Documente Academic
Documente Profesional
Documente Cultură
The Philosophy
ofPhilip Kitcher
Editedby MarkCouch
and
Jessica Pfeifer
1
1
Oxford University Press is a department of the University of Oxford. It furthers
the Universitys objective of excellence in research, scholarship, and education
by publishing worldwide. Oxford is a registered trade mark of Oxford University
Press in the UK and certain other countries.
135798642
Printed by Sheridan, USA
CONTENTS
Contributorsvii
Introduction1
Mark Couch and Jessica Pfeifer
1. Kitcher against the Platonists 14
GideonRosen
Reply toRosen
2. Kitchers Two Design Stances 45
Karen Neander
Reply to Neander
3. Proximate and Ultimate Information in Biology 74
Paul E.Griffiths
Reply to Griffiths
4. Bringing Real Realism Back Home:APerspectival Slant 98
Michela Massimi
Reply to Massimi
5. Unificationism, Explanatory Internalism, and Autonomy 121
James Woodward
Reply to Woodward
6. Special-Science Autonomy and the Division of Labor 153
Michael Strevens
Reply to Strevens
7. Toward a Political Philosophy of Science 182
JohnDupr
Reply toDupr
8. Kitcher on Science, Democracy, and Human Flourishing 206
LorraineDaston
Reply toDaston
9. Deliberating Policy:Where Morals and Methods Mix 229
Nancy Cartwright and Alexandre Marcellesi
Reply to Cartwright and Marcellesi
10. Function and Truth in Ethics 253
MichaelSmith
Reply toSmith
11. What to Do While Religions Evolve before Our Very Eyes 273
Daniel Dennett
Reply to Dennett
References289
Index301
[vi]Contents
CONTRIBUTORS
1.
* For help with this volume we are grateful to a number of people, including Kyle
Stanford, Peter Godfrey-Smith, Laura Franklin-Hall, Jim Thomas, Silvena Milenkova,
Philip Kitcher, the contributors, Seton Hall University, University of Maryland,
Baltimore County, and discussions with several others. Aspecial thanks to our editors
at Oxford University Press, Peter Ohlin and Emily Sacharin, for their guidance dur-
ing this project, and Nisha Dayalan and Judith Hoover for their help in preparing the
manuscript. Philip Kitcher would also like to say thank you to the contributors for each
of their chapters.
This volume provides an examination of various areas of Kitchers phi-
losophy. In this introduction we provide some background for the reader by
describing a number of his major works and how his interests have devel-
oped over the years. Given the many works that have been written, there
wont be an attempt to cover everything.
A good place to begin is with the first book Kitcher published, which
was on the issue of scientific creationism. In the 1970s some of the cre-
ationists in the United States were becoming more open with their view
that the scientific evidence that existed did not undermine the creation
story in the Bible. This culminated in a number of books defending scien-
tific creationism. In 1982 Kitcher wrote Abusing Science:The Case against
Creationism in reply to the emergence of this view, arguing that the cre-
ationist authors were bending the science to support their positions. In
this work Kitcher carefully explains how to think about scientific evidence
and the content of evolutionary theory and other issues that were rel-
evant to the creationists approaches. He patiently describes their mis-
takes in trying to make the science appear to support their positions. He
explains that the problems with their approaches become apparent once
one is clear about what the science actually says and how to understand
it properly. Several of the ideas Kitcher touches on in this book he would
develop further at a latertime.
In other work during this period Kitcher turned his attention to issues in
the history and philosophy of mathematics (which was the area of his doc-
toral work). While the focus was different, this raised issues related to what
came before. In The Nature of Mathematical Knowledge (1983), Kitcher aimed
at understanding the historical development of mathematical knowledge.
His interest here is with understanding how mathematics developed as a
practice and how an appreciation of this informs our account of mathemat-
ics itself. He takes the view that mathematical knowledge is empirical in the
tradition of John Stuart Mill. We should understand mathematical practice
as depending on perceptual experiences in the origins of the subject and
developing from there through a sequence of rational transitions to more
complex parts of mathematics. Such an approach does not involve making
reference to abstract objects to explain mathematics. Characterized in this
way, mathematics takes its place alongside other empirical disciplines that
serve to improve our knowledge.
Around this time Kitcher also devoted several years to gaining a better
understanding of the sciences and their features. One area on which he
focused was the notion of scientific explanation. The proposal he offered
was an account of explanation that drew on earlier work in the area. The
distinctive feature of the account was to appeal to the notion of unification
Introduction [3]
to provide us with objective knowledge about the world. To explain this
Kitcher offers his readers improved accounts of such notions as explana-
tion and rationality as they occur in the practice of science. Once these are
understood, we are in a better position to understand the development of
scientific knowledge. We can recognize the critics concerns without having
to give up the view that science develops in a largely rational manner.
After this Kitcher continued to develop his work along these lines. While
he worked on other issues, he never strayed far from his interest in under-
standing the sciences in general. Several years had passed since the publica-
tion of The Advancement of Science before Kitchers next book on the subject
appeared, and in the intervening years he had come to think that some of
his earlier views needed to be revised. In a sequence of works he explains
what should be retained from this previous work and how it should be
modified to better reflect the practice of the sciences. In Science, Truth, and
Democracy (2001b), he makes a number of suggestions. Whereas previously
he had thought there was a single, overriding goal of the sciences, he came
to think this was misleading. The sciences develop in their individual ways,
and we should recognize a plurality of practical and epistemic aims that
exist since this view is more consistent with how the sciences have arisen
historically. In addition, Kitcher argues that more attention should be paid
to the issue of how the sciences fit into society, particularly its democratic
aspects. It is evident that science develops in relation to the broader society
of which it is a part. For this reason part of understanding science involves
understanding the role of science in society and how to think of the relation
between scientists interests in their research and the needs of the larger
society (e.g., think of how some scientists are interested in pure research
and how this may differ from citizens interests in solving particular medi-
cal conditions). The suggestions Kitcher made tried to be sensitive to the
interests of both and describe the proper role of science in society. The view
he presented he called well-ordered science.
The details of this view are further developed in Science in a Democratic
Society (2011b). Here Kitcher provides more specific accounts of what
a well-ordered science would look like under his conception, making his
suggestions more concrete and working out how they apply in the circum-
stances. There is a balance that needs to be struck between the expertise of
the scientist on factual questions relevant to public policy and the recogni-
tion that value-laden decisions about which policies to pursue should be
informed by input from the public. How this balance should be struck, and
the difficulties involved in it, are addressed with a number of proposals.
These works focus on understanding the role of science in society and
the value of scientific research. But the specific issue of value and its
Introduction [5]
important problems and repay careful reading. The second point relates to
the breadth of Kitchers knowledge, which becomes apparent as soon as one
considers the several areas on which hes worked. The range of his contri-
butions to traditional areas of philosophy, as well as other areas of broader
interest, is seldom found today among scholars in any field. We would sug-
gest that anyone who believes philosophy has become narrow in scope should
see in Kitcher an example of someone who has avoided this sort of parochial
perspective.
2.
The above review has described some of the central areas of Kitchers
large body of work. While it would be impossible to cover all the areas in
which Kitcher has made a significant contribution, we have tried to cre-
ate a volume that represents the breadth of his research. The contributors
have been asked to raise critical issues about different aspects of this work,
and Kitcher has been given the opportunity to reply. The remainder of this
introduction provides a summary of the chapters.
Gideon Rosens essay, Kitcher against the Platonists, is a critique of
Kitchers anti-platonism in mathematics that focuses on Kitchers (2012b)
more recent work. In both that work and his earlier work (1983), Kitcher
attempts to avoid platonism by arguing that the truth of mathemati-
cal claims does not require the existence of mathematical objects of any
sort. Kitcher (1983) defends a non-face-value semantics for mathematical
claims, arguing that the subject matter of mathematics is actually a hypo-
thetical collecting activity of idealized agents. Later (2012b) he defends
the view that mathematics is a collection of games and that mathematical
claims are not descriptions and dont have a subject matter of any sort. He
provides a novel defense of this formalism by arguing against any view that
attributes a subject matter to mathematics, including platonism. In par-
ticular he argues that platonists have no good explanation of how symbol
manipulation could lead to the discovery of new abstract objects. There is a
gap between the basis for mathematical claims and the ultimate standard
of correctness in mathematics on the platonist view. In response, Rosen
argues that a version of platonismmoderate platonismcan answer
Kitchers charge. Unlike Benacerraf (1973), Kitchers critique of platonism
does not rely on a general constraint on knowledge. Rosen argues that
without such a general constraint Kitchers conclusion does not follow for
a moderate platonist. The moderate platonist can accept that there is a gap
Introduction [7]
others have been engaged in developing. He distinguishes between proxi-
mate and ultimate information and describes new accounts of each that can
be used to help characterize gene-environment interaction in a way that
respects Kitchers causal democracy principle. He argues that we can com-
bine the insights of interventionist accounts of causation with Shannons
information theory to develop an account of proximate information in
terms of causal specificity. Moreover by revising Sheas (2013) account of
ultimate, teleological information, Griffiths defends a notion of biologi-
cal teleology that can figure in proximate explanations of development.
This also allows him to show how the teleological notion of information
might be more closely aligned with the proximate account he discusses.
Both notions of information are consistent with Kitchers principle since
both leave open which causal factorsgenetic or environmentalmight
be carriers of information. Hence rather than being a barrier to under-
standing gene-environment interaction, Griffiths argues, biological infor-
mation might prove useful in vindicating Kitchers argument that a correct
response to genetic determinism requires patient, empirical study of the
relative importance of various causal factors in development in a way that
respects the principle of causal democracy.
Michela Massimis Bringing Real Realism Back Home: A Perspectival
Slant is an attempt to rescue Kitchers (2001a) Real Realism from an inad-
equacy she believes it faces by bringing it back to Kitchers earlier Kantian
roots. Massimi focuses on Kitchers response to Laudans (1981) histori-
cal argument against realism, and specifically his use of the distinction
between working posits and idle wheels of theories. While she considers
Kitchers argument one of the most persuasive replies to such challenges,
there are historical cases wherein Kitchers approach seems inadequate. She
diagnoses the problem as resulting from his stringent notion of success.
She distinguishes success from above (which might be Nagels view from
nowhere or the real realists view from now) and her own preferred success
from within. She argues that our current vantage point is not privileged; it
is just one perspective among many. Rather than assess past theories from
our current perspective, we should assess them using their own standards
of success, but from other subsequent or rival perspectives, which include
the richer information such perspectives have at hand. She maintains that
false claims could not satisfy such a criterion of success. Her perspectival-
ism thereby provides the Real Realist with an alternative route to defend-
ing realism without privileging our own perspective. Where Kitcher relies
on our own perspective to pick out those parts of theories that are deemed
true from our own perspective, her perspectival realism identifies claims
that we have reason to believe are true, since they are justifiably retained
Introduction [9]
approach of compartmentalization involves plugging black boxes into
a systems inputs and outputs, while stratification involves black-boxing
lower-level phenomena and building a model of the system out of the black
boxes. Stratification makes clear how explanatory autonomy is compat-
ible with reductionism. While objective irrelevance can lead to functional
stratification, scientists often decide to black-box lower-level phenomena
that are objectively relevant as a way of efficiently dividing cognitive labor;
in such cases the lower-level phenomena are objectively relevant, but prac-
tical considerations about how to efficiently divide labor entail that they
are contextually irrelevant. Such practical considerations do not entail that
contextual irrelevance is merely pragmatic or observer relative but depends
on what Strevens calls functional difference-making. Hence the world
allows for functional stratification, which enables scientists to efficiently
divide cognitive labor. Explanatory autonomy is thereby preserved in a way
that is consistent with explanatory reductionism.
In Toward a Political Philosophy of Science, John Dupr directs his
attention to Kitchers notion of a well-ordered science. While he thinks
Kitchers goals are laudable, he is less sanguine about whether or to what
extent well-ordered science is achievable and skeptical that Kitchers pro-
posed methods for realizing it are the most fruitful. Dupr focuses on two
main issues:how we ought to decide which research to fund (or even allow
to be pursued) and how democratic decisions should be made about the
application of science to public policy. He argues that implicit in Kitchers
work is the idea that science, democracy, and ethics are all social technolo-
gies. What he considers especially enlightening in Kitchers work are the
ways that science and democracy can come into conflict, which Dupr
sees as especially problematic in the information age. However, Dupr is
skeptical that Kitchers proposed solutions to ill-ordered science are either
workable or helpful. He argues that it is unclear how Kitchers idealized
conversations can be harmonized with actual conversations. He also ques-
tions the relevance of such idealized conversations for addressing the dis-
cord between democracy and science, given that such discord is a problem
of social technology. In addition, while he thinks the citizen juries that
Kitcher recommends are perhaps successful in some cases, such juries are
often ill-suited to the task. Given the current social system we inhabit, more
systematic political changes are needed. Where Kitchers focus is primarily
on equality of voice, Dupr argues that well-ordered science is hampered by
the inequality of resources that our current social system promotes.
Lorraine Daston, in Kitcher on Science, Democracy, and Human
Flourishing, focuses on Kitchers attempt to reconcile science and democ-
racy and his use of history in defending his views. First, she questions
Introduction [11]
and think that causality is linear and God-given. Causal relations are far
more complex, while the objective relations we discover through RCTs are
local, surface-level, and expressible only in language specific to the RCTs.
Instead of using other types of investigation that would be a better guide
to causal structure and hence a better guide for policy decisions, we over-
generalize from a few objective RCTs without adequately addressing the
moral ramifications of doing so.
In Function and Truth in Ethics, Michael Smith raises concerns about
Kitchers (2011a) account of ethical truth as developed in The Ethical Project.
Kitcher builds ethical truth out of ethical progress. Ethical rules count as
true if they are retained as ethical codes progress. Smith argues that this
account of moral truth leads to problems once we realize that progress is
to be understood in terms of promoting ongoing cooperation. On Kitchers
account there is a gap between the ethical rules we need to adopt in order
for ethical practice to serve its functionwhich Smith argues Kitcher must
understand as promoting ongoing cooperationand the moral beliefs
many of us (including Kitcher) hold. Hence Kitchers views about the func-
tion of ethical practice, together with his pragmatic naturalist account of
the truth of ethical claims, entail that many of our ethical beliefs are false.
Moreover ongoing cooperation is sometimes aided in crucial ways by the
fact that such false beliefs (beliefs that are false by Kitchers lights) are
widely shared. Fortunately we can accept Kitchers account of the function
of ethical practice without adopting his account of ethical truth. Smith
considers two alternatives he maintains are preferable:noncognitivist and
Kantian accounts of ethical truth. He defends both of these possibilities
against Kitchers objections. Either would also allow us to disambiguate the
causal question of why we have adopted the rules we have and the justifi-
catory question of what rules we ought to adopt. We can thereby accept
Kitchers account of the function of ethical practice, while leaving open
what function ethical practice ought to serve and what moral beliefs we
can legitimately assert aretrue.
