct, and that kind of development occurs in
two ways: first, when we go from knowing wholes in a confused manner
to knowing their parts and then return to the wholes with a more distinct
understanding; and second, when we proceed from considering the general
to the more specific by way of distinguishing the species or subdivisions
under the genus or more universal class. We see that these modes of
development are natural to the mind, for they are ubiquitous and
unavoidable: they stem from the constitution of the intelligence. Yet we
must not forget that we are talking about systematic or scientific
knowledge.
E. Imperfection and error: an emphasis
It is important to emphasize that whenever we know something

confusedly or indeterminately, we are not on that account in error. Even
color blindness, which many take to be a state of deception, is only a
failure to distinguish, and it cannot rightly be considered a permanent state
of error. We do not take animals that see only black and white to be
constantly in a state of error; nor do we consider ourselves to be seeing
false images when we look at black and white movies or telecasts. Plainly,
then, the failure to sense certain distinctions is a state of imperfection or
incompleteness rather than a state of error.
The same point can be made about knowing the parts of natural entities
indeterminately, and about knowing causes or other entities generally and
not specifically. We are all in imperfect states of knowledge about many
things, but we are not on that account wrong about them. To take an
extreme example, God knows what we know about rabbits and squirrels,
but He knows them exhaustively. Can his superior knowledge rightly be
taken to mean that we live in illusion? That there may be more to the world
than our thoughts conceive or our language expresses does not of itself
imply that what we conceive or express is false or that an anti-realist
position is the only one tenable. So with these points now in mind, we shall
go on to say something about the way in which we are obliged to define.
Following that, we shall consider the character of philosophical first
principles
Notes
I. Op. cit., p. 173. This time we shall use only Piaget's summary remarks and not
the interviews themselves.
2. Op. cit., p. 194.
3. This more empty character of the general is of course why Hegel had to claim
to have discovered a "new" universal, the "concrete" universal, which supposedly
contained the differences within it. Without such a claim he could not have so
audaciously pretended to deduce the species from the genus.
Chapter 15
Natural Philosophy
A. The issue narrowed 1
Socrates questions Gorgias in the dialogue bearing the latter's name

with a view to discovering just what rhetoric is. He attempts to get
Gorgias to name the subject matter that belongs to rhetoric on the grounds
that the different arts are to be distinguished by their different subject
matters. But Gorgias is not able to do so; hence the nature of rhetoric is
left in obscurity. If now, on the basis of Socrates' principle, we are asked
to do the same and indicate a subject matter peculiar to natural
philosophy, we shall have to admit that we cannot, at least we cannot if
we are expected to name one that is sui generis, one that is generically
distinct from the subject matters of the natural sciences. Yet we must add
at once that the absence of such a generic subject matter does not preclude
natural philosophy's being a separate discipline that has something
worthwhile to say about the world of nature.
But to make clear the character of natural philosophy, we need to
make some preliminary remarks about why physics measures and so is
mathematical, because doing so will help to make clear just what sort of
a discipline natural philosophy is.
B. Preliminaries
Taking advantage of what we have said about the mathematical

character of physics as a systematic natural science, we must note that as
soon as one understands that physics measures, he is able to see its
inherent limitation. Although accepting numbers-pointer readings, as
Eddington said-as its official data is one of its strengths by obliging the
mind to be empirical, such data also impose a limitation. When numbers
are applied to physical entities, representing differences of more or less,

they presuppose a qualitative homogeneity on the part of those realities
that differ according to a more or less. For instance, motions that are
faster or slower are both changes of location; temperatures, whether
higher or lower, both pertain to the quality by which we say that
something is hot; two men, one of whom is more rational than the other
- to pick a familiar example-have the same kind of cognitive capacity
and so are considered according to that homogeneity. In short, differences
of degree are not differences of kind, and that is our point.
That quantitative differences presuppose qualitative homogeneity is
obviously true when one relates two or more instances of characteristics
such as distance, temperature, voltage, resistance, light intensity, volume,
etc. But, we may ask, do quantitative differences presuppose qualitative
homogeneity on the part of physical reality when we relate temperature
to volume, pressure to temperature and to volume, etc.? Temperature,
pressure, and volume are qualitatively-that is, formally-different
properties, two of which are causally related to the other: differences in
temperature and pressure effect differences in volume (through the stuffs
particles, or course). Temperature and pressure are also causally related
to each other when volume remains constant. But the causal relations that
exist among these properties (and the particles of the stuff) is the reason
why variations of degree in one or the other of them occasion variations
of degree in the other or others. Thus, the recognition of such differences
is a qualitative affair and is presupposed to an understanding of the
numerical correlations that exist among the properties according to their
variations. 2 For instance, that one property is recognized as causing and
the other is recognized as affected by the cause is not itself a distinction
of more or less; and when one asks whether the measurement of two
events is simultaneous or whether the time interval as measured from two
different coordinate systems is the same or different, he does not need a
technical definition of time insofar as it is seen to be a necessary property
of motion.
Although physicists assign numbers to motion (velocity), to time, to
distance, and to other realities, measuring some and calculating values for
others, they do not engage in reflective discussions on questions such as:
what is motion? what is time? what is location (place)? Yet such questions
can be examined systematically in a way that does not trespass on the
domain appropriate to physics. One can define time as a general notion on
the basis of ordinary experience without having to be expert on the use
Natural Philosophy 187
physics makes of the notion and without being an expert on the subtleties
of attempting time's measurement; and of course to measure time we must
have at least a general understanding of what time is. We must add, too,
that other qualitative distinctions pertaining to common natural properties
are presupposed to the investigations of the natural sciences, distinctions
that need not be understood by such sciences according to anything more
than our common conceptions.
Perhaps it is even more obvious that considerations on what motion is
or what change in general is in no way affects mechanics or any other part
of physics, or any part of the other natural sciences, for that matter.
Furthermore, if a scientist makes a general distinction between categories
of change on the ground that some are chemical and some are physical,
then in the process of distinguishing them he will have to make use of the
notion of substance, since a chemical change involves a change of the
latter. But just what is involved when a modification affects a stuff rather
than one of its modifying properties is a matter that requires systematic
examination; and here, too, it may be said that the consideration in no way
affects the hypotheses that the chemist may have formed about the
processes by which the changes in stuff are thought to be accomplished.
In sum, the sorts of questions described above are the sort of thing natural
philosophy can and ought to do, and they separate it from the
experimental sciences themselves.
To continue the point, biology is a natural science that deals with
living entities; yet the biologist does not systematically examine, for
example, whether organisms are aggregates of particles, what an
aggregate is, what internal states as distinct from external operations are,
etc. These issues do get discussed, not by the biologist but by
philosophers.
And so having made these preliminary remarks, we may now ask, the
question, "What is natural philosophy"?
C. The character of natural philosophyl
As everyone knows, parts of some natural sciences are in large

measure non-mathematical, for example, ethology, the science of
instinctive animal behavior. In the main ethology is concerned with
explaining qualitative characteristics of animal behavior. Its fundamental
data are not numerical, nor are its principal considerations mathematical. 4
Can one imagine trying to explain instinctive animal movements by trying
to make the primary focus of his labors differences of more or less? But
then why, if the distinction of the natural sciences from philosophy is
found principally in the use of measurement, is ethology not a part of
philosophy? An ethologist must, to be sure, make more observations than
a philosopher, but can one look at that as an essential characteristic?
We must concede at once that no strong case can be made for putting
ethology on one side of the fence rather than the other if we depend on
such grounds. If ethology starts from observations, if it argues
systematically, defines, classifies, etc., by taking into account similarities
and differences that are not reducible to measurements, then little
substantive ground can be found for distinguishing it from natural
philosophy, especially if the latter considers all "mental" activities of
organisms. Quite the contrary, on the basis of an absence of measurement
a strong argument could be made for saying that ethology is an instance
in which the names "science" and "philosophy" can be applied to the same
discipline. But such an identification is unlikely to meet with favor from
ethologists, owing to impressions they may have as to what "philosophy"
does in practice, and owing also to the insistence on keeping philosophy
and biology distinct in accordance with current academic custom; and
assuming that such a longstanding custom cannot be wholly in error, the
objection is understandable. Moreover, the distinction between them must
in some measure be accommodated according to the points that have been
already made; for the considerations of the ethologist are more particular,
while those of the philosopher are more general. That is central to the
distinction between natural science and natural philosophy.
In the area of sociology and anthropology Lewis Mumford regards
himself as a generalist, and the name suggests how he views his function.
It indicates the kind of considerations Mumford makes on the detailed
investigations of his colleagues. He compares the particulars, the details,
to discover the general regularities that might belong to many distinct
societies. Similarly, one might describe the function of the natural
philosopher in that way: he is a generalist insofar as he considers the
causes and properties common to all natural things; and the distinctions
he makes are evident on the basis of ordinary experience, however much
he may also avail himself of the data and regularities uncovered by the
natural sciences.
It bears emphasis that generalizing supposes data and information that
are sufficient to supply an adequate inductive base for the generalizations.
Thus, like the generalist in sociology, the natural philosopher needs to be
reasonably acquainted with the data and the more important observational
regularities the scientist comes to know, although the natural philosopher
need not be a scientist nor have a scientist's understanding of his subject
mater. In our time, acquiring knowledge of natural regularities is for most
people a vicarious business. We live in artificial environments that
preclude our observing the behavior patterns and traits of many things in
nature, and from this point of view the American Indian was considerably
our superior. It is easily seen, then, that the natural sciences provide the
philosopher with a vicarious observational contact with reality, a not
insignificant service. But we must consider the point more determinately.
The general principle according to which one orders the parts of a
science which have the same category of subject matter is revealed in the
following quotation:
... in a certain way the sciences follow on things; for [operational] habits
are distinguished according to their objects, from which the habits derive
their species. But the natural scientist considers motion and mobile things.
The Philosopher [Aristotle] says in the Physica, Bk. II, that anything
which moves belongs to the investigation of some physical science.
Therefore it is necessary to distinguish the parts of the science of nature
according to differences in motions and mobile entities. Now the first kind
of motion is that which is local [change of position]; it is more perfect
than the other kinds and is common to all natural bodies .... Hence after
the consideration of motion and moveable entities that is common and that
was taken up in the Physica, [Aristotle] first considers bodies according
as they are moved by local motion, and he does this in the De Caelo,
which is the second part of natural science. There remains, then, the
consideration of other motions which follow, those motions that are not
common to all bodies .... Among these, production and destruction
[generatio et corruptio-what we today call chemical change] have the first
place. Alteration [what we now call "physical change"] is ordered to
production as to a goal. Growth also is consequent upon production and
destruction [chemical change], for growth does not occur without certain
productions and destructions by means of which nutritional materials are
converted into that which is nourished by them; for as the Philosopher
says in the De Anima, Bk II, just as food nourishes insofar as it is
potentially flesh, so, too, it increases magnitude as it is potentially
quantified flesh. Thus it is necessary that these motions be considered at
the same time because in a certain way they follow on production and
destruction. 5
The principle according to which the sciences are to be ordered for the
purposes of systematic considerations6 is brought out in the text quoted