Daniel Dennetts essay, What to Do While Religions Evolve before Our
Very Eyes, focuses on Kitchers (2011c) essay Militant Modern Atheism,
in which he argues that the New Atheists fail to account for the positive
role religion can play in peoples lives. Consequently their militant athe-
ism is likely to be counterproductive in the end. Kitcher argues that it
is possible to maintain a religious life even in the face of criticisms the
modern atheists have effectively wielded, and for at least some people it
is beneficial to do so. He distinguishes between the belief model of reli-
gion and the orientation model, arguing that the orientation model opens
up such possibilities and more adequately accounts for the aspects of
Introduction [13]
CHAPTER1
Mathematics is replete with results that affirm (or seem to affirm) the exis-
tence of mathematical objects. For example:
1. There are several senses in which an object may be said to be abstract. This
usage follows a tradition deriving from Frege ([1918] 1984), but the objects of pure
There are exactly three ways to resist this argument. You can take the
eccentric view that numbers and the rest are (despite appearances) con-
crete entities (Forrest and Armstrong 1987). You can step back from ordi-
nary mathematics and hold that while the existence theorems may be good
mathematics, they are not true and so cannot serve as premises in a sound
argument (Field 1980). Or you can holdand this is trickierthat while
the existence theorems are true and so fit to serve as premises, their truth
does not require the existence of mathematical objects of anysort.
This last position is tricky for obvious reasons. It is a plain contradiction
tosay:
There are prime numbers greater than 15, but there are no numbers.2
mathematics are presumably abstract in every sense if they exist at all. See Rosen
(2014) for discussion of the terminologicalpoint.
2. Compare the closing sentence of Benacerraf (1965, 73):If truth be known, there
are no such things as numbers; which is not to say that there are not at least two prime
numbers between 15 and20.
3. One important feature of this argument for present purposes is that it is not a
semantic argument. It does not assume a Tarskian account of mathematical truth,
or any other such determinate account. The argument uses, but does not mention,
mathematical vocabulary. It thus puts pressure on any theorist who is happy to use
mathematical vocabulary in the usual ways, regardless of his or her semantic views. For
a more complete statement of the argument, see Rosen and Burgess (2005).
K i t c h e r a g a i n s t t h e P l at o n i s t s [15]
Kitcher has never been tempted to identify the objects of mathematics
with concrete things or to dismiss ordinary mathematics as a false but use-
ful fiction. His view has always been that settled mathematics is just fine as
it is, but that its claims, properly understood, do not concern a domain of
mathematical objects.
Kitchers 1983 book, The Nature of Mathematical Knowledge, defends a
version of anti-platonism according to which the subject matter of math-
ematics is not a domain of entities but rather the hypothetical collecting
activity of an idealized human agent. Like many anti-platonist strategies
from the period, this one works by constructing a nonface value seman-
tics for (part of) the language of mathematics. Such semantic theories yield
a mapping S S* from ordinary mathematical claims like There is a prime
number greater than 15 to claims in a modal language that do not seem to
require the actual existence of abstract entities:roughly, claims of the form
If the concrete world had been thus and so, then such and such would have been
the case. The mapping is designed to associate truths (falsehoods) in the
mathematical language with truths (falsehoods) in the modal language in
such a way as to preserve intuitive entailment relations among claims. But
more than this:the mapping is supposed to give the meaning of the original
mathematical claim and so to show it to be the sort of claim that does not
require the existence of mathematical objects for itstruth.
Since Kitcher has abandoned this approach Iwill not dwell on his par-
ticular version of it. But it is worth asking how views of this sort respond
to the quick argument for platonism sketched in the previous section. The
reductive nominalist cannot deny that there are prime numbers greater
than 15, since his view will map this mathematical claim onto a modal
claim he accepts by means of a semantic mapping that is designed to pre-
serve truth value. Nor can he deny that there are numbers, since the lat-
ter claim is a logical consequence of the first and his mapping is designed
to preserve logical relations.4 Instead he must sayand this is the tricky
bitthat while there are indeed infinitely many numbers of various sorts,
it is a kind of nonsense to ask whether these numbers are abstract or con-
crete, whether they exist in space, and so on. The paraphrase procedure that
gives meaning to statements in the language of mathematics associates
each ordinary mathematical statement with a definite (modal) content; but
4. These points are emphasized in Alston (1958). The neglect of this paper in the
literature on mathematical platonism is striking, especially in view of the fact that
Alstons paper was reprinted in the field-defining collection Benacerraf and Putnam
(1964). Astriking exception is Wright (1983), the first important work to emphasize
the significance of Alstons point for the metaphysics of mathematics.
5. To my knowledge the reductive nominalists have not made this point explicitly,
but Ibelieve it is the only way for them to evade Alstonspoint.
K i t c h e r a g a i n s t t h e P l at o n i s t s [17]
derivable in chess, that is, whether it can arise through play accord-
ing to the rules. But it makes no sense to ask whether a configuration of
chess pieces correctly represents its subject matter. According to the view
Kitcher now holds, mathematics is a collection of games for transform-
ing strings of nonrepresentational squiggles. These games are governed
by rules implicit in our practices and sometimes known explicitly to
mathematicians. We can make combinatorial, metamathematical state-
ments about these strings and the rules that govern them. These state-
ments constitute what Frege ([1903] 2013, 93) calls the theory of the
game, and for all Kitcher says, they may be genuinely representational.
But There are prime numbers greater than 15 is not a metamathemati-
cal statement about the game. It is a configuration of pieces within the
game. The rules governing the manipulation of these pieces may give the
string a kind of meaning. But the string and its parts do not stand for
anything, so it makes no sense to ask what the objects it describes are
like, or whether it describes them correctly.
How does this position block the quick argument for platonism
sketched at the outset? Kitcher does not say, but the answer must be
this: Whereas a string like 17 is a prime number may be derivable in
the game of arithmetic and hence assertible in a sense, metaphysical
statements like The number 17 is not in space or The number 17 is an
abstract object arelike
is green.
VARIETIES OFPLATONISM
Before we turn to the argument, Ishould say a word about its official tar-
get. As is well known, the view we have called platonism comes in two
flavors (Chihara 1973). Both hold that mathematics is concerned with a
domain of immaterial abstract objects. The hardcore platonists distinctive
claim is that these objects play something like a causal role in mathemati-
cal practice:that mathematicians are somehow aware of them or sensitive
to them, and hence that our mathematical beliefs are sometimes shaped
by the objects they represent. The moderate platonist denies this, insist-
ing that abstract objects do not impinge on us in any way. To put the con-
trast dramatically, the hardcore platonist holds that if the numbers had not
existed (per impossible, but so what?), the history of mathematics would
have been quite different, whereas the moderate holds that it might have
unfolded just as itdid.6
I mention this familiar contrast because Kitcher often writes as if hard-
core platonism were the only form of platonism on offer. Thus after an ele-
gant review of the history that led to the acceptance of imaginary numbers,
K i t c h e r a g a i n s t t h e P l at o n i s t s [19]
a history in which certain early figures (Cardano and Bombelli) stumbled
on the complex roots of cubic equations only to dismiss the new numbers
as subtile and useless, Kitcher (2012b, 182)writes:
More generally Kitcher supposes that any platonist must hold that when
the domain of mathematics is extended, new objects are discovered
or worse, detectedin roughly the sense in which Mendel discovered/
detected genes. Thus after entertaining the view that the sort of symbol
manipulation that led to complex arithmetic simply counts as a way of
detecting new abstract objects, Kitcher responds:
In none of these instances do we have any serious account of how the symbolic
manipulations serve as a way of detecting the alleged abstract entities. In the
Mendelian case, its possible to provide a positive causal explanation for why
the detection via pea plants works. Mendel himself saw part of this, and thats
why he could take his observations of the pea plants to be ways of detecting
underlying factors. He could justifiably use his instrument because he had
an account of how the phenomena he was trying to detect were related to the
properties he was able to observe.... Imagine properly educated counterparts
of Bombelli, Euler, Hamilton and Lagrange who fully subscribe to the Platonic
wisdom. Like Mendel, they would surely reflect on how their instruments, in
this case their symbolic practices, enable them to detect the underlying entities,
and platonic wisdom would supply them with no answer. Thus if they had what
is supposed to be the correct philosophical view of the matter, they would not
have been able to proceed as they did. (18485)
I will return to this passage later. The point to emphasize for now is that this
talk of detection makes sense only if the view under discussion is hard-
core platonism. Moderates deny that mathematical objects are detected, on
the ground that detection is a causal process, and so reject the demand to
say how the complex numbers were detected.
Of course the moderate platonist does believe in mathematical discovery
in a bland sense. To discover a fact in the bland sense is simply to come to
7. Notably Quine 1961; Putnam 1971; Wright 1983; Burgess 1983. Burgess and Rosen
(1997) is a full dress defense of moderate platonism (there called antinominalism).
K i t c h e r a g a i n s t t h e P l at o n i s t s [21]
of these new symbols and about the ultimate fruitfulness of their introduc-
tion. These doubts were fully allayed by the beginning of the nineteenth
century as the fruitfulness of the new apparatus was placed beyond doubt
and the now standard interpretation of complex arithmetical operations
as operations on points in the Argand plane guaranteed the consistency
(relative to established analytic geometry) of the game involving them.
Kitchers key claim is that this standard history is satisfactory, not just in
the sense that it tells us who did what and why but in the further sense that
it vindicates the introduction of complex analysis, showing it to have been
warranted by every pertinent epistemological or methodological standard.
With this paradigm of a satisfactory historical account in place, Kitcher
(2012b, 17273) presents four claims about mathematical discovery that
any platonist should be tempted to accept:
These claims are obviously inconsistent, so the platonist must reject one
of them. Kitchers argument is that unlike the formalist, who can happily
reject (A), the platonist has no good options.
Moderate platonists should accept (A), provided discovery is under-
stood in the bland sense given earlier. Mathematical innovation may not
involve the detection of new objects. But if to discover a thing is just to
come to know for the first time that it exists (or that it has the features
that make it interesting), then of course mathematicians discover new
abstract objects from time to time. Needless to say, the discovery will often
be spread out over time and over many people. (Asking Who discovered
zero? is like asking Who discovered the stone ax?) But if there was a time
when no one knew anything about X and a later time at which important
facts about X are clearly known, then we can infer that X was discovered
along theway.
(B) is likewise unexceptionable, properly understood. The metaphor of
access is of course misleading, since access to a thing normally involves
spatial or causal contact, and the moderate platonist will deny that we have
When the claim is put like this, the platonist must accept it. But he will
instantly add that it is somewhat unclear what it means to offer an account
of how people first came to know that X exists. Any complete account will
have two aspects:there will be an account of how people first came to believe
that X existsan historical account of the various psychological and social
processes that led to the formation of this beliefand then there will be
an epistemological gloss on this account designed to show that the belief in
question, formed in this way, was not a mere opinion but a case of knowl-
edge (or reasonable belief). Now in practice any account of this sort will
leave a great deal unsaid. The textbook account of Le Verriers discovery of
Neptune is in one sense satisfactory as it stands. It lays out Le Verriers evi-
dence in such a way as to make it clear that one might come to know about
the existence of a planet on such a basis. But it does not include a worked-
out theory of evidence, or of inference to the best explanation, or what have
you. Ahistorical account of this sort will make it clear that Le Verriers evi-
dence was in fact good evidence. But it need not answer the philosophical
question Why is the evidence that led Le Verrier to posit Neptune good
evidence for the existence of a thing like Neptune? We could insist that a
satisfactory account of an episode of discovery always include a philosophi-
cal account of why the discovery counts as a discovery. But if we do, then
Kitchers (D)will be clearly false. We should therefore understand a satis-
factory account of a discovery as one that presents the grounds on which
the new existential belief was formed in such a way as to make it clear to
the sensitive reader that these grounds were in fact good groundsideally
grounds on the basis of which the proposition in question might have been
known. And if we understand the word in this way, (B*) is unassailable.
Passing over (C)for a moment, we can agree that given this last point,
(D)is likewise unassailable. Standard historical accounts of mathematical
discovery tell us in more or less detail how mathematicians came to know
the existence theorems, which is to say that they tells us how mathemati-
cians came to believe these claims in such a way as to make it clear that these
beliefs were reasonable. Suppose a mathematician discovers a new solution
to the field equations of General Relativity by writing down an expression
for the solution and then proving that it is in fact a solution. The historian
can tell us in more or less detail how this happened, reconstructing both
K i t c h e r a g a i n s t t h e P l at o n i s t s [23]
the context of discovery and the proof itself, filling in the gaps, and so on.
But qua historian she will not broach the question Why is mathematical
proof of this sort a source of knowledge? That doesnt mean that the ques-
tion does not arise. The point is simply that ordinary historical accounts
presuppose that proof is a source of knowledge, leaving it for philosophers
and foundationally minded mathematicians to explore this presupposi-
tion. Asatisfactory historical account of discovery need not include a com-
plete philosophy of mathematics, though of course it must be consistent
with some such account if its explanations are to be correct.
Now Kitchers examplethe discovery of the complex numbersis
much more interesting, since this discovery was not a matter of proving a
theorem in an existing framework but rather of devising a new framework
(or extending an old one) with new rules of proof and calculation. The pla-
tonist must indeed regard this as a discovery in the undemanding sense. In
1500 no one knew the key existence theorems of complex arithmetic. By
1850 they were common knowledge. So over the course of this period the
relevant objects must have been discovered. Kitchers historical account
lays out the key mathematical developments in such a way as to make it
clear to the sensitive reader that mathematicians were reasonable in pro-
ceeding as they did. The account is therefore satisfactory in the sense
outlined earlier. Of course it is also incomplete. Not only does it leave out
details in the historical sequence, as is inevitable; it does not even begin to
address the epistemological question Why was it reasonable for mathema-
ticians to proceed as they did? This is a philosophical question, and Kitcher
himself clearly believes that it needs an answer that the historical account
itself cannot supply. After all, Kitchers formalist philosophy of mathemat-
ics is designed to supplement the historical account at just this point. So
Kitcher cannot object if the platonist must also provide a supplement:an
account of why the historical processes that led mathematicians to believe,
for example, that every polynomial with complex coefficients has a com-
plex solution counts as a way of coming to know this proposition.
The platonist must therefore reject (C), as Kitcher anticipates. Of course
if we rewrite (B)as (B*), we should rewrite (C)as(C*):
REFRAMING THECHALLENGE
Of course there is still a perfectly good philosophical question for the pla-
tonist:Why is that sort of history a way of coming to know (or reasonably
believe) a proposition about abstract objects? This is analogous to the ques-
tion one might ask after absorbing the standard historical account of the
discovery of Neptune:Why does that evidence count as evidence for the
existence of a distant planet? It is no knock against the historical accounts
K i t c h e r a g a i n s t t h e P l at o n i s t s [25]
that they do not answer this question. But of course it must have an answer
if the platonists rejection of (C*) is to be tenable.
The dialectic here is subtle, so let me clarify the point. Suppose the pla-
tonist concedes that at this point he does not have a satisfactory philosoph-
ical account of why the considerations that led mathematicians to accept
the existence of complex numbers count as a route to knowledge of abstract
objects. To concede this is simply to concede that he has not solved every
philosophical problem that his view raises, and that by itself is no objection
to his view. The platonist may be within his rights to say, in a Moorean vein,
Mathematicians certainly do know that x3 + 1=0 has three complex solu-
tions. And since the only thing to think about these solutions is that they
are abstract entities, their grounds for accepting this claim must amount
to a way of knowing about abstract objects. Since Ilack a general theory of
knowledge, Icannot explain why these grounds suffice for knowledge. But
they clearly do, and that is enough to answer Kitchers challenge. On the
other hand, if it could somehow be shown that there can be no adequate
platonist answer to this question, then the platonists Moorean response
would fail and Kitchers challenge wouldstand.
This is of course the standard gambit on the anti-platonist horn of Paul
Benacerrafs (1973) famous dilemma for theories of mathematical truth.
But to sustain this argument it is not enough to ask the platonist for his
epistemology. One must wield an epistemology of ones own to show
that if the subject matter of mathematics is as the platonist takes it to
be, then the usual ways of coming to believe a mathematical proposition
could never amount to ways of knowing. This is how Benacerraf himself
proceeds, invoking the causal theory of knowledge, and this is how Hartry
Field (1989) proceeds, invoking an explanatory constraint on knowledge.
Kitchers argument resonates with Benacerrafs, as Kitcher notes, but it is
distinct on precisely this point. Kitcher does not wield a general constraint
on knowledge against the platonist. If my analysis of the argument is cor-
rect, however, some such constraint is needed at just thispoint.
Moderate platonists will differ in their positive accounts of why the
grounds cited by the historian (and the mathematician) constitute adequate
grounds for believing the existence theorems, understood as claims about
abstract entities, but let me sketch the response Iprefer. The key thought is
that the ordinary internal justification for a mathematical claimwhether
it amounts to proof in an existing system or the sort of informal argu-
ment that justifies the acceptance of a novel axiomalways amounts to
adequate justification for the claim in question, the sort of justification
that suffices for knowledge when all goes well. To ask why the mathemati-
cal case for p is a good case for p is like asking why the inductive case for
K i t c h e r a g a i n s t t h e P l at o n i s t s [27]
would surely reflect on how their instruments, in this case, their symbolic
practices, enable them to detect [better:discover] the underlying entities, and
platonistic wisdom would supply them no answer. Thus if they had what is sup-
posed to be the correct philosophical view of the matter, they would not have
been able to proceed as they did. They would have had to hesitate, as Bombelli
did, although for considerably different reasons.
Looking past the rhetoric of detection, we may put the point as follows.
According to the platonistmoderate or otherwisethe truth of a math-
ematical claim is one thing; its acceptability by mathematical standards is
another. Mathematicians who took this view would therefore presumably
want some sort of proofor at least reason to believethat acceptability
by mathematical standards is a guide to truth. They would want a sound-
ness proof, in effect, not just for the system they presently accept but also
for the informal procedures that lead them to modify that system, as in the
case at hand. But platonism provides no such proof. This shows, first, that
these mathematicians were not (consistent, thoughtful) platonists, since
they saw no reason to demand proof of the soundness of their procedures.
But it also shows, more importantly, that platonism cant be the right phi-
losophy of mathematics since it would force us to represent this episode
in the history of mathematics as a matter of mathematicians blundering
around as sleepwalkers without adequate justification (Kitcher 2012b,
169), when in fact it is one of the great success stories in the history of the
subject.
My moderate platonist answers that since intramathematical justifica-
tion is justification enough, the heroes in Kitchers story were fully justified
in proceeding as they did (insofar as their procedures were mathematically
unexceptionable), and would have proceeded just as they did even if they
had been apprised of the platonistic wisdom. Just as astronomers can pro-
ceed by means of epistemic rules that they cannot justify independently, so
mathematicians can proceed by means of whatever principles guide their
practice without any independent proof of soundness. This does not mean
that astronomy does not aim at (transcendent) truth; the same goes for
mathematics, mutatis mutandis.8
and the rest thus cannot fail:if the posited structures are consistent, they exist. This
principle is notoriously hard to formulate; this is an important unsolved problem in
the philosophy of mathematics. (See Maddy 1997 and Balaguer 1998 for discussion.)
A fully adequate platonist response to Kitchers challenge would include a formula-
tion of this principle and then the claim that this principle stands to mathematics
as the metaphysical presuppositions of perceptual knowledge, inference to the best
explanation, and the rest stand to natural science. If settled science simply takes it
for granted, without independent justification, that the external world exists, that the
laws of nature are as simple as they can be given the phenomena, and so on, then it is
rationally permissible for natural scientists to take these things for granted. The mod-
erate platonists view is that the same goes for the principle of plenitude presupposed
by the method of free postulation in mathematics.
9. We might conceive mathematics as modal logic, or the study of structures, or
as the idealized science of human operations. Some of these conceptions dont make
much headway with respect to Benacerrafs original dilemma but none of them has
any advantages over Platonism with respect to the challenge of fitting the ontology
to the evolution of mathematical practice. For all of them say too much, introduc-
ing shadow entities behind the languages with which the mathematicians play their
increasingly sophisticated games (Kitcher 2012b,185).
K i t c h e r a g a i n s t t h e P l at o n i s t s [29]
Kitchers (2012b, 186)positive view is that mathematics is a game for
the manipulation of meaningless symbolsor better, a family of evolving
and overlappinggames:
This is a version of what Michael Resnik (1980, ch. 2)calls game formal-
ism. Kitcher cites Wittgenstein as a predecessor, though of course the view
isolder:
For the formal conception, arithmetic is a game with signs, which one may well
call empty, thereby conveying that (in the calculating game) they do not have
any content except that which is attributed to them by their behavior with
respect to certain combinatorial rules (game rules). Achess player makes use
of his pieces in a similar fashion; he attributes certain properties to them that
constrain their behavior in the game, and the pieces are only external signs of
this behavior. (Thomae 1898, quoted in Frege [1903] 2013, v.2,9798)
K i t c h e r a g a i n s t t h e P l at o n i s t s [31]
governing truth as we ordinarily conceive it that a statement is true if and
only if things are as it says they are. But according to Kitchers formalist
approach, a mathematical statement does not say anything. We do not call
a configuration of chess pieces true when the rules permit its derivation
from the starting position. More generally we do not call a move in a rule-
governed practice true simply because it is permitted by the rules. Better
to say, as Kitcher is clearly tempted to say, that like configurations in chess,
mathematical strings cannot be true or false. They can be derivable or not, and
that isthat.
The main argument for this sort of formalism is that it provides the basis
for a fully vindicatory reading of the history of mathematics in which epi-
sodes of radical innovation, like the introduction of complex numbers (or,
as Kitcher might have said, Cantorian sets) do not involve a bold leap into
the beyond but rather simply the introduction of new games constrained
by norms of consistency and fruitfulness. Even those of us who resist for-
malism should feel the force of this consideration. Occasional dissenting
voices to the contrary notwithstanding, it is a defining feature of modern
mathematical practice that mathematicians are free to innovate as they
likeintroducing new assumptions, objects, and structures, constrained
only by (hard-to-articulate) norms of consistency and fruitfulness. Any
philosophy of mathematics that cannot ratify this Cantorian idea is false
to the best modern self-understanding of the subject. Kitchers formalism
is expressly designed to satisfy this desideratum.
I have sketched a version of platonism that is also (I think) well placed to
satisfy it, so Ideny that this argument favors formalism. But even if formal-
ism were better placed to accommodate the fact of Cantorian freedom, this
would not settle the matter. It remains to consider the objections to this
sort of formalism, since they may tip the scales in the opposite direction.
One familiar objection comes from Gdels theorems. Let G be a Gdel
sentence for (say) ZFC of the sort that figures in the proof of the first
incompleteness theorem. Afamiliar sort of reflection persuades us that G
must be true (if ZFC is consistent, as we think it is). But G is not deriv-
able in ZFC or in any other accepted theory, so truth and derivability must
diverge.
Kitchers (2012b, 191)response is that the sequence of worthwhile sys-
tems proceeds indefinitely:
One of the directions in which it can extend consists in the addition to any for-
mal system adequate to arithmetic of the pertinent Gdel sentence, to yield a
new formal system for which the same extension can be carried out . . .. We
learn from Gdel that there will be no first order system adequate for the whole
K i t c h e r a g a i n s t t h e P l at o n i s t s [33]
a question mark (if only a very faint one in this case). Unlike the axiom of
infinity or the axiom of choice, Con(ZFC) is not an acceptable, fully detach-
able resource for proving theorems. So if mathematical truth simply con-
sists in derivability in a system that is fully acceptable for this purpose,
Kitchers formalist must say that Con(ZFC) is not true. Of course this is not
to say that Con(ZFC) is false. The formalist can say instead that Con(ZFC)
is like the continuum hypothesis as he presumably understands it:neither
provable nor refutable in any fully acceptable mathematical system, and
hence, by the formalists lights, neither true norfalse.
It would be disastrous if the formalist were forced by this admission
to the metamathematical conclusion that there is no fact of the matter
whether ZFC is consistent. But Kitcher is not quite forced to this conclu-
sion. The metamathematical, combinatorial claim that there is no proof of
0=1 in ZFC is, for all Kitcher says, a fully contentful claim whose truth
does not consist in its derivability. That claim might therefore be true, in
the ordinary sense, even if the number-theoretic statement Con(ZFC) is
neither derivable nor refutable, hence neither true nor false in the only
sense appropriate to it. The difficulty is that this severs the link, essential
to foundational research in mathematics, between metamathematical
claims of consistency and derivability, on the one hand, and the ground-
level mathematical claims that we normally take to formalize or code
them. This part of mathematics is predicated on the assumption that
we can convert modal or combinatorial claims about the consistency of
formal systems into mathematical claimsclaims about the existence of
models or about the existence of (numbers coding) formal derivations.
The objection is that Kitcher-style formalism would call this aspect of
mathematical practice into question. Understood as a modal/combinato-
rial question, the consistency of ZFC appears to be a factual question with
an answeralbeit a question we cannot answer within established math-
ematics. By contrast the mathematical question whether there exists a
number that codes a proof of 0 = 1 in ZFC, or the question whether
there exists a model of the ZFC axioms, must be understood as a question
that has no answer at all, since the candidate answers are underivable in
every authorized game and hence untrue in the only sense pertinent to
such claims.
Be all this as it may, the main source of resistance to formalism does
not come from these somewhat recherch considerations but from a
more basic source. Mathematics presents itself as a domain of inquiry.
As we normally think, mathematicians raise questions, entertain con-
jectures, make assumptions, engage in reasoning, and often come to
K i t c h e r a g a i n s t t h e P l at o n i s t s [35]
platonist by contrast can take these aspects of mathematical practice at
face value, and this must certainly be reckoned a clear advantage for the
platonist.
This brings us to the last and most important objection to formalism,
first pressed by Frege ([1903] 2013, 91) and widely regarded as deci-
sive.10 Suppose Iwant to know how many socks to buy and reason as fol-
lows:Each kid needs six pairs of socks, and Ive got 12 kids. 6 12=72.
So Ineed 72 pairs of socks. This line of reasoning is clearly cogent, and we
could easily represent it as a valid argument if we wished. The standard
(Fregean) reconstruction would involve both mathematical and non-
mathematical premises, together with bridge principles like The num-
ber of socks=n iff there are n socks, where the left-hand side involves
a numerical singular term and the statement on the right involves the
(readily analyzable) adjectival use of a number word. According to the
platonist, this argument supports its conclusion in the familiar way, that
is, by showing that it is impossible for the premises to be true and the
conclusion false. For the formalist, by contrast, the mathematical prem-
ises arelike
strictly incapable of ordinary truth, and the bridge principles are hybrids
involving meaningful English expressions on the right and meaningless
squiggles on the left. An argument of that sort cannot support its conclu-
sion in the familiarway.11
Of course the formalist can try to recast the reasoning as involving a
truth-evaluable metamathematical premise at the crucial point:
mathematics, and so their truth (or acceptability) cannot consist simply in their deriv-
ability in some formal game. Rather the norms governing their use must derive in
part from these formal rules and in part from the independently given meanings of
the nonmathematical vocabulary they contain. To my knowledge, formalists have had
nothing to say about how these two constraints might interact.
K i t c h e r a g a i n s t t h e P l at o n i s t s [37]
formalist wisdom, treating meaningless strings of squiggles as if they were
premises in arguments for factual conclusions:
Like Mendel, they would surely reflect on how their instruments, in this case,
their symbolic practices, enable them to arrive at knowledge of these ordinary
matters of fact, and formalist wisdom would supply them no answer. Thus if
they had what is supposed to be the correct philosophical view of the matter,
they would not have been able to proceed as they did. They would have had to
hesitate, as Bombelli did, although for considerably different reasons.
Reply toRosen
PHILIP KI TCHER
K i t c h e r a g a i n s t t h e P l at o n i s t s [39]
why the pragmatic considerations about fruitfulness, so prominent in the
history of the acceptance of complex numbers, are relevant. Moreover,
how exactly do the symbol-manipulating processes of the mathematicians
involved provide information about an abstract realm?
Lets start with a concrete example of the sort that plainly moved
Benacerraf. You buy an expensive present for someone dear to you. You go
to great efforts to ensure that it is well hidden. Yet before you spring the
great surprise the dear one offers a compelling demonstration of advance
knowledge, a description of the present both richly detailed and accurate.
When you ask how this knowledge has been obtained, you receive an aston-
ishing response:I sat down and thought very hard about what you might
give me, exploring the possibilities on paper. They led me to the description
Ive just given you. Isuspect that, were this to happen, you would not be
convinced by the explanation. Youd suppose that your best-laid schemes of
concealment had gone (as Burns puts it) agley.