above; and it is: when one is engaged in systematic considerations he
ought first to consider that which is common before he moves on to that
which is less common, lest, as Aristotle often remarks, he be obliged to
repeat many times that which belongs to the genera of the realities under
consideration. (Of course one cannot follow an ideal procedure when he
first instructs beginners.) On that basis, then, because movement
according to place is common to all natural entities, one first considers
this kind of motion according to those principles which are themselves
common to all things. In modem physics that is the task of mechanics; it
investigates the local motion of bodies according to the principles-mass
and gravity-by means of which such spontaneous, natural movements are
brought about. 7 Following such investigations, one must consider
chemical changes and the alterations, the physical changes, which are
necessary for them, namely, changes of temperature and pressure.
Although reproduction as a biological activity is a kind of process that
brings about another substance, it is not a part of chemistry; for
reproduction is sui generis and not common to everything that can
undergo chemical changes. Thus, reproduction gives rise to the science of
biology.
In addition to the order of the natural sciences according to the relative
community of the motions formal to each, we must also recognize that the
natural movements have .a causal order among themselves. Chemical
reactions are necessary for biological operations and the production of
organisms; physical changes, especially changes of temperature and
pressure, are necessary for chemical reactions; changes of temperature
and pressure in nature are the consequence of solar radiation, which in
tum varies at a given location on the surface of the earth according to the
relative motion of the earth in its orbit, together with the rotation of the
earth on its axis. Considered according to this serial relation, then, local
motion of the earth in regard to the sun is causal in relation to physical
changes, which in tum are causal in relation to chemical reactions, which
again in their tum are causal in a certain way in relation to reproduction
and other biological operations. Hence the order of the natural sciences
according to the community of the kinds of motion is also an order
according to which one kind of motion is causal in relation to another.
In summary we may now say that natural philosophy is related to the
natural sciences according to a principle that is responsible for the order
of them all in relation to each other. Natural philosophy comes first in a
systematic arrangement because it considers that which is common to all

mobile entities. It considers the principles of natural entities together with
their properties. But to avoid confusion, let us note that natural philosophy
when it considers motion looks at what is common in an analogous what
to every kind of change, every kind of motion, whether it be a change of
location, change of quality, etc.
D. Natural philosophy and philosophy of science
The reader is certainly aware that some of the issues that we have
claimed to be the province of natural philosophy do get discussed under
the title of philosophy of science. When one "explicates the concepts of
the natural sciences" he can hardly avoid touching some of the topics that
belong to natural philosophy. Nonetheless, these two parts of philosophy
ought not to be regarded as equivalent because questions of scientific
methodology, which form a large part of the philosophy of science, do not
fit into the same category as substantive questions that bear directly on
natural entities and their properties. Nor do the epistemological topics,
those that bear formally on the kind of knowledge each is, belong to
natural philosophy. In short, the philosophy of science has its own
character and its own role to play, and one ought not to confuse it with
natural philosophy.
In closing, I would like to note that although the foregoing discussion
is but an outline, it does indicate what, in the minds of many, natural
philosophy ought to do. It would seem evident that a complete
understanding of nature requires, to the extent possible, a systematic
examination of the general character and attributes of natural entities
insofar as they are qualitatively distinct and insofar as their differences are
not reducible to differences of degree (a state of affairs that precludes
philosophy disappearing into science). The latter task, I have argued,
belongs to philosophy, even though one cannot claim that its contem-
porary mainstream succeeds in or even attempts to accomplish that task.
But the actual state of affairs matters little for the theoretical issue, and the
work is there to be done.
E. Epilogue
Earlier we made the point that philosophy starts from and is concerned
with regularities, as are the sciences of nature. But before we leave the
issue of where philosophy stands in the spectrum of disciplines that

investigate nature, we must note that some regularities pertain to less than
the species. For instance, a part of the human race is light-skinned, while
other parts are darker skinned. Furthermore, anthropologists ask
themselves how such differences in pigmentation is adaptive, and they
offer explanations based on environment. People with darker skin adapt
better to tropical and semi-tropical climates, they say, than people who
have lighter complexions and who, in contrast, are better adapted to the
more northern regions of the earth, as for instance the countries of
Norway and Sweden. But our aim is not to try to be an amateur
anthropologist but only to indicate that some regularities pertain to
properties and behavior that are less than species-wide; and similar
regularities are to be found in parts of animal biology, when, for instance,
one is concerned with varieties of organisms. Philosophy, on the other
hand, when it investigates nature, does not concern itself with traits or
behavior that is less universal than the entire class, category, or species
that it might be considering. The difference is important, for it must be
taken into account when one considers not only the place of philosophy
in regard to the natural sciences but also when the kind of explanations
philosophy employs are at issue. The kind of argument Aristotle called
"demonstration through proper causes" does not have to do with proper-
ties or behavior that belong only to subdivisions of species. On that
account philosophical explanation will ground its arguments on causes of
the genera or species as such, not on causes that are outside the genera or
species and so affect only individuals possessed of traits that are not
rooted in the nature as such.
Notes
1. The position presented here regarding the relation of natural philosophy to the
natural sciences, except for the shortcomings that are mine, is substantially that
of the late Charles DeKoninck, of Laval University in the city of Quebec.
2. An ordinary understanding derived from our common and indeterminate notion
of these properties founded on common experience suffices for such purposes.
3. What we say here about why physics measures and how natural philosophy is
to be distinguished from the natural sciences is treated at length in our essay, "The
Character ofNatural Philosophy," in The New Scholasticism, Vol. LI, 3, Summer,
1977.
4. Ethologists do measure some phenomena to find out, for example, if one mode
ofbehavior is more efficient than another, and such differences of more or less do
occasion some measurements. But these do not tell against our point.
5. Thomas Aquinas, Proemium to his commentary on Aristotle's De Generatione

et Corruption, Leonine edition. This principle of order is discussed elsewhere, for
instance, see Aristotle, De Partibus Animalium, Bk. I, c. 1.
6. We must note that the order considered here is not to be cnsidered a
pedagogical order according to which one must expose the young to the subject
matters of the natural sciences. One does not first expose young people to the
natural sciences according to the demands of a perfectly rigorous and systematic
treatment.
7. It is important to understand just what the natural is, and we consider the
definition of the natural S&MS.
-~~~-~----
Chapter 16
Definitions
A. Introduction
Definitions are answers to the question "what is it?" and our desire to
defme arises from an obscurity in some class of things, an obscurity that
we want to overcome. The procedures the mind employs in arriving at
defmitions results, however, from the way the mind develops, and that is
what we wish to show. But first, some preliminaries.
B. Packaged formulas
Usually the ultimate goal of a defming procedure in a scientific

discipline is to produce a "packaged" formula that succinctly tells us what
something is. The package ordinarily is the result of much preliminary
consideration and discussion, and sometimes it comes about as the
consequence of many minds meeting on the problem, which means that
scientific defmitions must be arrived at with care. This appears to oppose
the commonly held view that one may defme as he pleases, which, if true,
would not require determinate procedures of any kind, much less
procedures that are imposed on the intelligence by its very nature. It
would seem plain, then, that the word "definition" has more than one
meaning; so it is important for us to begin by distinguishing the relevant
senses of the term, but to do so we must first make some remarks about
the functions of definitions.
Because defmitions are instruments of the intelligence, and because
instruments are known and characterized by what they facilitate, our first
task is to distinguish the functions definitions can have. Such functions
depend, however, on words, which signify realities to the mind because
they name the realities. That is to say, if a word signifies something to the
mind, then by that fact it is said to name it. Then starting from this very
general notion, we may distinguish words further according to a naming

function or according to a describing function. Once again: we can
distinguish words according to whether they name or whether they
describe something. 1
Words such as "water," "man," "pulsar," "circle," "viscosity," to
mention only a few examples, are words that ordinarily have a naming but
not a describing function. Yet "wet" in "wet dog" has a describing
function, as does "viscous" in "viscous substance. "2
In our earlier discussion of naming we noted that the signification of
words depends upon how proximate to observation various realities are.
We drew attention to the fact that the source from which a name is taken
and that upon which it is imposed are usually different. Especially is that
true when the realities named are substances, for the only way substances
can be appropriately labeled is through observable properties from which
their names are taken. "Grasshopper" is one of the illustrations that we
employed earlier, and it makes our point clearly. What this way of naming
implies, however, is that the describing function words can have arises
from the fact that some realities are observable while others are not; or
more generally, some realities are more easily known by us than others.
The very need to describe arises from the fact that some realities are
obscure while others are less so or not at all. Were everything in nature
capable of affecting the senses, there would be no need to describe, for
nothing would be more knowable to us than anything else; all things
would be equivalently intelligible and none would need to be known
through an inferential process of any kind. Using another philosopher's
term, none would be known "derivatively." Hence the difference between
the naming function and the describing function is rooted in the unequal
knowability of various realities in relation to the intelligence. To be sure,
a word can have two different functions in two different contexts without
changing its linguistic appearance. For example, if someone points to a
swatch and hears "red" in reply to the question, "What color is that?" the
function is one of naming. But "red" describes when one says, "The apple
is red." On the basis of these two functions, then, we may distinguish two
distinct kinds of definition.
C. Nominal definitions
Without doubt when the answer to the question, "what is it?" yields the
name and only the name of the object of inquiry, we are confronted with
Definitions 197
the most common function of definition. That, of course, is why we have

dictionaries: their primary reason for existing is to perform a labeling
function, to point out to us the realities to which names are attached. Of
course, definitions of this sort obviously must make use of describing
words to accomplish their function. Yet through the descriptions which
they provide, realities are distinguished sufficiently to have a name
attached to them; and as we said, that is the primary purpose of
dictionaries. Technical definitions can also be found in dictionaries, but
they are secondary. In addition to an ostensive action, there are, of course,
many ways (etymologies are one) to attach a name by using other words
to do the job, 3 but thought they all impose or attach a name to some
reality the mind has hitherto not identified in that way. And so we shall
call this sort of verbal expression a "nominal definition"; and let us
emphasize that its name comes from its function, not from the fact that it
is composed of words. 4
If the function of a nominal definition is to attach a name, and if names
are assigned to realities by convention, then it is plain that in an important
way nominal definitions are arbitrary and necessarily so insofar as no
particular sound of the voice naturally and unavoidably signifies one and
only one reality. The vocal sound "man" could have been used to signify
an elephant. But because linguistic conventions are determined by the
public as such, or by the public's duly appointed representatives, then
even nominal definitions are not arbitrary in the fullest sense of the word.
Speaking strictly, there can be no such thing as a "private language," since
language is an instrument of communication and therefore public by its
nature. Yet we must not be understood too strictly, for within reasonable
limits one who speaks or writes may legitimately impose limits on how
he uses words or extend them for the purposes he has in mind, assuming,
of course, that his limits and extensions are made known to the audience
whom he intends to address. Hence with these qualifications understood,
we may say that nominal definitions are arbitrary, or stipulative, if one
prefers.
D. Systematic definitions
Speaking of fluids, the authors of a physics text published a few years

ago have the following to say:
The term fluid refers to a substance that does not have a fixed shape but
that is able to flow and take the shape of the container .... 5
That is the first line of the chapter on the statics of fluids. Then later the
i
authors say something else:
For our purposes we adopt the following definition: A fluid is a material

substance which in static equilibrium cannot exert either tangential or
tensile forces across a surface, but can exert merely pressure (compressive
force normal to a surface).
Having given the definition the authors then add a comment:
From this definition we shall be able to derive all the laws that govern the
experimental behavior of ordinary liquids and gases in static equilibrium.
Plainly the first defmition the authors give does no more than tell the
meaning of the word "fluid" and therefore is a nominal defmition stated
in terms of its characteristic observable property, the flowing motion
common to liquids and gases. However, the second definition is
substantially different from the first, for it is a principle of inference
insofar as it implies the laws (the regularities) which reveal what fluids do
under conditions of static equilibrium. Again: the second defmition allows
one to infer theorems that are conclusions derived from the defmition as
an explanatory principle. It allows the physicist to see why fluids under
static conditions behave as they do, which is the ground for our calling it
a "systematic definition." By calling it that we focus attention on the role
of the defmition as an explanatory principle, and it has that role because
it tells us what a fluid is in such a way as to allow us to account for other
behavioral properties a fluid exhibits.
Elaborating a bit, we wish to add that although the definition as stated
does not declare that fluids flow, nonetheless it presupposes such a
motion. Only because a fluid flows is it unable to exert tangential or
tensile forces, etc., and so the nominal definition is presupposed to the
systematic, and it provides the proper starting point, by reason of its own
light. Hence to account for the reaction of the fluid to exterior forces
exerted on it is to account for its passive behavior on the basis of a
fundamental motion that characterizes the substance. 6 But we do not
pretend to pass a physicist's judgment on the merits of the defmition
quoted above. Our aim is only to note the existence of a kind of defmition
which tells what something is as well as possible, providing thereby a
Definitions 199
foundation for inferences that explain certain properties of the defined.

Another example can be provided from chemistry; for usually an acid
is introduced to a student of the science as "a substance with a sour taste
that turns red litmus blue," or something similar, an expression which is
plainly a nominal definition. But when the chemist wants to account for
the acid's behavior in reactions, he defmes it as a "substance that increases
the hydrogen-ion content in water," or something equivalent; and this
second definition is appropriately called "systematic. "7
Joseph Schumpeter provides us an example of this sort of defmition,
insofar as it is attainable in human affairs, when he defines imperialism. 8
Having made a number of preliminary remarks which prepare for the
defmition by contrasting imperialistic aggression to aggression which is
on account of a concrete interest, Schumpeter then goes on to define
imperialism as the objectless disposition on the part ofa state to unlimited
forcible expansion. But the definition appears to include the first property
which follows from it, namely the unlimited character of the expansion.
That is to say, the definition itself would contain only the phrase,
"objectless disposition on the part of a state to forcible expansion," while
"unlimited" is a property that follows on the expansion as objectless.
Evidence that Schumpeter himself recognized unlimitedness as a property
is contained in his own words when he says "such expansion cannot be
explained by concrete interest. ... Hence the tendency of such expansion
to transcend all bounds and tangible limits." He seems to introduce the
objectlessness as the cause of the unlimited character of the expansion,
but because it is a negation, one might better say that "unlimited" is
conceived in the mode of a property rather than say it is a property.
In a right triangle the hypotenuse squared is equal to the sum of the
squares of the other two sides, and this (positive) relation of equality is a
true property of right triangle. Obviously, being a triangle with a right
angle is not the same as being a relation of the hypotenuse to the sides,
just as being objectless (in the sense of not having a concrete interest) is
not the same as being unlimited.
One last point. A defmition that is systematic and tells us as well as
possible what some natural species is so that we may account for the
relevant properties of the natural species is plainly not arbitrary. The
constitutions of natural substances and entities are not produced by us;
hence we investigate them with the aim of exposing what they are in
themselves. Stated another way, once a nominal definition has pointed out
the reality to be investigated, then the systematic definition is subject to
200 From Observab/es to Unobservables
the constraints imposed by the object being investigated, by the attributes

of the reality that has now been pointed out by the name. There are of
course problems when things are so obscure that we cannot obtain a
defmition from experience and have to resort to hypotheses and
postulations. But that is not germane to the general point being made here.
We need to remark that our discussion of defmitions has been limited
to those that reveal or name natural entities. Defmitions of artefacts or of
anything that is the product of human causality are excluded, and for
determinate reasons. In a paragraph above we alluded to the fact that
natural entities are not made by us, on account of which they are objects
of investigation and not of manufacture. They therefore function as the
measure of what our intelligences conceive. On the other hand, the objects
we make, whether they be tools, houses, musical compositions, ballets,
civil societies or what have you, because they are the products of the
human intelligence, must be measured by it. In nature we take the natural
entity as the standard against which we compare our concepts, but in the
realm of the artefactual the designer's conception is the measure to which
the product is compared. Consequently, because a productive conception
is prior to its reality, in a certain sense the definition is independent ofthe
reality; and so the reality must exist according to the formulation of the
defmition. Furthermore, since the "natures" of artefactual realities are
determined by us they are variable in a way natural things are not, on
account of which permanence and determinateness of application are not
to be found in their defmitions. For reasons such as these, then, we have
considered only those definitions that pertain to the world of nature. To
run the two realms together cannot but produce confusion.
E. A summary of the foregoing
Our point has been to distinguish two main kinds of defmition on the
basis of two distinct functions they can have. One function is to attach a
name; the other is to say what something is. As we mentioned above,
there are many kinds of nominal defmition, some of which approach more
closely to being systematic than others. Often, too, they ground
inferences, although not of the explanatory kind. 9 In fact, the reader can
see that any defmition which falls short of being systematic in the sense
described above ought to be classified as a nominal definition. 10
Definitions 201
F. Definition and the natural procedures of the mind
Returning, now, to the goal we declared in our frrst paragraph, namely,

to show that the way we define follows on the human mode of knowing,
we can see that the procedure we follow when we begin by imposing a
name through a nominal defmition and only later arriving at a systematic
defmition is a consequence of the mind's knowing unobservables through
observables. That a nominal defmition has to come before one that is
systematic is fundamentally the result of having to understand that which
escapes observation through that which does not, a noetic relation
obviously true of substances, which, after all, are the principal realities
that we investigate.
Of course, observables themselves at times can be defined both ways
at once, which is to say, one definition can have both functions. For
example, the definition, .figure contained by three sides, not only attaches
a name, it also provides a basis for inferring theorems (in Euclid) about
triangles; so the definition plays both roles at once. Figures are
observables, and the imagination as well as the senses represents them as
they are, which means the mind has no need to hunt further to determine
what a triangle is.
Of course definitions often are not given in the neat packages we used
as illustrations above. Frequently they occur as lists of unordered
attributes, yet such lists do amount to definitions insofar as they are able
to fulfill one or the other of the functions discussed above. Furthermore,
we often cannot obtain genuinely systematic definitions, and we ought not
to think they are always available. 11
The difference between an initial understanding of things based on
observable properties and behavior and one that comes after investigation
and inference from observation ought not to be overlooked, especially
when we are considering descriptions in works that do not pretend to be
systematic treatises. For instance, if a hare is described as a cud-chewing
animal in some literary work, the description ought to be kept in
perspective. It comes directly from the observation of lip-motion and is
not intended to be a systematic classification (even though it might be
defended as such). In short, definitions that are preliminary or nominal,
definitions that start from ordinary experience and intend only to point to
the object, are accurate as far as they go and suffice for ordinary purposes.
A treatise on history or ethics, for instance, does not need to defme
organisms or other natural entities in the way required by the natural
sciences. Hence the categories or classes that are a part of ordinary

experience ought not to be criticized or "corrected" because scientists
have on systematic grounds placed certain commonly known entities
elsewhere, for common experience is not so much inaccurate as
incomplete or imperfect. Fishermen, for example, often use the word
"minnow" to mean any small fish used for bait, but to a fish biologist the
word means something more restricted. But that is enough on these
matters, and with these basic delineations in mind, we may now go on to
talk about the sort of principles philosophy employs.
Notes
1. We hardly need say that we are nottalking about prepositions, conjunctions,

etc.
2. In what follows, we shall be interested only in those names that signify a class,
that is, a common property or nature, or in an older language, a universal. We are
not interested in proper names or equivalent expressions.
3. We limit our considerations here to attaching names through the use of words.
If someone points to an object and then gives it a name, he is said to have defined
ostensively, and that can be considered equivalent to a nominal definition.
4. We ought to note, too, that "working definition" is usually equivalent to
"nominal definition."
5. This and the quotations that follow are from George Shortley and Dudley
Williams, Physics (New York: Prentice-Hall, Inc., 1950), c. 4, pp. 73 and 75.
6. Actually the definition as stated appears to present a property which follows on
the characteristic behavior. If one may rightly say that a fluid is unable to exert
tangential or tensile forces, etc., because it is able to flow, then he has given a
cause of the behavior. Strictly speaking, then, the definition is not fully
fundamental; but that does not affect the general point we are discussing.
7. It is my understanding that chemists are not in perfect accord on the best
definition of acid; but that does not affect our point. Whatever formulation is
presented, it is presented because it is thought to account for behavior in a more
satisfactory way than other formulations.
8. See his Imperialism and Social Classes, Meridan Books, 1955.
9. Operational definitions, although they enter into mathematical calculations and
derivations, and although they ground mathematical inferences, must nonetheless
be classified as nominal definitions; for example, K.E.=l/2 mv2 The function of
such operational definitions is certainly not to say what some physical reality is.
1O.lt is perhaps worthwhile to note that historically what we here call a
"systematic definition" was sometimes spoken of as a "real definition" or an
"essential definition," names which we think are misleading, at least in a
contemporary context. Moreover, sometimes the manuals that employed such
Definitions 203
terms did not, in our opinion, delineate the kinds adequately.

11. This issue is discussed at some length in our S&MS, where the limits of the
human capacity to define systematically are discussed more determinately and
in more detail.
Chapter 17
Analytic Propositions
A. Introduction
One hardly need argue that systematic considerations depend on first

principles and that the careful theorist will attempt to reduce all his
arguments to them. A convenient illustration of such a principle, one that
does not require technical expertise to be understood, is provided by a
meteorologist when he writes:
All of the phenomena that we call weather, from the most gentle of
breezes to the driving winds of a storm, or from clear skies to massive
downpours of rain or snow, are the results of the unequal distribution of
heat energy in the atmosphere.'
According to the author, weather phenomena, from the greatest to the

least, are ultimately accounted for by the statement which says they all
result from the unequal distribution ofheat energy in the atmosphere. And
if we ask where this principle or premiss comes from, then we have to say
first that it is not proved in meteorology; rather, it is taken as a starting
point of the science's explanations. Of course meteorology is a part of
physics, and so its first principles or first premisses are not first in the
same way as are the laws of mechanics, for example. Nonetheless, the
proposition above illustrates well the notion of a first principle; as far as
weather phenomena are concerned, it is the ultimate court of appeal.
Everyone knows that there have to be first premisses in every
systematic discipline. If a proposition (a theorem or conclusion) r is
deduced from another, q, and if q is deduced fromp, etc., then at some
point we must come to a premiss that is not deduced (though it might be
obtained inductively) from some other proposition; and that sort of
statement is what we mean by "first principle." Furthermore, we must
recall from our considerations in chapter 12 that first principles are not
necessarily hypotheses; on the contrary, they can be empirical regularities
obtained inductively from observation. But the issue is not whether there
are first principles, for everyone knows that they occur in every science;
the issue is whether they are anything more than regularities obtained
from observation.
To say that the nature of first principles which are not hypotheses is a
somewhat confused issue is not to overstate the case, and an illustration
of how propositions are thought to be divided may help to set up the
problem we wish to discuss:
What sorts of features ofmicroeconomic general statements ought one to