We dont believe that people can obtain knowledge about physical
objects unless (to use Benacerrafs phrase) the space-time worms of
knower and object intersect (or, more exactly, come close to intersecting).
Thats what motivates the causal constraint. Moderate platonists will not
be moved by this, seeing it as a crude and misleading analogy. Yet a puzzle
remains: there seems to be a gap between the mathematician (Bombelli,
Euler, Hamilton) at the writing desk and the realm of abstract objects. How
do the scribblings performed at the desk, or the thoughts that generate
them, yield knowledge of the abstract realm? How do the cognitive rela-
tions change during the process of discovery?
Really modest platonists might try to close the gap by force. The really
modest platonist declares that all there is to being an abstract object is
to be susceptible to being known through various intellectual activi-
ties (expressed, perhaps, in experimental scribblings). I strongly suspect
that any attempt to provide a semantics for talk of abstract objects that
would articulate the declaration would lead to difficulties parallel to those
my critics discerned in my own earlier efforts (Kitcher 1983; see Burgess
and Rosen 1997, 2023). But in any case, the only residual issue between
really modest platonism and my position would be a dispute over preferred
modes of speech.
Rosens preferred moderate platonism attempts to do more than this.
He sees the task as one of explaining why the grounds cited by the his-
torian (and the mathematician) constitute adequate grounds for believ-
ing the existence theorems, understood as claims about abstract entities
and takes this to be part of the practice of mathematics in just the way
that using inductive generalization is part of the practice of the natural
K i t c h e r a g a i n s t t h e P l at o n i s t s [41]
you had better have a sample of instances that clearly have that property,
from which you can generalize.
As I noted, a part of my view on which Rosen doesnt comment (2012b,
18385) explores various possibilities for the moderate platonist. I dont
see how Rosens preferred epistemology improves on these.
But what of my rival account, seen by Rosen as game formalism? He
is entirely correct to suppose that game formalism dates back to the nine-
teenth century and that it provides one way to elaborate the approach I
favor. Yet there was reason to invoke Wittgenstein (rather than Thomae).
For Wittgensteins later writings contain three themes from which I draw
inspiration. One, explicit in my 2012b paper, is the emphasis on language
games directed toward different functions (not in the description busi-
ness). A second, related concern rejects the idea of some problematic depth
behind the surface phenomena: Wittgenstein excoriates the idea of some
hard reality behind mathematical statements, seeing that reality as idle
(as, I think, it becomes in Rosens moderate platonism). The third, of which
I made little use in my paper, is to separate the idea of meaningfulness
from correspondence-truth conditions in favor of a connection with use.
Taking that third theme seriously, I reject the inference from the premise
that mathematical terms lack reference to the conclusion that mathemati-
cal statements lack meaning.
An alternative way to articulate my approachmathematical pragmatism
is to start from the idea that the meanings of mathematical terms are con-
stituted by their conditions of use, both within mathematics and in the
uses to which mathematical vocabulary is put in scientific inquiry. (More
on Rosens penetrating questions about this practice later.) The history
of mathematics is a process in which mathematicians not only prove new
theorems using established language but also periodically extend their
languages. Mathematical progress consists in finding answers justifiable
by the settled rules to questions recognized as significant (and we should
allow for progressive revisions of the rules and for progressive reworking
of the notion of significance). The mathematical truths are the stable ele-
ments that emerge as we continue, indefinitely, to make progress. (I obvi-
ously echo thoughts of Peirce and James.)
Mathematics obtains its initial content from the ways arithmetic and
geometric vocabulary is used in operations of collecting, combining, and
comparing. Here my old Millian account (1983, ch. 6) can be refined to
explain the content of elementary mathematics. The history of math-
ematics reveals how further expressions become connected with those
introduced earlier, with new patterns of usage fixing the content of mathe-
matical language. As some people who have responded to my proposals have
K i t c h e r a g a i n s t t h e P l at o n i s t s [43]
points out, mathematical statements are often deployed in conjunction
with other statements in reasoning about many aspects of nature. Any
account that either denies truth (formalism) or understands truth differ-
ently (pragmatism) must explain how reasoning goes when the two types
of statementsmix.
For elementary arithmetic and geometry this is relatively easily done,
and my preferred way of doing it is along the lines of my reconstruction
of Mill (Kitcher 1980, 1983, ch. 6). For the higher parts of mathematics
I would adapt a classic logical empiricist idea: the symbols of the math-
ematical systems are given concrete applications in context, by linking
them to physical objects, properties, and operations on those objects and
properties. Newton connects the language of the calculus to velocities
and distances, conceived as measured through performing various opera-
tions using clocks and rulers. The mathematical systems are interpreted
anew in the contexts of various types of inquiry, so that applied (or
mixed, to use an old term) mathematical statements are candidates for
correspondencetruth.
To leave matters there is only to point in a direction along which an
answer to Rosens criticism might be found. Actually to give that answer
would require another historical enterprise. What is required is to show
how mathematics can be applied in measuring the world in a succession of
reinterpretations of the available formalisms, where schemes of measure-
ment introduced at earlier stages are taken for granted as further steps are
taken, so that we move from the simple (Millian) ways in which elemen-
tary mathematics is used to the most sophisticated applications of com-
plex analysis, tensor calculus and linear algebra in contemporary physical
theory. Some of my own early work in the history of mathematics (Kitcher
1973)inspires me to think this is not impossible (at least not for the intro-
duction of the calculus). But, to the best of my knowledge, nobody has
yet offered a detailed history of measurement practices that would settle
theissue.
Metamathematics should be viewed from the same perspective, as the
construction of a system for applying mathematical language to processes
that occur in the practice of mathematics itself (notably, the production of
proofs). Iwould hope that the historical approach to measurement would
enable us to see clearly just how a formalist/pragmatist can reconstruct the
reasoning mathematicians engage in when they reflect on the incomplete-
ness theorems, and thus address the questions Rosen poses. But for the
present Ican offer only this outline of a way of responding to some of his
doubts rather than providing fully developed answers.
P erhaps everyone who has thought much about the issue is now a func-
tion pluralist in some form or other. However, the interesting and sub-
stantial controversy concerns the nature and theoretical role of each notion
of function. Philip Kitchers (1993c) paper Function and Design develops
a pluralist view that in some ways is close to my own, but in other ways it
remains an influential challenge to the view that Idefend. Iwelcome this
opportunity to clarify how Kitchers theory differs from the theory Ifavor
and to explain why his argument for the form of function pluralism that he
supports is at best inconclusive, even though it raises important challenges
that need to bemet.
Kitcher (1993c, 379) begins his paper on functions by telling us that
there is some unity of conception that spans attributions of function
across diverse contexts. These contexts include those in which functions
are attributed to artifacts in everyday life, as well as to parts or processes
in organisms in creationist as well as contemporary biology. This unity, he
writes, is founded on the notion that the function of an entity S is what S
is designed to do (379). This might suggest that Kitchers analysis simply is
an etiological analysis.
Most etiological analyses tell us that an entitys function depends on
its history of selection. As Kitcher (1993c, 379) says, when we attribute
the function of releasing the metal bar to the mousetraps lever, we do
so because we believe that it was put there to do just that. And when
Harvey attributed the function of circulating blood to the heart, he prob-
ably believed that the wise and beneficent designer foresaw the need for
a circulation of blood and assigned to the heart the job of pumping (380).
As Kitcher further comments, it is also generally recognized that both
intentional selection and natural selection can ground functions.
However, Kitcher contrasts his analysis of functionsneedless to say
favorablywith etiological analyses. The main difference is that, on his
analysis, the links to intentions and to selection can be more or less direct
(1993c, 380). He is, it emerges, supporting a pluralist proposal in which
more than one notion of function is employed in attributions of the form
The (or a) function of S is to do Z. He calls one a strong etiological notion;
the other is less demanding, with close ties to Robert Cumminss (1975)
notion.1
According to Cummins, the (or a) function of a part of a system is to do
Z if it contributes Z-ing to a complexly achieved capacity Z* of a containing
system that a researcher is trying to explain by means of a functional analy-
sis, also known as a mechanistic or operational explanation. The crucial
difference is that, on Kitchers analysis, the relevant complexly achieved
Z* capacities are not pragmatically determined but are capacities for which
the system is designed.
Unlike Cummins, Kitcher thinks that the etiological notion of function
also has an explanatory role in biology. However, Kitcher argues that the
etiological notion is too onerous, ontologically and epistemically, to be of
much use in explaining how organisms operate or function. In my view
Kitchers argument is more worrying than the one originally offered by
Cummins, which was a critique of Larry Wrights (1973) formulation of the
etiological notion.2
To assess Kitchers argument for function pluralism, I first need to
explore an interpretive question concerning his more and his less demand-
ing notions of function. Precisely how demanding is each of them? Ishall
argue that his more demanding notion is more demanding than it needs to
be and that his less demanding notion is open to two interpretations:on
one it is not demanding enough for the explanatory role that Kitcher
assigns it; on the other it is a middling-strong etiological notion that might
play both of the explanatory roles that Kitcher identifies. Ishall also argue
that Kitchers argument for function pluralism in any case leaves it open
whether there is a middling-strong etiological notion of function that
might play bothroles.
1. Kitcher uses the phrase strong etiological conception several times (e.g., 1993c,
389). He refers to a less demanding account of functions once (1993c,388).
2. For other arguments for function pluralism, see Amundson and Lauder 1994;
Godfrey-Smith 1993; Brandon2013.
K i t c h e r s T w o De s i g n S ta n c e s [47]
negative view of natural selection, good designs arise as a result of random
processes (such as mutations, genetic recombinations, and environmen-
tal changes that affect development). Then they are passed on through the
mechanisms of inheritance and development. Selection is only responsible,
on this view, for preserving and distributing good designs in the popula-
tion and for eliminating bad ones. In my view this is correct, except for the
word only. The negative view ignores the role of cumulative selection in the
evolution of complex adaptations, which do not arise in a single saltation.
So it fails to recognize how Darwin answers Paleys question.6 On the posi-
tive view, simple adaptations can arise without natural selection. Complex
adaptations could possibly arise without natural selection. But complex
adaptations are vastly more likely to arise if selection is part of the evolu-
tionary process, with its repeated rounds of fresh variation, selection, fresh
variation, selection, and so on.7 The chance of a random adaptive alteration
to an existing mechanism M (an earlier version of a wing, for example)
increases as a result of selection of M, since the more M is replicated, the
greater the chance that a lucky alteration to M (an improvement on the
wing) will arise. Of course selection of M will also increase the probability
that maladaptive alterations to M will arise too. But selection can eliminate
these, as well as pick up and run with the improvements. Complex adapta-
tions arise through this mix of selection and randomly arising alterations.
By eliminating bad designs and multiplying the instances of good ones,
selection channels evolution.
Kitcher anyway does not seem to use the term design to denote a process
that necessarily involves fashioning. So anyone who (in my view wrongly)
rejects the positive view of natural selection in favor of the negative view
need not reject his account of functions on that ground. Kitcher sometimes
speak of natural selection as fashioning traits, but he also distances this
manner of speaking from the positive view of selection.8
The other reason someone might object to Kitchers assumption that nat-
ural selection is a source of design is that it might seem anthropomorphic,
like talk of Mother Natures intentions. This might be thought acceptable
for casual talk but not for serious analysis. Of course Kitcher fully under-
stands that natural selection is a purposeless process that lacks foresight or
any other kind of sight. So this objection is a mere verbal quibble. Perhaps
CONSEQUENCE ETIOLOGIES
K i t c h e r s T w o De s i g n S ta n c e s [49]
One of Wrights examples is of a belt buckle stopping a bullet and sav-
ing the life of the soldier wearing it. Wright says that the belt buckle serves
or performs the function of stopping the bullet and functions as a bullet
stopper, but it does not have the function to stop the bullet and save the
soldiers life, or not in the sense in which he is interested. Wright need
not suppose that we would never speak of the belt buckle as having the
function to stop the bullet, if we were (say) analyzing a hit by friendly fire
for the purpose of a military trial. Rather he is drawing our attention to a
certain notion of function by means of a locution that marks it out when
the contrast between the two locutions is explicitlydrawn.
Wrights main interest is in a teleological (or, if nonintentional, teleo-
nomic) notion of function. He tells us that the claim that the function of
the belt buckle is to hold up pants (and not to stop bullets, even if it did
both) is equivalent to saying that the belt buckle is there in order to hold
up pants (and not to stop bullets). In Wrights view all teleological explana-
tions invoke consequence etiologies. When an entity has a consequence
etiology, a consequence of the entity explains the entity. More specifically,
on Wrights analysis, an entity, S, has the function to do Z if and only if
(a)S does Z and (b)S is there (i.e., where it is and/or in the form that it is
in) because it does Z. Wrights hope is that this will work equally well for the
functions attributed to coffee mugs, calculators, and candelabra in every-
day life, as well as the functions attributed to biological entities, such as
hearts and hemoglobin, in pre-Darwinian biology and contemporary biol-
ogy. In this way its scope is ambitious.
However, Wright leaves certain details to be completed in context-
sensitive ways. For instance, he tells us that the formula is tenseless. In
other words, we may use whichever verb tense is appropriate in a given
case. Intentional design involves foresight, but natural selection does not.
So while a creationist might think that God gave mammals hearts with the
intention that they pump blood in the future, just as a potter gives a mug a
handle with the intention that it save drinkers from burning their fingers
in the future, a contemporary biologist will appreciate that natural selec-
tion cannot select traits on the basis of future advantage. Thus the second
requirement of Wrights formula is very schematic. In effect it says that if
S has the function to Z, S must be there because it does Z, or because it did
Z, or because it will do Z (or it is hoped that it will), depending on thecase.
Wrights first requirement says that S does Z, if doing Z is Ss function. It
is unclear if Wright intended this to be tenseless too, but he intended it to
deal with vestigial loss of function. It will, for instance, preclude the emus
vestigial wings from having the function to fly if it takes the present tense.
An emus vestigial wings are there (where they are and to some extent in
10. Ifirst defended a pure etiological analysis, which dropped the first requirement
and explicitly appealed to selection, in a widely circulated paper first presented to the
New Zealand Division of the AAP in 1980 as well as in my PhD dissertation, submitted
in 1983. See Neander (1991) for a more easily accessible early version.
K i t c h e r s T w o De s i g n S ta n c e s [51]
and (b)the hole remains in the pipe (where it is and in the form that it is in)
because the gas knocks out the scientist, who then cannot repair thehole.
A second counterexample from Boorse is of a man who kicks a dog,
intending to break its leg and cause it pain. If the man succeeds in his aim,
Wrights analysis entails that the break has the function to cause the dog
pain, for (a)it causes the dog pain and (b)it is there because it causes the
dog pain (that is, it was put there in order to dothat).
More recent etiological analyses usually drop Wrights first problematic
requirement (and seek other ways to handle vestiges). Plus they usually
require that the relevant history involve selection. The function of some-
thing is what an entity was (or entities of the type were) selected for doing.
This handles Boorses first counterexample, because the hole in the pipe was
not selected for leaking gas. More recent etiological analyses also usually
recognize that the details of the analysis need to be elucidated differently
for different kinds of selection. Most basically, while intentional selection
grounds artifact functions, it does not ground nonartifact functions. The
break in the dogs leg was intentionally selected to cause the dog pain, and
so it can have an artifact function to do so, on this type of analysis. But the
break in the dogs leg does not have a nonartifact function to cause the dog
pain, for there was no natural process of selecting leg breaks in dogs for
causing thempain.
Wrights aim to provide a univocal analysis for artifact functions and bio-
logical functions was abandoned. But even setting aside its other problems,
Wrights formula only anyway described a soft-focus unity of conception.
That is, it is vague or underspecified in certain ways that leave us with wriggle
room to fill out the details in different ways, as required for different kinds of
functions. The tenselessness of the formula is one instance of this. Its failure
to specify whether S stands for a token or for traits of a type is another. The
appeal to different tones of voice is yet another. In other words, recent etio-
logical analyses are explicitly acknowledging the disunity that was already
implicit in Wrights treatment. Recent etiological analyses can also equally
well accommodate a soft-focused unity of conception. All we need to do is
look less closely at the details. Whether artifact or nonartifact, an entitys
function is what it was (or entities of the type were) selected fordoing.
Further, there is no need for more unity of conception than this.
Scientists often refine everyday notions when they put them to use in spe-
cialized contexts. Scientists also often refine their technical terms in the
light of revisions to background theories. They can, moreover, be expected
to do so when the revisions are radical and the background theories are
central, in the way that the Darwinian Revolution was to the biologists
understanding of the functions of organismic traits.
While Kitcher thinks that etiological theorists have lost some valuable
unity of conception, he also seems to be pressing for more pluralism than
11. See Neander (1991) for a more developed version of this argument.
12. Not only do we tend to allow that functions can be grounded in intentional selec-
tion or natural selection, but we also generally allow that natural processes of selection
might include antibody selection, cultural selection, and some learning processes.
K i t c h e r s T w o De s i g n S ta n c e s [53]
was conceded in the previous section. He claims that two notions of func-
tion are used in ascriptions of the form S has the function to do Z, one
of which is a strong etiological notion, which he takes to be the one that
proponents of etiological theories support (I question this later), and a
less demanding notion. In relation to the less demanding notion, Kitcher
claims that an indirect connection to design suffices. But just how indirect
does he think it can be? Kitcher first introduces the less demanding notion
in relation to intentional selection.
Here is his leading example:You are building a machine for a purpose.
Lets say that you are building it to ferry people across a river. You combine
certain parts:a couple of cables, a barge that is to be pulled across on the
cables, various parts of a motor to do the pulling, and so on. These parts
all have jobs that you explicitly intend them to do. In addition you acciden-
tally drop a screw into the machines inner workings, where it luckily lodges
between two parts and makes a connection without which the machine
would not work. You never learn about the dropped screw. You do not know
that it is in the machine, that it makes the needed connection, or that such
a connection is needed. The screw was not put in place originally, nor left in
place later, with the intention that it make such a connection. Kitcher says
that the screw has the function to make the connection in the machine.
Unfortunately there are two ways to read his claim, because the example
can be further elaborated in two different ways. For the first reading, sup-
pose that the screw accidentally falls into a screw thread and works its way
in, screwing the two parts together. On an etiological analysis the screw
has the function to screw two parts together, since it was designed to do
this and it retains its original design function while lodged in the machine.
Moreover it performs this function in the machine, even though no one
intended it to perform it in this particular machine.13 If this is what Kitcher
has in mind by an indirect connection to design, the connection to design
is not much loosened.
For the second reading, suppose that the screw gets stuck between two
parts without working its way into a screw thread or screwing two parts
together. Instead it gets wedged between two parts, and it makes an elec-
trical connection. In this case, that it is a screw is irrelevant. On this read-
ing Kitchers claim is that the screw has the function to make the electrical
13. Had the screw not made the connection, the machine would not have worked. So
if the screw had not made the connection, the machine might have been dismantled
and rebuilt or been further worked upon. So maybe the screw is there because it makes
the connection. It is therefore unclear, on Wrights analysis (which does not require
explicit intention), whether the screw has a function in the machine.
the Creator intended that jackrabbits should have the ability to thrive in desert
environments... [but] that there was no explicit intention about the length of
14. The issue of how direct the connection between function and design is, in our
everyday conception of the functions of artifacts, should be settled with the help of
psychological investigations. See, for example, Matan and Carey2001.
K i t c h e r s T w o De s i g n S ta n c e s [55]
jackrabbits ears. Yet, because the length of the ears contributes to the mainte-
nance of roughly constant body temperature, and because this is a necessary
condition of the organisms flourishing (which is an explicitly intended effect)
the length of the ears has the function of helping in thermoregulation.(381)
Kitcher contrasts his less demanding notion of function with a strong etio-
logical one. Doesnt this exclude an etiological reading of his less demand-
ing notion? It would, except that Kitchers strong etiological notion is
16. There is another reason someone might want to privilege maintenance selec-
tion. One might raise Cumminss (2002) objection that selection is too fine-grained to
ground (appropriate) etiological functions. Cummins argues that, in order for wings
to have a function in flight, on an etiological analysis, there must have been a popula-
tion in which individuals had wings that allowed flight while other individuals lacked
wings that allowed flight. Cummins says that there was, rather, a gradual evolution of
wings, during which variants were selected for more effective flight or a different kind
K i t c h e r s T w o De s i g n S ta n c e s [57]
event it is not necessary to privilege maintenance selection in order to
accommodate vestiges and exaptations. True, a modern traits function
need not be the same as that for which ancient homologues were selected.
But think of the feathers of the flightless dinosaurs. Selection spread muta-
tions in the population that enhanced the ability of proto-feathers and
early feathers to contribute to thermoregulation. Their function in thermo-
regulation was grounded in this directed selection (as well as in any main-
tenance selection for them that was occurring at thetime).
We can handle changes in functions over time with a careful answer
to Kitchers second question, without restricting the relevant selection to
maintenance selection. The second question asks if the function-conferring
selection for a trait S of an organism O is (a)in the recent past, (b)in the
present, or (c)in the recent past and the present (relative to when O lives
and S comes into existence). Kitcher suggests that (b)is the popular choice
for biologists. The friends of the etiological theory generally choose (a)for
the reasons lucidly explained by Peter Godfrey-Smith (1994). I choose a
similar but longer answer that avoids the use of the vague term recent and
further obviates the need to privilege maintenance selection.17
Suppose we want to know if a feather has a function to assist in flight. In
principle the relevant history concerns the relevant lineage of historical homo-
logues. Clearly it matters if the feather belonged to a preflight dinosaur, an
emu, or an eagle, and it also matters what kind of feather it is (for example,
if it is a long tail feather as opposed to a down feather in either an adult or
a chick, since down is for insulation). Next the question to ask about this
lineage is if selection for assisting in flight operated on this lineage and, if so, if it
began prior to and continued up until the feather in question arrives on thescene.
of flight over other variants that flew less effectively or flew differently. One response
to this is that it ignores maintenance selection. Even when flight-enabling wings have
gone to fixation, highly deleterious mutations would undermine a capacity for flight
if not weeded out. This is a good reason not to ignore the importance of maintenance
selection, but it is not a reason to exclude directed selection from grounding functions.
Cumminss challenge deserves a longer discussion, but it seems to me to trade on shift-
ing inappropriately between coarse-grained and more fine-grained ways of speaking.
When we speak of wings having the function to enable flight, we speak in a very coarse-
grained way. Coarsely speaking, selection for flying more efficiently or selection for
flying in a different way is still selection for flying.
17. This follows Neander and Rosenberg (2012). Their proposal does not eliminate all
vagueness, but Ithink the remaining vagueness is unproblematic since it corresponds
to a plausible vagueness in functions. It is vague what counts as a cessation in selec-
tion. For how long must it cease? Aday or a season is too short. Are several genera-
tions enough? There is no sharp cut-off, but nor need there be. There is also vagueness
with respect to gradual transitions in selection for (say) leaping or gliding to flying.
However, this vagueness seems appropriate.
K i t c h e r s T w o De s i g n S ta n c e s [59]
seem bothered about its burden in discussing functions (although his argu-
ment is partly that) but that it is too onerous for some purposes, for reasons
that Idiscuss in the next section.
Before that we need to consider the second issue that Kitcher raises for
the proponent of the etiological theory:the issue of the alternatives. The
full burden of the etiological notion of function is revealed, Kitcher says,
when we consider the alternatives with respect to which an entity with a
function must have been selected. But this is where it becomes apparent
that Kitcher (1993c, 383)has an ultra-strong notion in mind. Here is a key
passage:
OPERATIONAL EXPLANATIONS
For most of Kitchers paper, his comments on his less demanding notion
of function seem to cast it as an etiological one, or as ambiguous in this
respect. At least on one reading, the dropped screw was doing what it was
designed to do (screwing two parts together). And, at least on one read-
ing the length of the jackrabbits ears was either thought by creationists
to result from a secondary source of design, or from Gods explicit and
advance intention regarding their length. This reading is also supported
by Kitchers Ultra-Strong Selection Requirement for his more demanding
notion of function. If this requirement is invoked in the case of the more
demanding notion, there is room for the less demanding notion to still
be etiological. The main difference between Kitcher and me, in that case,
would be with respect to whether the ultra-strong etiological notion is at
all useful. Idoubt thatitis.
However, when Kitcher gives a more general characterization of his
less demanding notion, the etiology seems to disappear, except insofar as
Kitcher clearly requires that the system as a whole must be designed (to
ferry people across the river, to thrive in a desert environment, and so on).
This is clearest in the comparison that Kitcher draws toward the end of
K i t c h e r s T w o De s i g n S ta n c e s [61]
Function and Design between his less demanding notion of function and
the one defined by Cummins (1975).
As mentioned earlier, Cummins claims that the (or a) function of a part
of a system is what it contributes to a complexly achieved capacity Z* of a
containing system, where Z* is a capacity that a researcher is interested in
explaining by means of a functional analysis, otherwise known as a mecha-
nistic or an operational explanation. In such an explanation, diverse func-
tions (Z1 Zn) are ascribed to the diverse parts (S1 Sn) of a system,
and it is the performance of these functions (in the appropriate spatial and
temporal order) that explains the systems abilitytoZ*.
Kitcher suggests that the relevant complexly achieved Z* capacity is
one for which the system is designed. When contemporary biologists are
explaining how organisms operate or function, the relevant Z* capacity is
one for which natural selection designed the organism. In general they are
designed to survive and reproduce in the environment in which they are
found, within the constraints imposed by hard-to-change features of their
Bauplan. This is the sole modification that Kitcher mentions making to
Cumminss account of the weaker notion of function. Thus he here seems
to come down decisively in favor of there being no need for the parts of
the organisms to have been selected for their functions. They need only
be responses to selection pressures insofar as they happen to be adaptive.
The modification to Cumminss account is a major improvement. For
one thing, Cumminss proposal is too liberal because it implies that some
entities that lack functions have functions. For example, it implies that the
various entities that contributed to the formation of the solar system had
the function to contribute to its formation. It implies that stars have the
function to send heavy elements into the interstellar medium to help form
vast clouds of molecular matter, that preplanetary clumps of matter have
the function to collide and accrete into larger clumps, and that various dif-
ferent elements and compounds have various functions to behave in the
ways they needed to behave if solar systems like ours are to form. Those
who defend Cumminss proposal might contend that such a system is insuf-
ficiently complex, but it is quite complex.18 Kitchers revised notion does
not suffer from this problem.
18.One might also wonder if the simplicity of some artifacts is a problem for
Cumminss proposal: cups, spoons, and doorstops have functions. Cummins can
respond that they are parts of larger systems that include their users. This puts a lot
of weight on how systems are demarcated. On Kitchers analysis functions belong to
parts of systems only if the systems were designed. So this helps. It takes the weight
off the need for complexity and the delineation of systems in determining which parts
have functions and whichdont.
K i t c h e r s T w o De s i g n S ta n c e s [63]
we need to separate his claim concerning how the functions are analyzed
from his claim concerning the importance of this explanatory role. It is
apparently tempting to bundle them together. For example, in introducing
the etiological account and a Cummins-style causal-role account, Philippe
Huneman (2013, 2)says:
19. For a more extended treatment of the points made here, see Neander (2015).
K i t c h e r s T w o De s i g n S ta n c e s [65]
Does (a)or (b)best capture the physiologists notion of the system that
is functioning normally or properly?
(a) Asystem that functions as designed, with each part able to do what it
was selected to do (along the lines of the middling-strong etiological
notion).
(b) A system that thrives anyhow, in its environment, within the con-
straints imposed by general hard-to-change features of the organisms
Bauplan.
My answer is (a) and not (b). But neither Kitchers arguments nor mine
here settle this question.
If malfunction is possible, S has the function to do Z is compatible with
S lacks the capacity to do Z, if S stands for a token. If we use an etiological
notion, a token trait can malfunction because its function does not depend
on its own capacities but on what its historical homologues were selected
for. Kitchers analysis of his less demanding notion (on the less demand-
ing reading of it) seems to preclude the possibility of malfunction. Like
Cumminss analysis, it seems to entail that the functions of token traits are
among the traits actual causal capacities. (Readers will recall that a similar
problem arose in relation to Wrights analysis.)
However, if Kitcher allows S to stand for traits of a type (i.e., for Ss),
then there is room for individual instances to malfunction.21 But even so
there is a problem. Now the problem is that the analysis is underspecified.
We need to know (i)how many current tokens of a given type of trait must
have the adaptive capacity in question, (ii) how traits are typed for this
purpose, and (iii) in what environment(s) the relevant traits must be adap-
tive. (It cannot simply be whichever environment an organism is currently
found in; at any rate that does not seem to be the physiologists notion
of normal or proper function.) If the analysis is also supposed to apply to
artifacts, we will need to know how one-of-a-kind artifacts can still have
idiosyncratic functions with respect to which they might malfunction, and
how generic artifacts (such as accidentally dropped screws) are now to be
treated. Until the analysis is further specified, it cannot be fully assessed.