inquire about? Perhaps the first question that comes to a philosopher's
mind when presented with a body of general propositions is, What is the
"logical status" of the statements under examination? This is usually taken
to mean are they "analytic" or "synthetic"? That is, are they disguised or
undisguised definitions, or the consequences of definitions that are simply
true in virtue of the meanings ofthe terms they employ? Or are they ampl-
iative, contingent, empirical propositions, whose truth (or falsity) hangs
not merely on the meanings of their constituent terms, but on
extralinguistic, factual considerations, and, it is often claimed, ultimately
on observable facts? 2
We have, the author says, two kinds of propositions, one that is

synthetic, contingent, and ampliative-that is, informative about the real
world-and another that is analytic and true solely by virtue of the terms
in which it is expressed. Propositions of this second kind cannot function,
we are told, as first principles because they are not explanatory. The
author makes this point when he says:
An analytic proposition is true, come what may. It is consistent with the

occurrence or nonoccurrence of any logically possible event, state, or
process. But if a proposition is consistent with anything at all happening,
then it could hardly be used to explain what actually does occur; nor could
it play any role in the prediction that some particular event will occur.
After all, ex hypothesi, it is consistent with anything at all happening,
while actual occurrences exclude some of these possibilities. It is for such
reasons that propositions which are true, come what may, are described as
telling us nothing about the world. Because the purpose of a scientific law
is to make a claim about the world, it is important to exclude analytic
propositions from the body of such laws.
Analytic Propositions 207
Rosenberg tells us that frrstprinciples must be either empirical regularities

or hypotheses. Analytic propositions are not a real choice, it seems,
because they are empty and cannot on that account explain anything at all
in the real world; after all, they are consistent with anything at all
happening. So our problem will be to consider the nature of those first
principles that are indeed analytic, for it will be our contention that
philosophy starts from them. We have already said that philosophical
explanation cannot employ hypotheses as typical starting points, after the
fashion of the natural sciences; nor can it start from those empirical
regularities which require explanation, for such propositions (as everyone
knows) are by nature conclusions. But we have said enough about his
text; so let us take up the issues. To begin, then, let us look at some
propositions that are recognizably instances of first principles that
everyone will recognize.
B. Instances of first principles: those that are common
We are all aware that there are some first principles which are
employed by every systematic discipline. The statement, "Anything either
exists or does not exist," or what is the same, "It is impossible for
something to be and not to be at the same time and in the same respect
(the principle of non-contradiction)," is the most obvious case. The mind
cannot take a single mental step without recognizing, at least implicitly,
that the principle is true; for anything whatsoever that we consider either
does or does not exist. Another example, one that is sometimes not
understood properly is, "A whole is greater than any of its parts." Still
another, less frequently articulated common principle, is "Anything
received in any way whatsoever is received according to the limitations
of that which receives." These statements are clearly true, and they clearly
apply to every category of reality. Wherever there are existing realities,
or wherever there are wholes and parts, or wherever something passively
receives an action by another, these principles apply.
Still another example of the same sort is "Two things equal to a third
are equal to each other," a statement upon which the whole of
mathematics depends. 3 Of course such propositions do not enter directly
into the arguments of the sciences; they do not provide explanations
proper to the special subject matters, but instead form part of the
background. 4 Nonetheless they are recognizably indispensable first
principles, and we understand such propositions to be necessary. But that

is what raises the problem: if they are first principles and at the same time
are necessary, what sort of propositions are they?
C. Analytic propositions: preliminaries
The burden of the first part of this essay has been to show how the
intelligence starts from observables and moves from them to an
understanding of unobservables. The position defended is realist, and
consequently we argued that philosophy bears on regularities derived
from sensory data. We have maintained that the origin of our knowledge
is singular entities, since they are what actually exist and what actually
affect the sensory powers. As a consequence of the mind's dependence on
sensory data for the origin of its concepts, we do not possess concepts that
are unrelated to observable properties of singular physical entities.
Furthermore, if the first propositions that we know are derived from
experience and observation, then they are affirmative and have what in
the past has often been called "existential import"; or as the medieval
Aristotelians would say, they have "personal supposition." One might
state the matter still another way and say that such propositions have
"referents." We also maintain that negative propositions have "existential
import" insofar as they say something about the real world; for one must
keep in mind that negation presupposes affirmation. Thus, if negative
propositions are to be formed at all, they must deny in some manner or
other that which something else has affirmed and therefore is taken to
exist. That is how a negative proposition has "existential import." Using
the Aristotelian vocabulary, we may say that affirmative propositions that
are non-supposing (that do not have referents or existential import) are
false. Only a negative proposition can be true when non-supposing, that
is, when it is not applicable to some particular entity or class of entities
"out there." On that account, the square of opposition for the real sciences
is that which Aristotle presented, and it may not be denied that the
systematic disciplines (we except mathematics) do indeed deal with
propositions that have personal supposition or existential import or
referents.
But if this is the position one takes in regard to propositions in general,
it is also the one he must take in regard to first principles. That is to say,
first principles in genuine disciplines cannot be of such a nature as to deny
their relation to real, individual things; they cannot be propositions that
deal with concepts alone in separation from the things to which the
concepts may be applied.

No one ought to think that we are denying the mind's ability to
examine its concepts separately, that is, to examine them as apart from
and distinguished from the singular entities in which they are embodied.
On the contrary, a sophisticated intelligence must be able to do just that:
it must be able to distinguish that which belongs to the concept as
something existing mentally from the reallity in which the concept is
embodied. But the consideration of our concepts according to their status
in the mind is reflective; it is not a priori. Such a status is necessarily
posterior to and in no way independent of the concept's realization in
singular entities. That is what the realist position we are defending entails.
That the first principles we presented as examples are not hypotheses
need not be argued, and it seems to be equally clear that they are not
empirical regularities of the sort that count as scientific laws. But then we
must ask: just what sort of propositions are they? Despite Rosenberg's
assertions, the likely candidate seems to be analytic proposition, mainly
because it seems to be the only option that is left. If, however,
philosophy's first principles are analytic propositions, then it would seem
that they cannot be the sort of thing Rosenberg has claimed them to be; so
once again, what kind of proposition are they?
D. Analytic propositions: the Kantian view
The reader probably knows that the notion of analytic proposition

which Rosenberg mentions is representative of what is current among
philosophers, and he probably knows as well that the view is essentially
Kantian; so it would seem desirable to consider the doctrine of analytic
propositions in its genesis, that is, to consider its original formulation as
it was given by Kant. For that reason we wish to quote the following text:
In all judgments in which the relation of a subject to the predicate is

thought (I take into consideration affirmative judgments only, the
subsequent application to negative judgments being easily made), this
relation is possible in two different ways. Either the predicate B belongs
to the subject A, as something which is (covertly) contained in this
concept A; or B lies outside the concept A, although it does indeed stand
in connection with it. In the one case I entitle the judgment analytic, in the
other synthetic. Analytic judgments (affirmative) are therefore those in
which the connection of the predicate with the subject is thought through
identity; those in which this connection is thought without identity should
be entitled synthetic. The former, as adding nothing through the predicate

to the concept of the subject, but merely breaking it up into those consti-
tuent concepts that have all along been thought in it, although confusedly,
can also be entitled explicative. The latter, on the other hand, add to the
concept of the subject a predicate which has not been in any wise thought
in it, and which no analysis could possibly extract from it; and they may
therefore be entitled ampliative. 5
According to Kant/ when we form a proposition of the sort he called

analytic, we merely say in the predicate what has already been thought in
the subject, and the predicate merely reflects the mind's breaking up of the
subject into its constituents. Hence today the analytic proposition is
represented as being of the form, xy is y, which means that the necessary
truth of analytic proposition depends on the impossibility of contradiction;
that is to say, it depends on the necessity of something existing or not
existing at any one time, but not both at once. Thus, if something is an x,
it is necessarily an x as long as it exists, and it cannot be a non-x, all of
which is obvious. In short, analytic propositions are taken to be statements
of identity the denial of which is most obviously contradictory.
As Rosenberg indicates, a more contemporary view of analytic
propositions has it that they are no more than linguistic conventions
determining how we use certain words, of which "A bachelor is an
unmarried male" is a well-known illustration. Analytic propositions are
thought to be linguistic identities because their subjects and predicates
signify the same thing, saying essentially no more than statements such
as, "John is John" or "A horse is a horse," neither of which adds a whit to
our understanding.
That the two views described above are not incompatible and that they
are true if one grants what Kant has said about analytic propositions, need
not be elaborated. We might say that given the Kantian definition, the
modem critiques are correct. But there is an overlooked qualification in
Kant's description which points to another exegesis of what may
legitimately be called "analytic proposition," the overlooked qualification
being his remark that the predicate is contained "covertly" in the subject,
that the predicate is all along thought in the subject, "although
confusedly." Now it seems to me that these qualifications are correct and
that Kant recognized something he could not exploit, given his views of
the a priori and of experience. But if we consider the qualifications
carefully in the light of what we have done in previous chapters, they lead
to a legitimate but non-Kantian doctrine of analytic propositions.
Therefore we shall take up the issue and present what we think is the
position of the ancient Greeks, a position which says that although
analytic propositions are true by definition, and although subject and
predicate signify the same thing, the propositions themselves are neither
conceptual nor linguistic identities. First, however, we must undertake
some considerations on necessity in things; for if analytic propositions are
not linguistic or conceptual identities, and if somehow they depend on
realities, then in one way or another there must be necessity in contingent
entities. So to begin, let us look once more at the examples of common
principles we introduced earlier. 7
It is not hard to see why the principle of non-contradiction has to be
true, for "exists" and "does not exist" are absolutely incompatible in the
real world. The issue is not just one of words or an identity of concepts,
for non-existence is the absence "out there" of something existing. First
we note the difference between existing and not existing and then we
attach the two names. Nor is it much more difficult to see that "Every
whole is greater than any of its parts," is also necessarily true and that
"whole" and "part" are attached to extra-mental realities. And since these
statements originate from our experience of real things, they are first
obtained by an induction from observation, even though the ease with
which we see them puts them, paradoxically, outside the list of inductions
we most quickly recognize. To be sure, the inductions themselves do not
show that the propositions are necessary; yet we nonetheless recognize
that they are. Their necessity must come, therefore, from the realities that
ground the inductions, and the necessity must be accessible to the mind
through something other than the inductions themselves.
At this point we must note apropos "The whole is greater than any of
its parts," that there is a tendency for some to think that "whole" and
"part" can be used for any arbitrary association whatsoever, even an
association that is wholly mental. Grand Central Station and a patch of
blue sky do not, to take an example, constitute a whole, except insofar as
"whole" has been given an arbitrary extension that does not respect the
original attachment of the name. Convention first imposes the words
"whole" and "part" upon physical units that are integrated out of what we
recognize as subunits constituting the complete units themselves.
Sometimes the units are amounts of homogeneous stuffs that are divisible
into homogeneous subunits, while at other times they are made up of
heterogeneous parts. But in either case it is upon such realities that we
impose the names "whole" and "part." A random collection such as an old
beer bottle, an egg shell, a tree branch, and a bird bath, even though they
are all found in the same back yard and make a "pile," they cannot be
integrated into a whole, except in the most incidental way. Thus, to extend
the names "whole" and "part" to anything whatsoever we wish to consider
as "united" whether or not they are in the world outside is to distort the
language. It is to ignore the restraints placed upon the intelligence by the
first impositions of words according to which the words are attached to
independently existing realities. Put another way, we must respect the
noetic and hence linguistic limitations imposed upon us by the realities
that are signified, and we must not think that we may without sacrifice
extend a name to whatever suits our fancy. Contrarily, if the imposition
of words were purely arbitrary, if it were completely ad libitum, then the
mind would indeed construct what it knows.
Perhaps it is unnecessary to say that our discussion of first principles
does not include practical rules that tell us how to design, make, produce,
or otherwise cause or carry out some work of human invention. As we
said when speaking of defmitions, works of human origin are excluded
from our considerations, an exclusion that applies here, too; for
disciplines that tell us how to produce or do something provide
immediate, practical principles that are unlike the kind of first principles
nature provides for investigative, purely theoretical sciences. On the other
hand, since practical sciences presuppose natural entities either as
materials or as agents, they make use of some first principles that belong
to the natural sciences and are not properly their own. But putting that
aside, let us face the basic issue directly and ask how the necessity in
things that founds analytic propositions is accessible to us? How do we
obtain necessary propositions from contingent entities? What is our
ground for denying Hume?
E. Necessity in natural things
As the reader well knows, beginning with John Locke and David
Hume, philosophers have been inclined to think that, since natural entities
are contingent, the mind cannot derive necessary propositions from
sensory impressions. A necessary proposition about a contingent reality
is taken to be a contradiction in terms; so that would seem to be the first
point to consider: how can there be necessary propositions about
contingent realities?
To start, let us note that there is a difference between the contingent or
possible that is purely logical and the contingent or possible that is