I have long believed that serious attempts to answer questions along
these lines push us toward an etiological notion of function and to dif-
ferent precisification for artifacts and the natural functions of parts and
processes in organisms. In my view, in relation to (i), the normal or proper
CONCLUDING REMARKS
K i t c h e r s T w o De s i g n S ta n c e s [67]
Reply toNeander
PHILIP KI TCHER
K i t c h e r s T w o De s i g n S ta n c e s [69]
and Lewontin, she has no interest in conjectures about the course of evo-
lutionary history. I look at the animals as they are now, she declares,
and Ifigure out the ways in which, given their major traits and environ-
ments, they face problems in surviving and reproducing. I then identify
the ways in which particular molecules, intracellular structures, organs,
and systems respond to those challenges. If she has read the philosophi-
cal discussions of the concept of biological function, she may add an ecu-
menical remark: So I agree with Cummins that my attributions rest on
causal analyses, based on information Ihave about the animals in the here-
and-now, but Cummins fails to see that the problem background arises
from Darwins fundamental ideas about lifefrom his hostile forces; the
Darwinian connection adds the idea that etiological theorists like Neander
highlight, but it does so in a less specific way, freeing me from any burden
of commitment to a particular style of selective history. With respect to
adaptationism, my conscience is clear.
Now to the details of Neanders critique. Consider first the dropped
screw. Neander rightly distinguishes two scenarios, one in which the screw
drops into a threaded holefunctioning as a screw!and one in which it
does something different (making an electrical connection). Ihad the sec-
ond of these in mind and should have been more explicit. On this second
scenario the placement of the screw was not selected, or even intentional;
its presence in the machine is a lucky accident. It has the function to make
the connection because the machine is supposed to do a certain job, and,
given major features of the structure, that requires an electrical connec-
tion between two separated bits of metal; without the fortunately placed
screw the connection would not be made. Ihope the parallel with the cas-
cade of more specific demands generated by Darwins fundamental chal-
lenge (Reproduce!) is evidentand Ialso claim that the screws function
has nothing to do with the history of its placement.
The second major ambiguity Neander rightly discerns in my account
concerns my remarks about the demands of selective history and the
extent to which natural selection has to pervade the history underlying
the emergence or maintenance of a function. This is a deep and important
critique, and the presentation in my 1993c paper is inadequate to meet it.
It now seems to me that the inadequacy stems from my having taken for
granted some morals Idrew from Gould and Lewontin that Ifailed to make
explicit. Inow hope to remedythat.
Let me begin with the line of thought most obvious in my 1993c paper.
If you contend that the function of X in Os is F because recent selection for
F explains the presence of X in current Os, you need to decide just how pow-
erful the role of selection must be:Under what styles of history involving
I entirely agree with Neander that (1)is too strong. But what about(2)?
Theres an obvious tendency to think that (2)can fail, even though selec-
tion does a significant amount of work in explaining the contemporary
prevalence of X among the Os. Imagine that a few rivals to X with supe-
rior expected reproductive success actually emerged, but that these were
fortuitously eliminated from the population. With respect to most of the
alternatives, however, their inferior contributions to reproduction doomed
them. Here we have a scenario in which selection is largely responsible for
Xs success, even though chance helps out from time to time. (2) would
then be false, even though the role of selection in the history seems sub-
stantial enough to ground the attribution of the function. Recognition
of that motivates Neanders proposal of a middling-strong etiological
notion that doesnt live up to the demand imposed by(2).
But there are complications. Suppose chance figures differently in the
story. The genome of the Os has hotspots at which mutations would allow
for all sorts of superior alternatives to X. There are lots of them, and the
probability that a significant number of them would not appear is tiny. Yet
by a bizarre quirk of fate, none of the beneficial mutations ever occurs. This
is one of those occasions on which the vastly improbable happens. Instead
X beats out the (relatively few) inferior competitors that doarise.
This latter story deploys the notion of potential helpful mutations,
introduced in too simple a form in the overstrong (1). Yet although the
scenario satisfies (2), it still seems inadequate to explain the presence of
X as the result of selection. For a vital part of the story is the exceptional
luck X experienced:there were all these close possibilities that would have
dethroned X. We should no more explain Xs presence by citing a history of
selection than we should explain a childs passing an exam on the basis of
her deep mathematical knowledge, when the questions posed were drawn
from a list that includes large numbers of problems on topics about which
she knows nothing, but the actual draw generated only the simplest types
of arithmetical calculations.
K i t c h e r s T w o De s i g n S ta n c e s [71]
Thus (2) can sometimes be too strong and sometimes too weak. The
forms of the evolutionary histories prove crucial. Notice also that chance
can enter in both the modes considered so far. Amend the scenario about
the favorable mutations at the hotspots to allow some of them to emerge.
By coincidence, however, when those mutations arise they happen to
cluster in particular locales, and freak storms in these locales generate
catastropheshere an avalanche, there a number of falling trees, elsewhere
an incursion of predatorsand in every case these chance events elimi-
nate all the superior rivals. Once again the presence of X isnt explained by
recent selection but rather by the infrequent arising of beneficial mutants
and by their unfortunate bad luck on the few occasions on which such
mutants do appear.
I anticipate the response that the scenarios just envisaged are too con-
trived to dislodge a middling-strong etiological conception of the style
Neander favors. But they are only the entering wedge to appreciating the
complex combinations of evolutionary forces that can easily occur in the
history of a lineage. Central to the Gould-Lewontin critique of adaptation
is the thought that traits are bound together in packages. The genetic basis
for a focal trait gives rise in development to the acquisition of other char-
acteristics. Instead of selection for F we should think of selection for
F & G1 & G2 & Gn, where the Gi are the ancillary traits with which F is
developmentallybound.
Consider another type of example. X delivers F, but a rival X* does F
better than X; X* also manifests a deleterious trait G, and the selective dis-
advantage of G outweighs X*s superiority with respect to F. X wins out in
history because it is selected over X*, but this isnt selection for F. Rather
X is selected because it doesnt show G. Nevertheless F may still be the
functionofX.
Imagine an extreme case. Long ago mammals had blood- pumping
devices of greater efficiency. Unfortunately the genetic basis for developing
the organs also generated a physiological condition under which there was
a relatively high chance of a fatal response to some modestly uncommon
environmental factor. Agenomic reshuffling produced a new developmen-
tal program, yielding the ancestral mammalian heartsignificantly less
durable and less efficient than the older device but happily free of the phys-
iological side-condition. Ever since, in all descendant mammalian lineages,
the significant competition has been with genomic reversals that produce
the old developmental program. Recent competition and recent selec-
tion have pitted the older pumping device against the familiar mamma-
lian heart. The latter has wonhas been selectednot because it pumps
K i t c h e r s T w o De s i g n S ta n c e s [73]
CHAPTER3
1. This paper appeared in Singh et al. (2001) before being reprinted in Kitchers
(2003a) collected papers. Kitcher himself cites the paper as 2000 (Kitcher 2003a, 13),
which was the physical publication, as opposed to the imprint, of the original volume.
Kitchers principle has been widely misrepresented. As the quotation
makes clear, causal democracy requires equality of opportunity, not equality
of outcomes. Nevertheless, like conservative political commentators, con-
servative philosophical commentators have represented causal democracy as
the demand that all causes be dragged down to the same level and the refusal
to acknowledge that some causes are more significant than others.2 But like
any good liberal, Kitcher is merely asking that all causes be given a chance to
reveal whether they play a significant role in development. He believes that
current empirical evidence suggests that genes and environment interact in
many different ways, depending on the phenotype being studied, so that the
relative significance of genes and environment must be assessed on a case-
by-case basis. He recognizes that the existence of powerful, standardized
techniques for investigating genetic factors provides a practical justification
for focusing on genes. But he thinks, or at least thought at that time, that the
focus on genetic factors to the exclusion of environmental factors is greater
than can be justified by these practical considerations.
Kitchers (2003a) key message in Battling the Undead is that patient
reiteration of interactionism and causal democracy are all that is needed
for a balanced and accurate assessment of the role of the genes in devel-
opment. He denies that there are any deeper conceptual reasons for the
persistent neglect of the role of the environment in development, contra
Richard Lewontin (1983), Susan Oyama (2000b), and Russell Gray and
myself (Griffiths and Gray 1994). Kitcher agrees that simplistic genetic
determinism can seem like a vampire, rising from the grave each time
it seems to have been dispatched, but he counsels patience rather than
searching for a conceptual stake in the heart (Kitcher 2003a,283).
The decade since Kitcher proposed the causal democracy principle has
seen greatly increased interest in the role of environmental factors in
development (Griffiths and Stotz 2013, ch. 5). Amajor reason for this has
been the rise of the developmental origins of health and disease paradigm
in medicine, which has redirected some of the vast resources available for
biomedical research. Research into obesity, for example, now targets not
2.See, for example, Franklin-Hall 2015; French 2012, 197; Okasha 2009, 724;
Rosenberg and McShea 2008, 174; Thornhill 2007, 206; Weber 2006, 607; Woodward
2011, 249. Most of these authors attribute the idea of causal democracy to Susan
Oyama (2000a, S333), citing her response to Kitcher where she says that she will not
adopt his phrase causal democracy because it introduces into already-complicated
discussions rather more additional baggage than is likely to be helpful. Perhaps these
authors have been misled by the fact that the phrase occurs in the title of Oyamas
paper. Kitchers democracy principle is not dissimilar to Oyamas demand for parity of
reasoning when comparing genetic and nongenetic factors in development, a demand
that is usually parodied in the same way as Kitchers democracy principle.
P r o x i m at e a n d U lt i m at e I n f or m at i o n i n B i ol o g y [75]
just obesity genes like LEPR but the epigenetic effects of maternal nutri-
tional state on offspring physiology and broader exogenetic pathways from
parent to offspring, such prenatal and neonatal influences on food prefer-
ences. In evolutionary biology there has been a parallel rise of interest in
adaptive phenotypic plasticity, both within and between generations. In
many species a significant component of fitness differences results from
transgenerational environmental influences, or parental effects. Even
behavioral geneticists, the target of some of Kitchers strongest criticisms,
have recognized the need to broaden their research to embrace an interac-
tive picture of behavioral development (Hamer 2002). Some of the power-
ful techniques for studying genetic factors that seemed to provide practical
reasons to focus on genes have simply been repurposed for studying non-
genetic factors. To take just one example, second-generation sequencing
can be used for high-throughput screening of epigenetic marks. The spirit
of democracy seems to have been handsomely vindicated.
This picture of how the gay gene causes same-sex preference does not
reflect the actual scientific content of behavioral genetic research (Hamer
etal. 1993), but it was clearly the picture operating in public discussion of
that work. The idea that genes are units of information meant the gay gene
was understood as an intentional causethe brain is constructed using
a homosexual blueprint, or it is instructed to be homosexual. But inten-
tional causation is utterly different from the interactive, context-sensitive,
difference-making role of genes envisaged by the interactionist consen-
sus (Sterelny and Kitcher1988).
In more recent work Iand my collaborators have tried to bolster such
anecdotal arguments with a program of experimental research on the folk-
biology of behavioral developmenthow development is understood by
people without formal education in biology (Griffiths etal. 2009; Linquist
etal. 2011). This work provides some empirical support for Oyamas con-
tentions that people hold a dichotomous view of development in which
some phenotypic traits express the organisms inner essence while oth-
ers are imposed on it by the environment; that traits that come from the
inside are thought to be resistant to modification by the environment;
and that this inner essence is nowadays thought to be in the genes.
In this essay, however, Iwant to consider how the concept of informa-
tion can play the opposite role, helping to vindicate the principle of causal
democracy.
P r o x i m at e a n d U lt i m at e I n f or m at i o n i n B i ol o g y [77]
variable on another, the kind of dependence that is measured in informa-
tion theory. One variable carries information about another whenever the
values of the two variables are systematically related. Intentional informa-
tion is the context-insensitive aboutness described earlier. Intentionality
was introduced into philosophy over a century ago as the distinctive fea-
ture of human thought and language. In recent decades, however, phi-
losophers and biologists have argued that intentionality can be created by
natural selection. On this teleosemantic view a variable carries inten-
tional information if it evolved for the purpose of representing another
variable (Maynard Smith 2000; Millikan 1984). Hence intentionality can
be ascribed to bacteria and to genes, not just to humans.
Causal and intentional biological information can be thought of as
proximate and ultimate information, respectively (Griffiths 2013; Mayr
1961). Causal biological information is a way to describe the causal struc-
ture of a living system: How do the parts of the system depend on one
another? It can be characterized mathematically using information theory.
Hence the study of causal information is part of proximal biology. However,
if intentional biological information is teleosemantic information, then it
is a way to describe the purpose for which some aspects of a living system
evolved:What is the evolutionary function of these parts of the system?
This is an ultimate or evolutionary biological question. Knowing how a sys-
tem works will not tell us anything about the teleosemantic information
it contains, unless we also know the selection pressures that created and
maintain the system.
In sections 4 and 5 I describe a new theory of causal/proximate bio-
logical information. This is in the spirit of Sterelny and Kitchers (1988)
difference-making analysis of the sense in which a gene may be a gene for
a phenotype despite the interactionist view that every phenotype depends
on many variables. The new theory uses the recent interventionist view
of causal explanation (Woodward 2003). Some of my earlier claims about
causal/proximate biological information need to be revised in light of this
recent work. Ipreviously argued that causal/proximate information could
not distinguish genetic from nongenetic causes because it is ubiquitous.
Any variable that has an effect on the development of a phenotype will con-
tain information about that phenotype in the sense of information theory;
knowing the state of the causal variable reduces our uncertainty about the
state of the phenotype (Griffiths and Gray 1994; Maynard Smith 2000).
In this more recent work, however, my collaborators and Ihave used the
concept of causal specificity to discriminate between causes that provide
information for their effects and those that do not (Griffiths etal. 2015;
Griffiths and Stotz 2013). This reintroduces the possibility that genetic
4.CAUSAL SPECIFICITY
It has long been argued that because the effect of an allele substitution
depends on many other factors, both other genes and the environment,
it is misleading to identify a single allele as the gene for a phenotype.
However, the fact that alleles produce phenotypes interactively does not
prevent their being salient causes of those phenotypes in the interactionist
picture favored by Kitcher. Alleles cause phenotypes by making a difference
to those phenotypes against a background of other factors. This idea was
spelled out in detail by Kitcher in an article with Kim Sterelny:
Gray and I offered two criticisms of this definition (Griffiths and Gray
1994). The first concerned the definition of standard environment, to
which Kitcher (2003a, 29192) responded with an amended definition. The
second criticism was that the definition could equally license epigenetic
P r o x i m at e a n d U lt i m at e I n f or m at i o n i n B i ol o g y [79]
marks for or incubation temperatures for phenotypes. Kitcher does not
regard this as a criticism and believes that we can and should treat genes
and environment symmetrically in this respect:Far from being a reductio
of the interactionist view, this point simply testifies to the democracy prin-
ciple introduced above(293).
The idea that there are genes (and other factors) for phenotypes is
closely related to the idea that genes contain information about pheno-
types. Sterelny and Kitchers difference-making approach can be readily
translated into information-theoretic terms. Their analysis identifies a
covariance between gene and phenotype when other factors are held con-
stant. We can regard the gene as a signal source, the phenotype as a signal
receiver, and the other factors as channel conditions. When the channel
conditions are stable, we can reduce our uncertainty about the state of the
phenotype by observing the state of the gene, so the gene carries informa-
tion about the trait (Griffiths and Gray 1994). Iobserved earlier that infor-
mation in this sense is ubiquitous. All developmental factors carry such
information. For many authors, the fact that causal information is found
in all factors affecting development is a reason to look for another kind of
information that is found only in genes, or in genes and some special selec-
tion of environmental factors. This has typically been teleosemantic inten-
tional information (Maynard Smith 2000; Shea 2007; Sterelny etal. 1996).
However, an alternative strategy is to develop a more discriminating causal
account of information using resources from the philosophy of causation
and information theory. It is this strategy that my collaborators and Ihave
pursued in our recentwork.
The influential interventionist theory of causal explanation provides
new resources for the study of causal difference- making (Woodward
2003). It provides formal criteria that distinguish causal from noncausal
relationships, based on the insight that causal relationships are relation-
ships that are potentially exploitable for purposes of manipulation and
control (Woodward 2010, 314). The theory treats causation as a relation-
ship between variables in a scientific model, using causal graph theory as
a canonical format in which to express these models. There is a causal rela-
tionship between variables X and Y if it is possible to manipulate the value
of Y by intervening to change the value of X. Intervention here is a techni-
cal notion with various restrictions. For example, changing a third variable
Z that simultaneously changes X and Y does not count as intervening on
X. Causal relationships between variables differ in how invariant they are.
Invariance is a measure of the range of values of X and Y across which the
relationship between X and Y holds. But even relationships with very small
ranges of invariance are causal relationships.
P r o x i m at e a n d U lt i m at e I n f or m at i o n i n B i ol o g y [81]
Mutual information is symmetrical:I(X; Y) = I(Y; X). So variables can have
mutual information without being related in the manner required by the
interventionist criterion of causation. However, our measure of specificity
does not simply measure the mutual information between C and E. Instead
it measures the mutual information between interventions on C and the
variable E. This is not a symmetrical measure because the fact that inter-
ventions on C change E does not imply that interventions on E will change
C:in general, I(C ; E) I(E; C), where C is read do C and means that the
value of C results from an intervention on C (Pearl2009).
Any two variables that satisfy the interventionist criterion of causation
will manifest some degree of mutual information between interventions
and effects. If the relationship C E is minimally invariant, that is, invari-
ant under at least one intervention on C, then C has some specificity for E,
that is, I(C ; E)>0. Conversely, if C E has some degree of specificity, then
the relationship is invariant under at least one intervention on C.
Elsewhere we have argued that a causal relationship in biology should
be regarded as an informational relationship when it is highly specific
(Griffiths and Stotz 2013). We are not the first to draw a link between
information and specificity. Woodward (2010, 312n21) has written, The
ideas of causal specificity and information are obviously closely linked.
Biologists tend to think of structures as carrying information when they
are involved in causally specific relationships. Iregret that Ilack the space
to explore this connection in more detail. Sahotra Sarkar (2004) and Ulrich
Stegmann (2014) have also argued that the salient causes in development
are the most biologically specific causes. Sarkars set-theoretic analysis of
biological specificity is very similar to Woodwards idea of a bijective map-
ping, and Isuggest that biological specificity is simply causal specificity in
a biological system.
Using our measure, a causal relationship will be highly specific whenever
C and E can take many values and there is a high degree of mutual informa-
tion between them. In informal terms the cause can make the difference
between many different states of the effect and can be used to exercise fine-
grained control over that effect. This is actually what Francis Crick (1958,
153)meant by information when he introduced the sequence hypothesis
and the central dogma of molecular biology: Information means here
the precise determination of sequence. The distinction between instruc-
tive and merely permissive causal interaction in developmental biology
is also a distinction between more and less specific causes in our sense.
The proposal to identify biological information with causal specificity in
biological systems is thus a classic explication (Carnap 1950). We construct
Some authors have suggested that the very idea of causal specificity refutes
the principle of causal democracy (see note 2). This relies on the misrep-
resentation of causal democracy as the view that all causes are equally
significant, so it can be refuted merely by showing that it is possible to
discriminate. But Kitchers principle is that all causes should be given a
chance to show their significance. This is evidently compatible with the idea
of causal specificity and with other theories of causal selection.
When Crick advanced the sequence hypothesis and central dogma he
assumed that the sequence of the gene not only precisely determined the
sequence of the product but also completely determined it. The discovery
of alternative splicing in the 1970s showed that the sequence of the gene
can underdetermine the sequence of the product. Since then, alternative
splicing has turned out to be ubiquitous in eukaryotes and has been joined
by other mechanisms of pre-and post-transcriptional processing: mRNA
editing, co-transcription, programmed frame shift, trans-splicing, transla-
tional recoding, and protein trans-splicing. The transcriptomethe total
population of RNAs found in the cells of an organismis at least an order
of magnitude greater than the number of genes.
These mechanisms are employed because there are many different types
of cell, each of which uses the same genetic resources to make a different
set of products; even a single cell uses those resources differently at differ-
ent stages in its life cycle. This requires additional specificity of a kind not
captured by the original sequence hypothesis. Cricks biographer Robert
Olby (2009, 251, italics added)notes:
P r o x i m at e a n d U lt i m at e I n f or m at i o n i n B i ol o g y [83]
This point was immediately obvious to Cricks contemporaries and led the
ciliate biologist David L.Nanney (1958, 712)to introduce the idea of epi-
genetic control systems:
This view of the nature of the genetic material... permits, moreover, a clearer
conceptual distinction than has previously been possible between two types
of cellular control systems. On the one hand, the maintenance of a library of
specificities, both expressed and unexpressed, is accomplished by a template
replicating mechanism. On the other hand, auxiliary mechanisms with differ-
ent principles of operation are involved in determining which specificities are to
be expressed in any particular cell.... To simplify the discussion of these two
types of systems, they will be referred to as genetic systems and epigenetic
systems. The term epigenetic is chosen to emphasize the reliance of these
systems on the genetic systems and to underscore their significance in develop-
mental processes.
The philosophical literature on causal specificity has not been blind to this
aspect of specificity, and Woodward (2010, 3045, italics added) has noted
that specificity includes systematic dependencies between a range of dif-
ferent possible states of the cause and different possible states of the effect,
as well as dependencies of the time and place of occurrence of E on the time
and placeofC.
It is evident that the additional research questions to which Olby refers
concern additional sources of specificity. Our account of biological infor-
mation as causal specificity chimes neatly with the way many biologists
use information in these contexts. Regulatory mechanisms that affect how
coding sequences are used in a particular cell at a particular time have
been described as amplifying the information in those coding sequences.
Biologists in this field search for the target sequence specificity of
forms of editing (Davidson 2002)or search for the missing information
needed to supplement the information in the coding sequence (Wang and
Burge2008).
Nanney hypothesized that the utility of the epigenetic control systems
lies precisely in their ability to respond specifically to altered environmen-
tal conditions (1958, 713, italics added). He suggested that the influence
of these systems should be understood in terms of their specificity of
induction of developmental effects (715). We see the same language of
specificity employed in developmental biology when biologists distinguish
between the more specific instructive and the less specific permissive induc-
tive interactions.
6.ULTIMATE INFORMATION
To make the case against gene centrism, DST [developmental systems theory]
should be pointing to the undoubted specialness of genes and saying, You know
that property, the one that makes genes so special? Well that property is found
not just in genes but in several other factors in development. That special role is
to transmit information, generated through a process of natural selection, down
the generations to inform development. (Shea 2011,61)
P r o x i m at e a n d U lt i m at e I n f or m at i o n i n B i ol o g y [85]
(c) an evolutionary explanation of the current existence of the represent-
ing system adverts to Rs having carried the correlational information
that condition C obtains;and
(d) C is the evolutionary success condition, specific to Rs, of the output of
the consumer system prompted by Rs. (Shea 2013,5)
P r o x i m at e a n d U lt i m at e I n f or m at i o n i n B i ol o g y [87]
function will be incoherent seems to me to be a non sequitur. Here are four
versions of the argument:3
1. What a trait is adapted for is fully determined by facts about the past,
although admittedly these may be hard to discover. But there is no
determinate fact about what a trait is currently adaptive for. It may do
something for one organism but not another, or in one locality but not
another. Reply:What a trait is adapted for is defined as what it was
adaptive for in the past. So the first cannot be determinate if the sec-
ond is not. To identify the evolutionary forces now acting on a popula-
tion we have to look for general patterns that may not hold for every
organism, but we have to do that to identify the evolutionary forces that
acted in the pasttoo.
2. Fitness depends on the environment. When studying adaptation we
know exactly which environments are relevant:those in the actual past.
But when studying why a trait is currently adaptive we have no objec-
tive basis for rejecting abnormal environments. Some animals live in
zoosshould we include them? Reply:This is a more concrete version
of the first argument. There were animals in zoos in the past too. We
need to make decisions about what constitutes a single selective process
whether we are looking at the past or the present.
3. The evolutionary trajectory leading to the current population is deter-
minate, a matter of past facts. But its future trajectory is indeterminate
because future events may interfere with our best prediction. Reply:The
ahistorical account does not need to predict the future. It only needs to
establish the causes of current fitness. Just as a moving object has an
instantaneous trajectory in space, a population has an instantaneous
evolutionary trajectory. In experimental studies of evolution, it is this
instantaneous trajectory that we actually study. Even what might seem
to be essentially predictive traits, such as variance of offspring num-
ber, can be defined at an instantthat is why we can design games
of chance like lotteries rather than having to construct them by trial
anderror!
4. If functions are defined ahistorically, they will not explain why the traits
that have these functions exist. Reply:Ahistorical function is not meant
to replace historical function, any more than adaptive can replace
adaptation. The two are complementary. To explain why a trait exists
3. It has also been argued that we cannot describe organisms ahistorically because
their parts are defined by their adaptive function (Neander 2002; Rosenberg and
Neander 2009). For a refutation, see Griffiths (2006b).
P r o x i m at e a n d U lt i m at e I n f or m at i o n i n B i ol o g y [89]
This proposal can be made more comparable to the causal/proximate
notion of information described in sections 4 and 5 by stating it is as a rela-
tionship among three variables. C continues to denote Sheas environmen-
tal condition; R continues to stand for the state that signals this condition;
and Iintroduce E (effect) to denote the outputs of Sheas consumer system.
A causal variable contains adaptive informationif
7.CONCLUSION
4.This publication was made possible through the support of a grant from the
Templeton World Charity Foundation. The opinions expressed in this publication are
those of the author and do not necessarily reflect the views of the Templeton World
Charity Foundation.
P r o x i m at e a n d U lt i m at e I n f or m at i o n i n B i ol o g y [91]
Reply toGriffiths
PHILIP KI TCHER
The genomic research of the past two decades has yielded increasingly com-
plicated pictures of both development and evolution. Along with Karola
Stotz and John Dupr, Paul Griffiths has been at the forefront of attempts
to bring new discoveries about the dynamic genome to the attention of phi-
losophers. As Pauls illuminating essay reveals, even those who campaigned
in the 1980s and 1990s against the oversimplified scenarios often spun in
discussions of the biological basis of human behavior, tales woven by
biologists as well as by philosophers, underrepresented the wide variety of
ways in which epigenetic and environmental factors enter into proximate
and ultimate causation. Even the most ardent interactionists were unaware
of the intricacies of the interactions.
Griffiths sees how biologically informed philosophy can make far more
precise the principle of causal democracy I advocated in my critique of
three prominent ventures in articulating general models of development
and evolution that hoped to rival (and displace) the dominant gene-centric
approaches. Susan Oyama, Richard Lewontin, and Griffiths (along with
Russell Gray) aimed at a novel picture of ontogeny and of evolution, one
that would prevent the hasty and unwarranted forms of genetic determin-
ism, seemingly arising with monotonous and potentially harmful regular-
ity. Although I shared these thinkers concerns, I believed that no such
general account was needed. Recognizing causal democracy would suffice.
The principle of causal democracy was intended, as Griffiths sees, to
offer equal opportunity to a wide variety of potential factors. It allows dif-
ferent investigators to focus on a particular type of cause and to explore the
effects as other factors vary. It also encourages them to consider a variety
of potential causes. Perhaps at the end of the day, when different processes
are analyzed in terms of different types of causes, it will turn out that DNA
sequences play the most important role in a plurality of cases. Or perhaps
notsome other factor might dethrone the gene. Alternatively we might
discover that no single type of cause is the most fundamental across
any significant range of developmental and evolutionary phenomena. The
whole idea that some particular biological factor is more important than all
others might turn out to be a mistake.
Griffithss views have evolved since our debate of a decade agoand
so have mine. Although Icontinue to believe that the principle of causal
democracy offers valuable methodological counsel in the design of biologi-
cal research and the appraisal of biological claims, Inow think my earlier
P r o x i m at e a n d U lt i m at e I n f or m at i o n i n B i ol o g y [93]
are privileged because the genes are the sole information-bearing biological
entities hostage to debates about the future of semantics.
So Iprefer a different strategy. Sterelny and Kitcher (1988) and Kitcher
(2003a) eschew talk of information in terms of a more austere causal idiom;
as Griffiths rightly points out, those articles look at the differences causes
make, an approach elaborated and much improved in the recent work of
James Woodward and Kenneth Waters. Iregard the talk of information as
a metaphor, useful when it helps to make some biological issue vivid, but
potentially misleading, especially when users forget that the coinage is not
literal. The ultimate test of a use of the metaphor is to translate back into
the primitive austere idiom and to frame questions in terms of difference-
making causation. Causal democracy recognizes that there are all kinds
of difference-making causes: sometimes allelic differences are important
across a wide range of environments; on other occasions an environmental
factor (the absence of an important source of some nutrient, say) makes
a large and uniform difference across a spectrum of genotypes. The basic
causal idiom provides no basis for singling out some causes as always more
crucial than others. Treating informational language as metaphor thus
diagnoses gene-centrists as people who have allowed a figure of speech to
run riot in their thinking.