physical. 8 The logically possible is represented by the modal proposition
of that name, and it demands no more than the compatibility of subject
and predicate within a proposition. For instance, a unicorn appears to be
conceivable and hence logically possible, for the fictional unicorn does
not present any visible (to the mind) contradiction with the notion of
animal or any of the categories implied by animal. Yet no unicorn is
known to exist. Thus, the possible can be said of what exists and what
does not exist, and because this kind of possibility requires only that
subject and predicate be compatible, both contingent and necessary
propositions come under it in a way resembling that of a species under a
genus, or some subdivision under a main class. For instance, apples and
red are compatible, so one can say that their union in a proposition is
possible. But the same proposition is also contingent because in the real
world the predicate belongs to apples in some cases and in others does
not; both the affirmative and negative propositions are at times true. On
the other hand, a purely possible proposition that is not contingent, that
does not apply to existing things, could not be true, at least not according
to a correspondence theory of truth. And one must place the necessary
with the contingent, for the relation of necessity requires not only the
compatibility of the subject and predicate but their inseparability as well.
The proposition, 2+2=4, is always true, for its terms are not capable of
being separated. Hence their relation, though necessary, can be put under
the more general heading of possible.
Physical possibility, on the other hand, is more than a relation between
terms of a proposition; it is more than a description of what belongs to
things according to their mental status. As we all know, physical
possibility or potentiality is first of all an attribute of existing things. The
actually existing table is potentially a stool, a small chair, or a large
number of pencils, etc. Organic materials not yet organized are actual
stuffs that can be described as a potential or possible organism. On the
other hand, purely logical possibility is not a property of real things, but
a relation between internal parts of a proposition, a relation that of itself
is compatible with realities that are either contingent or necessary. But we
need not labor the point further. It will suffice to say that the contingent
to which the necessary we are now discussing is opposed is the contin-
gency of real things; it is the contingency that is found in physical entities.
But how, we shall ask again, can such things be necessary?
Entities are said to be contingent because they can exist and not exist;
r.
.
'
~
'
'
they can come to be and pass away. And if that is what contingent entities
are, then it seems that necessary entities must neither come to be nor pass
away and so be everlasting .. Such would seem to be one meaning of
"necessary," and certainly there is a clear opposition between "contingent"
and "necessary" as they are so described. And were that the the sense of
"necessary" we must use, then clearly only the conceptual can be in any
way everlasting; nothing in nature meets the description. We must not,
however, overlook another distinction that nature imposes on us.
If one starts not with entities but with properties, then he can see
another kind of contingency. Man, for example, is able to sit or to stand,
to walk or to run, etc. because he has operational capacities that are open
to alternatives. On that account we say that sitting and standing are contin-
gent-men contingently sit and contingently stand. Now this notion of
contingency does not have to do with the coming to be or passing away
of the entity itself, but with the coming to be and passing away of a
property in it. The opposite of this sort of contingency is a determination
inherent in the constitution of a thing which does not allow such
alternatives. We are not, for instance, able to opt for flying instead of
walking. Our operational capacity is limited, determined, to the action of
walking; it is not open to flying. In short, there is a necessity in things
which is none other than a determinate orientation to one property,
operation, or characteristic rather than to more than one.
An example of still another kind of contingency is skin color in the
human species. Skin color is not determined by man's nature; for if that
were so, we would all be alike, according to one pigmentation or another.
Yet skin color is not separable in a physical way from the individual, as
are standing and sitting. The contingency, then, is on the universal level,
insofar as the constitution of man which one conceives to be distinct from
color and other properties does not imply one pigmentation to the
exclusion of others. In contrast to these contingent properties stand others
which, although they belong to contingent entities, are not variable within
the species in either of the ways described above. Consider, for example,
a photon, which, physicists say, cannot stop moving at 300,000 kilometers
per second as long as the photon exists. To arrest it is to cause it to cease
to be. Moreover, it is not the case that some photons travel at the named
speed while others travel at one that is lesser or greater. Hence, although
the photon is a contingent entity, the relation between it and its motion is
necessary. That is to say, as long as they exist photons necessarily move
at 300,000 kilometers per second. But photons are theoretical entities;
they are postulated in hypotheses to account for esoteric phenomena not

a part of ordinary experience, and on that account we might be accused
of resorting to the purely conceptual to illustrate our point. We are well
advised, then, to look at other examples which reveal necessary relations
on the macroscopic level.
Ordinary physical substances such as water or salt or oxygen are
defined in terms of their physical properties. But interestingly enough,
these sets of physical properties are not open to variation. Under the same
conditions oftemperature and pressure all samples of water are the same.
Normally water is colorless, tasteless, a solvent, non-conductive, a liquid,
and under the same conditions we do not fmd some samples to be solid,
colored, have a flavor, not to be a solvent, to be conductive, etc. 9 Nor
under the same conditions of temperature and pressure is oxygen found
to alter its properties; all samples of the stuff are alike. And what we have
said about oxygen and water is true of all natural substances: the sets of
properties that define them are constant, invariant. In short, all the
evidence indicates such properties are not contingently related to their
substances; and we may rightly say that the relations between the
properties are necessary, not to mention the relation of the properties to
the substance they make known. 10
And so to summarize. Observation shows there is a difference in the
relation various properties have to the substances in which they are found:
some properties are related variably to their substances, while others are
related invariably. So although the entities themselves may be contingent,
connections among some properties of natural things are in many cases
invariant and so presumed to be necessary. Put another way, the absence
of what we know to characterize contingency in the forms described
above warrants the conviction that some properties are necessarily related,
although the full understanding of the relation cannot come without
actually seeing the necessary connection. For instance, one might have
noted as a regularity that the square of the hypotenuse of a right triangle
is equal to the sum of the squares of the two sides and not know why that
has to be. In brief, the proof of the theorem through a causal relation
brings the necessity of the connection fully into light, which is what we
have to do to show the necessary connections among properties and the
natural substances to which they belong. But an argument can show a
necessary connection only if the premisses themselves are known to be
necessary, and that is the issue we shall address in a moment. First,
however, an emphasis.
Though the point has already been made, it is worth repeating that not
every natural entity has a fixed set of properties which admits of no
variation; for what we have said about certain inanimate substances is not,
as everyone knows, true of organisms. Individuals within a species
obviously can vary, a state of affairs apparently not possible with
chemical elements. 11 Horses, for instance, can vary in size, color, shape,
running abilities, hair length, etc., without thereby belonging to different
species; and what is true of them is true of other organisms as well.
Nonetheless, there are certain attributes and certain operational capacities
without which one does not have the species, which means that some
attributes are contingent and others are necessary. There is a difference,
then, between the animate and the inanimate: the former have both contin-
gent and necessary properties.
The existence of necessary relations in contingent entities should not
be denied on the ground that individual organisms can depart from the
norm. Man, we say, is an animal that is able to speak; yet some
individuals are born dumb and remain forever so. But to hold therefore
that man is not necessarily a loquacious animal is to miss the point; for the
necessary relation is between the species-kind and the property that is
consequent upon the kind. Accidents can and do happen to individuals
both before and after birth, accidents that prevent the exercise of the
normal functions; and the ability to suffer accidents is indeed still another
kind of contingency. Yet, as we said, such states of affairs do not allow
us to say that man is both a loquacious animal and a non-loquacious
animal, for we cannot hold that both muteness and the ability to speak are
states natural to the species in the sense that they follow on it. Neither are
they contingent in the manner of alternatives, or in the manner of a
contingent property. In short, we recognize dumbness as abnormal and
hence as not rooted in the human constitution but as coming instead from
the capacity the materials have to be affected by foreign influences during
the generative process.
But this amounts to still a third kind of contingency, one that stems
from the inability of the constitution of the organism to prevent exterior
agents from introducing unwanted modifications into it. Such a
contingency is unlike both the capacity for alternatives and the possibility
of the species to receive more than one kind of property within its genus,
for instance, skin color. Thus, although accidents can and do happen to
individuals, we are nonetheless entitled to say that a necessary relation
exists between, say, the rational nature which defmes man and his ability
to communicate by artificial speech sounds. Indeed, the kind of

contingency that characterizes the abnormal individual presupposes a spe-
cies-determination that of itself is causal of an attribute, property, or part,
but is yet incapable of excluding another deviant attribute, property, or
part from appearing in the individual. And what is true of man is true of
other kinds of organisms as well.
So we see that although natural entities are contingent, necessary
relations may be presumed to exist between a substance and some of the
properties it exhibits, on account of which we may expect to form
immutable propositions about them. In sum, the contingency of natural
entities does not permit us to say that no necessary propositions can be
formed about them. Furthermore, if the necessity of some propositions is
seen only as the result of an explanatory argument, then as we have
already said, the origin of our seeing necessary connections in natural
entities depends upon our first seeing necessary connections in the starting
points; on them, everything else depends. So having said these things, let
us now move to our direct considerations on the kind of principle we
think analytic propositions are.
F. Analytic propositions: a realist, empirical view 12
To begin, let us note that analytic propositions are twofold, those that
are common, and those that are proper. Analytic propositions are common
when they are used by or presupposed to every science; but they are
proper when they pertain to a determinate species or genus of reality and
cannot be extended beyond that. An illustration of a common principle is
"Every whole is greater than its part," while an instance of a common
principle is "Man is a reasoning animal," a proposition that will be
discussed more at length later.
If we now ask the question, how do we know that the common
principle, "Every whole is greater than any of its parts," is necessarily
true, the answer has to be: we see its necessity as soon as we know what
wholes and parts are. That is not to say we see the proposition to be
necessarily true because in conceiving the subject we conceive the
predicate, nor is it to say that as soon as we know to what realities the
names are attached we know the proposition to be an identity, a tautology.
An analytic proposition is neither a conceptual identity nor a linguistic
convention. Rather, we may rightly say that we know the proposition to
be necessarily true as soon as we know what the terms mean, provided
that "knowing what the term means" is taken to say that in knowing the
meaning of the terms we actually get at what the realities are. And though
in the case of common principles such as those which served as
illustrations earlier, there is no difference between attaching the name and
knowing what the reality is, such a state of affairs does not always obtain,
a point we made at length in our chapter on definitions. Indeed, in the
case of proper principles outside mathematics, we do not know what
things are until after we have named them and investigated them at some
length. Yet analytic propositions are true by definition; and that, of
course, is what has occasioned both Kant's and the analysts' views. Put
another way, because our first definitions of common realities suffice to
tell us what things are, we precipitously tend to think that every necessary
proposition comes either from a noetic identity of subject and predicate
or from the arbitrary attachment of the name to the object signified. The
consideration of proper principles, however, shows how far from the truth
such a confusion is.
The classic example of an analytic proposition that is a proper and not
a common principle is the statement, "Man is a rational animal,"
"rational" having to be understood as "able to deduce conclusions from
premisses (or theorems from axioms)." 13 At first sight the example does
seem to be an identity in Kant's sense; it seems to be a statement in which
the predicate declares what we actually think in the subject and therefore
comes more from the intelligence than from induction and observable
realities. But the first impression is deceiving. Certainly the proposition
is an identity in the sense that what is signified by the subject term is
identical to what is signified by the predicate; the referents, one might say,
are the same. Yet there is an important difference, for the term "man" does
no more than name the entity; it does not signify distinctly what that sort
of animal is. In short, because the word "man" properly signifies an
unobservable substance, what man actually is cannot be known except
through the predicate, which is taken from an activity known by
experience, thereby revealing the character of the underlying substance.
Again: the predicate signifies more clearly because it comes from a reality
known by observation, both internal and external, whereas the interior of
the substance is known by inference from them. That, of course, is
precisely what Kant overlooked, and it is why he thought the predicate is
already conceived in the subject. Thus, although the subject and predicate
terms signify the same reality, the proportions of the subject and the
predicate to the human mind are considerably different, so different that
the mind cannot extract the predicate from the conception of the subject.
Certainly an analytic proposition is not of the form, xy is y.
The example, "Man is a reasoning animal," does present a certain
difficulty insofar as our familiarity with ourselves tends to make us look
on the defmition as nominal; it seems to be the phrase by which the name
of the species is attached. But a truly nominal defmition of man would be
"two-legged, featherless, TV watching animal," or something similar, in
which the describing traits would not be liable to confusion and would be
attained by external observation. Furthermore, when we understand "able
to reason" in a thorough way, we see in it why man is political, why he is
artistic, as well as why he has many other properties, none of which can
be read from a merely nominal definition. In short, the defmition of the
species man does what a systematic definition is supposed to do: it shows
why certain properties follow necessarily on his constitution. 14
It would seem to be plain by now that analytic propositions reflect the
mode ofknowing peculiar to the human mind, namely, that unobservables
(substances especially) must be known through observables (properties),
and the generic must be known (systematically, distinctly) before the
more specific. On that account the subject does not already- upon being
named.;_tell us what its predicates are; the subject does not contain them
noetically, though it does ontologically. Analytic propositions that are
proper principles, therefore, assign systematic definitions to their subjects;
nominal defmitions do not suffice. A systematic defmition, if it is
taken-as it ordinarily is-to be constituted by a category or genus along
with a difference that determines the species, cannot be formulated for
categories which are themselves first and most common and have nothing
more common over them. Although they are seen to be necessarily true
as soon as we know what the terms signify, what they mean, the
statements that contain them are not systematic formulations in the
ordinary sense but instead are circumlocutions; for common principles
cannot be stated in terms of a common genus divided by a difference that
is not intrinsic to the genus. And a single term applied to more than one
of them is necessarily equivocal, although in an orderly, non-incidental
way. Yet the awkwardness of speaking about such common categories in
circumlocutions does not prevent the mind's knowing what they are in
themselves.
And so to speak generally, we may say that an analytic proposition is
one in which the predicate enters into the definition of the subject in some
way, a formulation that covers both common and proper principles. We
must add that we cannot say the subject enters into the definition of the
predicate, though we cannot discuss the point here. Such a proposition is
not analytic but a priori synthetic, to use a term favored by Kant.
Furthermore, a systematic definition of a property demands seeing the
cause of the property in the subject; hence until such causes are known,
no analytic proposition about properties is truly possessed. 15 So to repeat:
the predicate of an analytic proposition must belong in some way to the
defmition of the subject; yet such statements are not redundancies or
tautologies or pleonasms, in which subject and predicate signify the same
thing without any noetic difference.
G. A summary of points made above
An analytic proposition is known to be necessarily true through the

terms intrinsic to it. It is an identity insofar as subject and predicate
signify exactly the same reality, but it is not an identity in the sense in
which Kant understood the term, nor an identity in the sense of the
analysts who take "A bachelor is an unmarried male," as a prototype of
such propositions. The subject and predicate are ontologically identical
but not noetically. Moreover, though analytic propositions are first known
by induction, the induction itself does not show the proposition to be
necessary. Understanding its necessity results from an analysis of the
proposition after experience reveals it as an invariant regularity. Thus we
come to see that fundamental to the character of analytic propositions
-especially those that are proper principles-is the mode of human
knowing: subject and predicate are not equally knowable, which means
that the mode natural to the mind requires that we form analytic
propositions by seeing the subject we initially do not know (except
superficially) through what we initially do know.
H. Considerations on some of Rosenberg's examples
At this point we must note once more that analytic propositions

properly belong to the realm of those realities not caused by men, and we
must make the point again because the confusion occurs in some social
sciences, which Rosenberg's examples illustrate. Disciplines whose aims
are to consider things we make or do insofar as they are caused or
controlled or performed by us do not have analytic propositions as they
have been described here, except insofar as the realities they talk about
are described or defined in terms that are common and related to the very
constitution of the human species; that is, to that in man of which he is not
the cause. But we need the examples to illustrate the point. 16
I. If income increases, prices remaining constant, and the increased income

is spent, then there must be an increase in consumption.
2. Two commodities are substitutes if both satisfy the same needs of the
consumer.
3. Two commodities are complements if they are consumed jointly in order
to satisfy some particular need of the consumer.
The first point to be made about the examples above is that they are
not expressed as simple, categorical propositions but as compounds, as is
evident from the conditional verbal form relating an apparent antecedent
to an apparent consequent. The first of the examples is truly a conditional
compound and hence not a first principle. Yet despite the way it and the
others are written, all the statements can be cast as categoricals and hence
cast in the mode proper to analytic propositions.
The first instance presents a difficulty, however, for restated in
categorical form it reads, "Every increase in income when prices are
constant brings about an increase in consumption." The difficulty is that
although the proposition is causal and on that account a candidate for
analyticity, it is not in the least necessary; for the relations expressed are
between contingent human activities: that of increasing income and that
of consuming more goods. One does not invariably result in the other. 17
The second example is a better candidate, for it can be restated as:
substitutes are commodities that satisfy the same needs. Here, however,
the terms of the proposition are obviously general; they are generic and
apply to the entire category of physical goods, if one takes commodity in
that sense of the term. The generic character of the terms, related as they
are to the common human constitution which requires physical goods, is
the reason why the proposition can be said to be analytic; and a similar
case can be made for the third.
I. Closing considerations
The medievals described an analytic proposition as per se nota,

literally, "known through itself," a phrase often translated as
"self-evident." The phrase, although defensible, is not a clear rendering
of the Latin, in evidence of which is the current use of "self-evident" to
mean "obvious." An empirical proposition such as "John is sitting in the
r'
comer" is self-evident in that sense of "self-evident," but it is by no means

analytic or per se nota. Hence we have avoided using the term, preferring
"analytic proposition," which says the same thing asperse nota.
The number of analytic propositions in any given discipline depends
on the number of distinct species found in it. If, for example, one is a
mechanist or reductionist, he would expect to fmd only a limited number;
that is, only those propositions that describe fundamental particles. (We
are ignoring for the moment that such propositions are hypotheses.) But
if the species are substances that exceed in their characteristics the
attributes that defme elementary particles, then the number of analytic
propositions will be larger; for each species can in principle if not in
practice give rise to at least one. 18 Of course, the more general categories
of natural substances can be defined so as to yield analytic propositions;
and there are others that relate first property-effects to causes. But their
consideration is worthy of an essay of its own; so we are able only to state
the point.
With this our discussion of analytic propositions is finished. There
remains to be noted only that they are the sort of principles which are the
foundation of philosophy; they are the kind of principle philosophy must
seek in the measure possible. That is not to say philosophy always and
everywhere attains such starting points, nor is it to say that no probable
premisses appear in philosophical considerations; for even hypotheses
may be employed. There are enough analytic first principles, however, to
make their pursuit worthwhile and to make plain that philosophy cannot
rightly imitate the natural sciences by using hypotheses as its
characteristic starting points. Such an error, we think, is destructive, and
were it true, better no philosophy at all; the sciences would suffice.
And so we have finished our task. We hope that what we have done
will be of use to philosophers and that they will attempt to show how
everything they do is contained in the observable facts from which they
start as their principles.
Notes
!.William L. Donn, Meteorology (4th ed.), (New York: McGraw-Hill Book

Company, 1975.), ch. 2.
2. Alexander Rosenberg, Microeconomic Laws (Pittsburgh, Pennsylvania:
University ofPittsburgh Press, 1976). This and the quotation that follows are from
chapter 2, pages 15-18.
3. A more generalized form of the proposition, one not tailored to the category of
the measurable, is: "Two things which are the same as a third are the same as each
other."
4. We shall say more about the metaphor, "background," when we talk about
hypotheses.
6. Critique of Pure Reason, tr. Norman Kemp Smith (New York: St. Martin's
Press, 1965), p. 48.
7. Henry Veatch has written a thoroughgoing critique of Kant's doctrine, leaving
nothing undone. His exposition is superior to anything I might do here, and I
recommend that the reader see his Two Logics, (Evanston, Illinois: Northwestern
University Press, 1969), chapter III.
8. What we shall say here about analytic propositions, propositions that have been
called per se nota is not new and will be familiar to many. We must take up the
issue, however, to relate such propositions to observation.
9. The terms "contingent" and "possible" are often taken as synonyms.
10. We need not say that we are excluding flavors, colors, and other properties
that come from the introduction of foreign substances into water to make a
solution or suspension.
II. The issue of the necessary relations among properties is discussed at length
in ourS&M.
12. It would seem that isotopes are not a problem; the reader will not find the
differences they represent to cause trouble.
13. The reader ought to see John D. Beach, "The Act of Analysis," The Thomist,
Vol. 44, No. I, Jan. 1980. This is a fine article and covers aspects of analytic
propositions we cannot take up here.
14. One must not take "rational" in some vague sense equivalent to an ambiguous
"intelligent," a word too often applied to non-human animals and covering many
cognitive operations that are not intellectual in the human sense of that term.
15. Actually seeing the cause-effect relation between rational and the ability to
speak, etc., entails a detailed understanding of the nature of the intelligence, an
understanding that is not had simply on the basis of our spontaneous familiarity
with our own internal processes.
16.The position we would defend regarding the relation of a property to the
subject for purposes of definition is only in a small way similar to the Kantian
view. We would agree with Kant that the concept of the subject does not include
in itself the concept of the predicate, even though the subject must be seen to be
the cause of the property signified in the predicate.
17. Rosenberg, Op. cit., pp. 17- 18.
18.It seems to us that this example is offered as an instance of a proposition that
is per se in the fourth mode, to use the Aristotelian description of the type. That
mode is the one in which the formality under which the subject is seem is that of
cause in regard to the predicate; the cause-effect relation is evident in the example
given.
19. For a fuller discussion of our ability to define natural species we refer the
reader to S&MS.
Chapter 18
Recapitulation
Having made our case as well as we are able, perhaps it is
advantageous to recapitulate what we have done in the chapters that have
gone before. The center of focus is of course the understanding-
particularly among those who are realists-that human intellectual
knowledge depends necessarily and in a fundamental way on the sensible,
observable traits of natural entities, not only for its understanding of the
sensible traits themselves, but also for its understanding of the
unobservable realities to which we constantly have recourse, not only in
the sciences, philosophy, and theology, but in our ordinary discourse as
well.
We began by posing the problem and then pointing to our commonly
accepted starting point, namely that we must reduce what we say to facts,
which are either empirically given or are inferred from empirical origins.
But that required that we first of all point to the properties of things that
are admittedly observable, and then show that the mind's first movement
to an unobservable depends on a causal relation between the observable
and the unobservable. Once unobservables come to be known to exist,
they can then be compared to observables on the basis of the similarities
and dissimilarities the we see in them.
Expanding on points already made, we saw the constitutional
dependence of the rational intelligence on observables reflected in our use
of analogies and metaphors, as well as signs and symbols, the principal
signs being words. Because language is our ordinary means of
communication, it must be carefully examined so that we can see the first
meanings of words, which ordinarily are those that are attached to
observable traits, either external or internal. Having looked at these issues
we then moved to broader considerations on myth and its role in the
development of human learning, after which we showed the errors of
those who claim that observations are theory-laden. Finally in Part I we
argued that the human intelligence cannot have innate or a priori ideas or
concepts.
Beginning with Part II, we looked at regularities as the observable
foundations of systematic sciences, moving from there to the regularities
that are the principles of the natural sciences. Following that with some
preliminary reflections on philosophy, we next looked at its principles,
which are analytic propositions and defmitions. That terminated our
labors.
BIBLIOGRAPHY
Anthes, Rudolf. "Mythology in Ancient Egypt," ed. Samuel Noah

Kramer. Anchor Books. Garden City, New York: Doubleday &
Company, Inc., 1963.
Aquinas, Thomas. Questiones Disputatae de Potentia. Rome: Marietti,
1949.
Commentarium in Aristotelis_De Generatione et Corruptione,
ed. Leonine.
Commentarium in Aristotelis Librum de Anima. Rome:
Marietti, 1948.
Aristotle. The Works of Aristotle, ed. W D. Ross. Oxford University
Press, 1928, Vol. V, De Partibus Animalium.
_Metaphysics.
Bernard, Claude. An Introduction to the Study of Experimental
Medicine. Trans. Henry Copley Greene, New York: Dover
Publications, Inc., 1957.
Bridgman, Percy. The Logic of Modern Physics. New York: The
Macmillan Company, 1927.
Brown, Harold I. Perception, Theory and Commitment. Chicago: The
University of Chicago Press, 1977.
Camap, Rudolf. "The Elimination of Metaphysics through the Logical
Analysis of Language," in Logical Positivism, ed. A .J. Ayer.
Glencoe, Illinois: The Free Press, 1959.
Connell, Richard J. Substance and Modern Science. Center for Thomistic
Studies, Houston, Texas. Notre Dame, Indiana: University of
Notre Dame Press, 1988.
DeTocqueville, Alexis. Democracy in America. Vintage Books. New
York: Random House, 1945.
Donn, William S. Meteorology (4th ed.). New York: McGraw-Hill Book
Company, 1975.
Duhem, Pierre. The Aim and Structure of Physical Theory. New York:
Atheneum, 1962.
Eddington, Sir Arthur. New Pathways in Science. Cambridge: At the
University Press, 1947.
Eliade, Mircea. Myth and Reality. Harper Torch book. New York: Harper
& Row, Publishers, 1963.
Feyerabend, P.K. Against Method. London: NLB, 1975.
Fodor, Jerry. "Observation Reconsidered," Philosophy of Science, 51,
(1984).
Galileo Galilei. /1 Saggitore. Reprinted as "Two Kinds ofProperties," in
Philosophy of Science, ed. Arthur Danto and Sidney
Morgenbesser. Meridan Books. New York: The World
Publishing Company, 1960.
Gordon, Cyrus H. "Canaanite Mythology," in Mythologies ofthe Ancient
World. Ed. Samuel Noah Kramer. Anchor Books. Garden City,
New York: Doubleday & Company, Inc., 1963.
Graves, Robert. The Greek Myths. Baltimore: Penguin Books, 1955.
Hanson, Norwood Russell. Patterns ofDiscovery. Cambridge University
Press, 1958.
Huygens, Christian. "The Wave Theory of Light," in Classics in Modern
Science. Ed. William S. Knickerbocker. Boston: Beacon Press,
1962.
Halliday, David and Robert Resnick. Physics. New York: John Wiley &
Sons, Inc., 1962.
Hamilton, Edith. Mythology. A Mentor Book. New York: The New
American Library,
Holton, Gerald and Duane Roller. Foundations of Modern Physics.
Reading, Massachusetts: Addison-Wesley Publishing Company,
Inc., 1958.
Jacobs, Noah Jonathan. Naming Day in Eden. New York: The Macmillan
Company, 1958.
Jammer, Max. Concepts of Force. A Harper Torchbook. New York:
Harper and Brothers, 1962.
Kant, Immanuel. Critique ofPure Reason. Tr. Norman Kemp Smith.
New York: St. Martin's Press.
Prolegomena to any Future Metaphysics. Library of Liberal
Arts. New York: Bobbs Merrill Company, Inc., 1950.
Kirk, G.S. The Nature ofGreekMyths. New York: Penguin Books, 1974.
Koyre, Alexander. Newtonian Studies. Cambridge, Massachusetts:
Harvard University Press, 1965.
Kuhn, Thomas. The Structure of Scientific Revolutions. Chicago: The
Bibliography 229
University of Chicago Press, 1960.

Lorenz, Konrad. On Aggression. New York: Harcourt Brace Jovanovich,
Inc., 1966.
Nagel, Ernest. The Structure of Science. New York: Harcourt, Brace &
World, Inc., 1961.
Newton-Smith, W.H. The Rationality ofScience. Boston: Routledge &
Kegan Paul, 1981.
Nozick, Robert. Philosophical Explanations. Cambridge,
Massachusetts: Harvard University Press, 1981.
Lemmon, E.J. Beginning Logic. Indianapolis: Hackett Publishing
Company, 1978.
Philips, O.M. The Heart ofthe Earth. San Francisco: Freeman, Cooper &
Company, 1968.
Piaget, Jean. The Child's Conception of the World. Trans. Joan and
Andrew Tomlinson. Totowa, New Jersey: Littlefield, Adams &
Co., 1965.
Pope, Alexander. The Poems of Alexander Pope. Ed. Maynard Mark.
London: Metheun & Co., Ltd., 1967.
Price, H.H. Thinking and Experience. Cambridge, Massachusetts:
Harvard University Press, 1953.
Quine, W.V.O. From a Logical Point of View. Cambridge,
Massachusetts: Harvard University Press, 1953.
"Posits and Reality," reprinted in Theories and Observation in
Science. Ed. Richard E. Grandy. Englewood Cliffs, N.J.:
Prentice-Hall, Inc., 1973.
Rodgers, Eric M. Physics for the Inquiring Mind. New Jersey: Princeton
University Press, 1960.
Rosenberg, Alexander. Microeconomic Laws. Pittsburgh, Pennsylvania:
University of Pittsburgh Press, 1976.
Rorty, Richard. Philosophy and the Mirror of Nature. Princeton:
Princeton University Press.
_Consequences of Pragmatism. Minneapolis, Minnesota:
University of Minnesota Press,
Shortly, George and Dudley Williams. Physics. New York: Prentice-Hall,
Inc., 1950.
Schlick, Moritz. Philosophy ofNature. Trans. Amethe von Zepplin. New
York: Philosophical Library, 1949.
Schumpeter, Joseph. Imperialism and Social Classes. Meridan Books.
New York: The World Publishing Company, 1955.
Spiegelberg, Herbert. The Phenomenological Movement. The Hague:
Martinus Nijhoff, 1960.

Stromberg, James S. "An Essay on Experimentum," in Laval Theologique
et Philosophique, Vol. XXIII, 1967, No.1; Vol. XXIV, 1968,
No.I.
Trollop, Anthony. Barchester Towers. New York: The New American
Library of World Literature, Inc., 1963.
Vander, Arthur J. and James K. Sherman and Dorothy S. Luciano.
Human Physiology. New York: McGraw-Hill Book Company,
1970.
Veatch, Henry B. Two Logics. Evanston: Northwestern University Press,
1969.
Wallace, William. "Are Elementary Particles Real," in From a Realist
Point of View. Lanham, Maryland. University Press of America,
1983.
Wartofsky, Mark W. Conceptual Foundations ofScientific Thought. New
York: The Macmillan Company, 1968.
Wilson, John. "The Origin of Life," in Did the Devil Make Darwin Do
It? Ames, Iowa: The Iowa State University Press, 1983.
Index
A Bernard, Claude, 130
Bossu, 91
a priori and intelligence, 118 bracketing existence, 56-57
astract and concrete, 105-109 Brahe and Kepler as seeing the
abstract terms, 72 same thing, 10 1
acceleration, 141-142 Bridgman, Percy, 36-37
active and passive, 115-116
activity implies an object, 50 c
affirming the conclusion, 231
Aristotle, 89 Carnap, Rudolf, 4, 12-13, 66-68,
sophistical argument, 173 70-71
theory of the cosmos, 149-150, cause,
180 as unobservable, 10-11
Aquinas, Thomas, 39, 151, 189 definition of, 11
analogies, common sensibles, 15
and unobservables, 34-38 contradiction as opposite, 22
and similarities, 36-42 contrary as opposite, 23
in genetics, 40 children and thinking, 43-44
in teaching of physics, 40 Cicero, 89
necessity of, 38-40 cognitive capacities,
analytic propositions , 208ff confusion of, 101-103
according to Kant, 209-210 cognitive powers, failure to
as identities, 210 distinguish, 103
the realist view, 217-220 contingency and language, 76-77
artefactual signs, 59-60 confused knowledge, 178-180
axioms, 121 Copernican Revolution, 147
Ayer, A. J., 13n3 cosmos, Aristotle's theory 149ff
Coulange, Fustel de, 88
B
D
Barrow, John D., iii
Beach, John D., 218n13 Dalton, 136-137
behavior, 148-150 deconstruction, 77
reflections on, 123-124 definition of artefacts, 200
Berkeley, George, 104 definitions, functions of, 196-198
relations to observables, 201 force, 106

DeKoninck, Charles, 185n 1 affecting us, 55
Descartes, Rene, ii, 115 as a metaphor, 139ff
development of mind, 177-181 as a phenomenon, 54
differences of degree, 152-154 as a T-terrn, 14-15
correlations of, 153 foreign cognates, 71-72
dissimilarity relation, see foundationalism, 5, 99
similarity free agent, 169
distinct knowledge, 178-171 function as activity, 46
Donn, William L., 205 fundamental relation, 21-24
Duhem, Pierre, 49-50 precise problem, 21
E G
Eddington, Sir Arthur, 15n7

effect, obscurity of, 141 Galileo, 109
elementary particles as a limit of generalist, 188
knowledge, 137 generic, or common 189-190
elementary particles as real, 137- God, divorced from observables,
138 13
Eliade, Mircea, 83, 86 Graves, Robert 87
emotion, signified naturally, 62-63 gravity, 140-143
empirical,
as something psychological, 2 H
English physicists and models, 49-
50 Hamilton, Edith, 83
ethology, 187-188 Hanson, Norwood Russell, 100ff
etymology, 73-75 Holton, Gerald, 146
experience, 127-128 Hume, David, ii, 2, 104
and problems, 128 Huygens, Christian, 35
as supplying premisses, 129 hypotheses, 133-135
experiment, 129-131 and predictions, 124
explanation, and theoretical entities, 136-
separated from argument, 15 8 138
existentential import, 208 as assumptions, 134-135
etymologies, 73 and modus tollens, 170-171
F I
fact, meanings of, 79-10 imaginative literature, 12

Feyerabend, P. K. 109-111 inflected language, advantage of,
fluid, two definitions, 197-198 76
first principles, common, 207
common principles, 193, 197 imitative art, 65-66
Fodor, Jerry, 5 imperfection not an error, 182
Index233 233
intelligence, Locke, John, 103-104

as correcting the senses, 109 logic, 170-173
as passive, 118 not a calculus, 170
as putting something into Logical Positivists, 4, 54
sensation, 103 logicians, see explanation
internal sensory impressions, Lorenz, Konrad, 153
concomitant use of, 51 Luciano, Dorothy S., 37-38n7
as illusory, 10
internal states, how defined, 45-47 M
interpretation, 107
confused with sensation, Ill Malinowski, B., 85
independent if a priori, 99-100 man as political, 168
as puting something into man, defined, 218
sensation, 99ff matter as particulate, 136
interpreting and inferring, 7-9 metaphor, 14
necessity of, 48-49
metaphysical approach to nature,
J 147
metaphysics, as nonsense, 12
Jammer, Max, 141-142 modus ponens, 171
judgments, modus tollens, 171
about unobservable entities, motion, new causal theory, 150-
107 154
mover and moved, 151
K Mumford, Lewis, 188
myth as metaphor, 81-82
Kant, Immanuel, 3, 209-210 and truth, 90-91
knowledge not empirical, 4 as early science, 83-84
Kepler, Johannes, 141-142 empirical basis and purpose,
Kirk, G. S., 84-87 82-87
Knickerbocker, William S., 35 classes of, 82
Koyre, Alexander, 147ff monolithic, 87
Kramer, Samuel Noah, 83-84 universal theories of, see Kirk
Kuhn, Thomas, 108-109 mythical deities not haphazard, 89
myths and truth, 90-91
L music as ordered sound, 107
language, obscurity of, i

contingency of, 76-77
language as a sign, 62-63 N
advantage of inflection, 76
laws, 124-126 Nagel, Ernest, 38n8
Lemmon, E. J., 170-171 names, sources of, 75
light, 3 5-3 7 order of, 74-75
Lindley, David, iii Natural Philosophy, 185ff
natural sciences, order of, 189-190 philosophers as believing, not

necessity, 212-213 knowing, iii-iv
necessity and real things, 28-29 philosophy,
negations, how known, 22-24 and explanation 166-170
Newton, Sir Isaac, see motion, and regularities, 159-162
new causal thesory and unobservables, 12
Newton-Smith, W. H., 13-15 distinction form science, 146-
nominal definition, see definitions 147
non-entity, incapable of affecting as empirical, 161
senses, 24 originally identical with
Nozick, Robert, 158n3 science, 145-146
separation from science, 157-
0 159
philosophy of science, 191
object and activity, 44-45 photon, 214
observable Piaget, Jean, 43-44, 47-49, 179-
defined, 15-17 180, 187, 188
as pointing to an physical objects as conceptual
unobservable, 21 entities, 2
second sense of, 29 physical possibility, 213
observable properties, 15 physical properties as not variable,
as reason for errors, 107 215
observables and ordinary physics as mathematical, 150-154
language, 29-30 planetary motion, 140
as intrinsically detectable, 14 Plato, 69-70
pragmatist view, 4 Pope, Alexander, 93
observation, possible, 212-213
as theory-laden, 99ff postmodemist, i
as theory-laden in secondary practical disciplines, 126
sense, 108 practical rules, 212
observation, priority of object, 44- predication, 171-172
45 premisses,
observation-term, and observable, supplied by experience, 129
107 Price, H. H., 60
occupational behavior, 166-167 primary qualities, 15
opposites, 22-23 principle, 70-71
order of natural movements, 190 a more geneic term, 33nl
ordinary man and experience, 2 principle of non-contradiction,
organization and lines, 10 1 211
privation as opposite, 22-23
p problem solving, 128
proper sensibles, 15,
particles, see elementary particles pseudo-concepts, 66
phenomenology, 53-57 psychic activities and objects, 46
Phillips, 0. M., 138n3
Jndex235 235
Q definition of, 61
imitate observables, 63-64
qualitative differences, 185-186 similarity relation, 33ff
quantitative approach to science, singular,
146-147 not object of science, 126
Quine, W. V. 0., 1-2, 47 sound, see wave
speech, modified vocal sound, 63
R Spiegelberg, Herbert, 53ff
spontaneous movements, 116
realist position, presupposed, ii state and activity, 46
recognition, primary and statue, Ill
secondary, 108 Stromberg, James S., 127n6
regularities, 124-126 subjective, 55
examination of examples, 220- substance, unobservable, 27-28,
221 105
regularities and experience, 124- known through observables,
126 27-28
Reichenbach, Hans, 115nl syllogism, 164-166
relations as unobservables, 26-27 systematic considerations
relations of predicability, 171 bear on regularities, 121
relative opposition, 23 systematic definitions, 197-199
Roller, Duane, 146n3 as explanatory, 198
Rogers, Eric M., 37n6 symbol
Rorty, Richard, i, 150-161 as distinct from sign, 60
Rosenberg, Alexander, 206 as likeness, 61
critique of, 220-221 implies human making, 61
s T
Schlick, Moritz, 51-52

Schumpeter, Joseph, 199 terms,
sciences distinguished by objects, formal relations of, 163-164
189 Thales, 145-146
secondary qualities, 15 theoretical entities, 136-137
secondary recognition, 108 real and fiactional, 140
seeing, as visual copy, I 02 Tipler, Frank J., iii
as an amalgam, 102 Tocqueville, Alexis de, 164
sensation , how defined, 46-4 7 Trollop, Anthony, 93-94
senses as active and passive, 16
twofoold passivity, 109-110 u
Sherman, James K., 37-38n7
Shortley, George, 197-198 unobservables,
signs, as involving negations, 39
arbitrary and natural, 61-62 how known, 21-22
as not purely arbitrary, 63 philosophy's preoccupation
with, 12
source of knowledge of, 11-13
v
Vander, Arthur J., 37-38
Veatch, Henry B., 170nl3
voice as a sign, 62
w
Wallace, William, 136n2
Wartofsky, Marx W., 4-5
wave, see light
whole and part, 217
Williams, Dudley, 197-198
words,
as pseudo-concepts, 66
cognates as obscure, 71
extensions of, 68-69
functions of, 195-196
kinds of existence signified,
162
many meanings of, 67-69
order of signification, 69-73
signify by convention, 63
signify observables in
relation
signify concepts, 62
to unobservables, 29
that from which and that
upon which, 74-75
..

Connell-From Observables To Unobservables PDF

Încărcat de

Informații document

Titlu original

Drepturi de autor

Formate disponibile

Partajați acest document

Partajați sau inserați document

Opțiuni de partajare

Vi se pare util acest document?

Este necorespunzător acest conținut?

Drepturi de autor:

Formate disponibile

Connell-From Observables To Unobservables PDF

Încărcat de

Drepturi de autor:

Formate disponibile

FROM 0BSERVABLES

University Press of America, Inc.

12 Hid's Copse Rd.

All rights reserved

Library of Congress Cataloging-in-Publication Data

ISBN 0-7618-1663-1 (cloth: alk. ppr.)

Chapter 1 The Problem 1

Chapter 2 Observables-What They Are 7

Chapter 3 The Fundamental Cognitive Relation 21

Chapter 4 The Similarity-Dissimilarity Relation 33

Chapter 5 More on Analogies and Metaphors 43

Chapter 6 Signs, Symbols and Words 59

Chapter 7 The Necessity of Myth 81

Chapter 8 Theory-Laden Observation 99

Chapter 9 The A Priori and Innate Ideas 115

Chapter 10 Regularities, Experience, Experiment 121

Chapter 11 Hypotheses and Reality 133

Chapter 12 The Mathematical Character of Physics 145

Chapter 13 General Considerations on Philosophy 157

Chapter 14 Philosophical Principles: Preliminaries 177

Chapter 15 Natural Philosophy 185

Chapter 17 Analytic Propositions 205

Chapter 18 Recapitulation 225

At this point we shall quote two different sources outside philosophy

Philosophers, these cosmologists think, are attached to their theories on

or see on the basis of evidence-what they hold. If that is true, then it is

shown about the empirical origins of concepts.

University of St. Thomas

I. Philosophy and the Mirror of Nature (Princeton University Press, 1979).

W. V. 0. Quine thinks himself an empiricist despite looking on physical

As an empiricist I continue to think of the conceptual scheme of science

as a tool, ultimately, for predicting future experience in the light of past

impressions. An object that cannot produce an impression cannot be

The peculiar characteristic of a science may consist of a simple difference

Unlike Hume, Kant held that we do have knowledge that is not

4 From Observables to Unobservables

who had some sound philosophical principles, nonetheless effectively

What is observable, in the pragmatist view, is what a community of

D. What will come

4 "Observation Reconsidered," Philosophy of Science, 51, 1984) p. 25.

B. The meaning of "fact"

Because every systematic discipline aims at understanding some

probably draw the conclusion "Therefore an opaque body is interposed

interpretation, both ofwhich involve more than a reporting of facts. The

reality, a shape, say, we become simultaneously aware not only of the

C. We do know some unobservables

There is no dpubt that systematic investigations (not to mention

ordinary conversation) concern themselves with many things that are

D. The source of our knowledge of unobservables

When we reflect, we readily see that the unobservables the sciences

which, indeed, do not have manlike bodies, yet manifest themselves

Currently the distinction between observable and unobservable

Those who have defended an 0/T distinction have sought to defend it as

According to Newton-Smith, there is no distinction in kind between

Newton-Smith tells us (with an appropriate reference) that according to

This characterization is entirely unenlightening unless it is supplemented

And that, he thinks, cannot be done. In his view, definitions of observable

then, need not concern us here.

magnitude of the measured, and the magnitude of the measure. 7 For

an object is actively determined by the animal, but the nature of its

18 From Observables to Unobservables

Our procedure in this chapter consisted in first pointing out certain

!.The conclusion-become-premiss is not an hypothesis, an issue we shall discuss