Griffiths touches on my second debate in his optimistic assessment of
the ways biomedical research is broadening its menu of options for explor-
ing the causation of disease. Iam less hopeful. In my judgment the narrow
focus on genes continues to dominate. Asymptom of that dominance is the
current fascination with personalized medicine. Many universities with
prominent schools of medicinemy own among themare giving very
high priority to the establishment of centers in which researchers will use
information about DNA sequences in attempts to identify drugs and other
forms of treatment that can be beneficial for people who carry a particular
sequence (or some member of a family of sequences). This form of gene-
centered inquiry is often heralded as the next phase in the progress of
medicine.
Who could complain? Surely it would be better if physicians knew in
advance which of a collection of medical regimes would be optimal for an
individual patient. Yet its worth asking who the beneficiaries are likely to
be. The universities who rush to achieve eminence in personalized medi-
cine are not driven purely by a laudable wish to ameliorate the human con-
dition. Personalized medicine and translational research are terms that often
appear in close proximityand the juxtapositions should remind us of a
basic fact:theres money in it. Visions of profitable patents already dance
before administrative eyes. When the universities receive the returns on
P r o x i m at e a n d U lt i m at e I n f or m at i o n i n B i ol o g y [95]
The principle of causal democracy descended from my efforts to expose
the flaws in the ambitious style of human sociobiology practiced in the
1970s and early 1980s (Kitcher 1985). The mainstream evolutionary psy-
chology of today rarely commits all the errors of its sociobiological ances-
tor, but theres often a simplified way of thinking about evolution and
development, akin to those Icriticized, that is essential to the attention-
grabbing conclusions. Were told that a particular behavioral propensity
the tendency of young females to find attractive those males who can
supply resources, saywould have been advantageous on the savannah.
This serves as the basis for a hypothesis about a modular psychological
capacity, which can then be seen as operative across a range of experimen-
tal or survey data. Contemporary womens behavior is thus interpreted
as expressing their savannah-selected propensity to be attracted to well-
provided (typically older) men. Add the thesis that selection can operate
only where there is genetic variation, and the way is open for the conclu-
sion that this propensity is hard-wired. At this point, of course, the jour-
nalists pick up thestory.
Even from the far more limited perspective of the 1980s, there were
many evolutionary and developmental alternatives. Today, with our richer
awareness of genomic complexities, the route to genetic determinist con-
clusions is even more crowded with rival scenarios, not to be dismissed out
of hand. If it is to fulfill its ambitions, evolutionary psychology must go
beyond the simple narratives that dominate manybut not allventures
under the Santa Barbara paradigm. It should emulate the approach of
Bateson and Nettle, with its thorough incorporation of causal democracy.
The simplifications of parts of evolutionary psychology are, however, as
nothing compared with the recent fad for evolutionary debunking argu-
ments in philosophy. Since the publication of Street (2006), meta-ethics
in particular has succumbed to a deluge of articles debating whether, if
moral realism is correct, a human capacity for morality could have evolved.
(Interestingly Streets original article is more sophisticated about evolution
than all those Iknow that have come after it.) Underlying the stream of
papers lies a common trio of ideas:evolution means evolution by natu-
ral selection; natural selection favors or frowns upon very specific traits,
things like a capacity for detecting moral truths; and underlying those
traits are genes forthem.
To hold any of these ideas you have to be very innocent with respect to
contemporary evolutionary theory. First, when the animal in whose evo-
lution you are interested is Homo sapiens, cultural transmission and cul-
tural selection can play a not inconsiderable part. Thanks to Robert Boyd
and Peter Richerson (1985), its been known for thirty years that cultural
P r o x i m at e a n d U lt i m at e I n f or m at i o n i n B i ol o g y [97]
CHAPTER4
1.INTRODUCTION
B r i n g i n g R e a l R e a l i s m B ac k H o m e [99]
(see Massimi 2015a). Hence my very own (loosely Kantian-inspired) per-
spectivalist slant to real realism.
Key to the Galilean strategyas I see it through Kantian lenses (see
Massimi 2010)is not just to deploy the telescope to overcome fictitious
boundaries (i.e., those between sea and land, Venice and Amsterdam,
Heaven and Earth) but also to approach nature through principles of rea-
son in one hand and experiments thought out in accordance with these
principles in the other hand, yet in order to be instructed by nature not
like a pupil, who has recited to him whatever the teacher wants to say, but
like an appointed judge who compels witnesses to answer the questions he
puts to them (Kant [178187] 1997, Bxiiixiv). This is how Kant famously
portrayed Galileos contribution to bringing natural science onto the secure
path of knowledge after groping about for so many centuries. It is this
further Galilean strategy that Iturn my attention tohere.
I cannot do justice to the breadth of the philosophical arguments that
Kitchers real realism has put forward. And much as Id like to discuss
Kitchers articulated response both to the epistemological empiricism
of van Fraassenean flavor and to the epistemological constructivism of
Kantian descent,1 Ihave to leave those for another occasion. For here Icon-
centrate on Kitchers influential response against the blockish holism
of epistemological empiricism in its historical form (best expressed by
Laudan 1981), which seems to assume that a theory is false because it is
not entirely true (RR 170). In reply real realism insists that the past suc-
cesses stem from parts of the theories that are approximately correct, (RR
170)namely from those hypotheses that are genuinely put to work (i.e.,that
characterize working posits), and are as such approximatelytrue.
In section 2 Ireview Kitchers famous distinction between working pos-
its and idle wheels in the context of his realist defense against the challenge
coming from the history of science. In section 3 Ifocus on the notion of sci-
entific success and distinguish between two variants:success from within
and success from above. In section 4 Isuggest a perspectivalist take on
real realism in the form of a notion of success from within, able to assess
success from a human vantage point and to capture truth across scientific
perspectives. Iconclude by considering possible objections and replies to
the perspectival view canvassed in section4.
B r i n g i n g R e a l R e a l i s m B ac k H o m e [101]
exported to other examplescan similarly be leveled against the real
realist. Consider, for example, Aristotles theory of free fall as accelerated
motion toward a natural place. The theory was undoubtedly successful
by its own lights at the time, and it provided a springboard for medieval
commentators (from Simplicius to Hipparchus and the Arabic commenta-
tors), whose views fed into the impetus theory of Buridan and Oresme, and
ultimately into Galileos early Pisan studies on free fall (see Massimi 2010,
2015b). What are the working posits in Aristotles theory of free fall? And
where do theoretical excrescences begin? Was Aristotles hypothesis that
bodies get heavier nearer the Earth, an idle wheel? Well, it provided an
explanationin Aristotles own scientific perspectiveof why free-falling
bodies accelerate (as opposed to decelerate or move with constant speed)
when moving toward their natural place (where, he assumed, bodies would
regain their form). Moreover it suggested that there might have been
forces acting on the body and pulling it either toward its natural place or in
some different direction (what Avicenna and Abl-Barakt called natural
and violent mail and what Buridan called impetus, as an intrinsic force due
to a natural gravity, which was in turn the ancestor of the early Galileos
gravitas as a weight-related concept and ultimately of Newtons gravita-
tional mass). Was there anything approximately (or even remotely) true in
Aristotles theory? Or should we conclude that Aristotles theory was quite
simply false? What has gone wrong with this example?
B r i n g i n g R e a l R e a l i s m B ac k H o m e [103]
Galileos kinematic studies exemplify the perspectivalist strat-
egy of engaging with the past from within (rather than from above).
Working with the Aristotelian tradition that goes from Hipparchus to
the Arabic commentators and Buridan and Oresmes impetus theory,
Galileo could operate within well-trodden paths. He could resort to
Archimedess buoyancy and Hipparchuss theory of free fall and intro-
duce gradual changes to key concepts. For example, the change from
impetus as an internal force propelling a body to momento (momen-
tum gravitatis) as an internal force that, after having propelled the
body, would gradually decay, causing the body to acquire degrees of
speed (celeritatis momenta) in its descent. Galileos breakthrough about
free fall did not happen by debunking the Aristotelian tradition (pace
Galileos own rhetoric against Simplicius in Two New Sciences). Nor did
it happen by selecting working posits in the Aristotelian tradition, for
even Archimedean buoyancy and Hipparchuss theory were inextrica-
bly entangled with idle wheels and not amenable to being imported
tout court into the Galileanstory.
Instead the Galilean kinematic strategy consisted in small theoretical
steps and subtle conceptual nuances that ultimately allowed Galileo to turn
the corner from the Aristotelian tradition. Galileos ability to interrogate
nature with principles of reason on the one hand (i.e., the indubitable prin-
ciples from which he demonstrated the law of free fall) and with experi-
ments thought out in accordance with these principles on the other hand
(i.e., both thought experiments with chords and real experiments with
inclined planes) made the revolutionary shift possible. The perspectival
realist can appeal to this Galilean strategy to bring the real realists notion
of success back home:from above to within.
But how should the perspectivalist notion of success from within be under-
stood? So far Ihave simply suggested that it should not be understood as
the ability of inquirers to identify parts of a theory that are essential to
success and hence approximately true. But this can hardly be enough to
understand the perspectivalist move Iam suggesting for real realism. We
need to unpack the slogan.
In what follows Itake my cue from broader discussions on perspectiv-
alism in contemporary epistemology to propose that success from within
should be understood as success with respect to standards of performance
adequacy appropriate to the scientific perspective of the inquirer when
2. See, for example, Sosas perspectival coherentism (part of his virtue perspectivism
in Sosa 1991), where the justification for beliefs is a matter of perspectival coherence.
Along similar lines, on perspectival justification for beliefs, please see Haack (1993)
and Rosenberg (2002, 149):The reason that we correctly judge that S does not know
that p is that, given our richer informational state, we recognize that what we are (stip-
ulatively) entitled to take to be Ss epistemic circumstances demand a higher level of
scrutiny than we are supposing S himself to have exercised. S therefore, has not satis-
fied what, from our perspective, are the standards of performance-adequacy appro-
priate to his epistemic circumstances, and hence, from our epistemic perspective, we
judge that, despite his not having acted irresponsibly given the information available
to him (judged from his own legitimate perspective on his epistemic circumstances),
he has not justifiably come to believe that p. In what follows Ilatch onto and expand
upon Rosenbergs appeal to standards of performance adequacy, but in a different con-
text and with a different purpose in mind. My goal is not to elaborate a perspectivalist
theory of belief justification but instead to elaborate a perspectivalist notion of success
from within that can serve the purpose of success-to-truth-inferences in the realism
debate.
B r i n g i n g R e a l R e a l i s m B ac k H o m e [105]
performance adequacy appropriate to the Galilean-Newtonian epistemic
context (e.g., How to think of Galilean free fall when sense impressions
about the free mobility of rigid bodies and paths of light rays get called
into question, as they were with Helmholtzs mirror sphere thought
experiment, for example? What becomes of Galilean-Newtonian gravity
in a non-Euclidean space?).
Success from within is then the ability of a theory to perform adequately
with respect to standards that are appropriate to the theorys wider
epistemic contextor scientific perspective, as I prefer to call itwhen
assessed from the point of view not just of the scientific perspective at stake but,
crucially, from the point of view of other scientific perspectives. Building on
recent important work in epistemology,3 I suggest the following definition.
A scientific claim (SC) meets the criterion of success from withiniff:
This definition of success from within vindicates the real realists expecta-
tions for successful posits to track truths in nature (via the first part of
premise (b)). Yet it is perspectival in giving up on both a Nagelian view from
nowhere and a convergent realists/real realists view from here now: it does
not take our currently successful scientific claims as the gold standard for
assessing past failures and successes.
Success from within is kosher to the Kantian spirit of perspectivalism
in giving due consideration to epistemic agents (or, I should say, scientific
communities) commitment to scientific claims (without dismissing them
out of hand as sheer errors of the past). Success from within does justice to
historians anti-W higgish plea for judging past theories in their own terms
and by their own standards (not by ours) when assessed from the point of view
of other (diachronically subsequent or synchronically rival) scientific perspec-
tives. At the same time, it avoids the perils of truth relativism by anchoring
3. Here Iwant to latch onto the helpful distinction between context of use and con-
text of assessment in discussions on relativized truth and faultless disagreement. See
MacFarlane (2005, 2009)and Marques (2014), among many others. By contrast with
MacFarlane, I will not be using this distinction to defend any notion of relativized
truth. Instead Imake use of MacFarlanes distinction between context of use and con-
text of assessment to provide a notion of success in science that does not beg the ques-
tion for scientific realism (i.e., that does not judge past theories on the basis of our
current successful theories).
B r i n g i n g R e a l R e a l i s m B ac k H o m e [107]
These examples show important features about the definition of success
from within Ijustgave:
B r i n g i n g R e a l R e a l i s m B ac k H o m e [109]
Consistent explanations first. Assuming caloric is an imponderable fluid
as eighteenth-century scientists did and our hypothetical scientists would
presumably also dowould immediately pose severe challenges to any
attempt to provide a consistent explanation of mechanical work and states
of aggregation. Mechanical work would require caloric to be consumed (pace
conservation of caloric), as much as turning water into ice would require
removing caloric (qua a shell of imponderable fluid surrounding waters
particles) and yet expanding the overall volume. How can waters particles
lose part of their volume (by releasing caloric), while also expanding their
overall volume? Caloric does not seem to license consistent explanations.
Simplicity next (a notoriously slippery standard, if any). Would caloric
provide a simple and elegant account of various phenomena? Caloric could
be squeezed out of particles volumes (assuming a Daltonian model) and
get reattached to them at ease. As simple as that? Well, assuming some
mechanism was in place to explain what held caloric attached to the parti-
cles of matter, what had the power to detach it from matter and reattach it
at will, and so forth. Perhaps some attractive and repulsive forces might do
the trick. Or perhaps electrical fluids. Or some ethereal substratum (along
the lines of Kants matter of heat). Simplicity is not within easy reach.
Acomplex story would have to be told about the mechanisms underlying
calorics behavior in all these phenomena, mechanisms that can potentially
be at odds with eachother.
Perhaps accuracy with the available evidence fares better than consis-
tency and simplicity when it comes to standards of performance adequacy.
Let us assume our hypothetical community has produced a system of scien-
tific claims that are accurate by the experimental standards available to the
community at the time. Such claims must surely be regarded as successful
(no matter how false caloric is from our current vantage point). An analogy
may help here. Suppose I have an accurate story about hedgehogs living
in my garden and creeping out at night to collect the mulberries that have
fallen on the ground. My story is so accurate that it tells me with precision
that hedgehogs come at night, between 1 and 2 a.m., from the far right cor-
ner of the garden, behind the hedge, and collect only the juiciest mulberries
they can get their spiky claws on. So my available evidence of red mulber-
ries on the ground seems to support the accuracy of mystory.
But is accuracy such a malleable standard? Surely, even my garden
hedgehogs would have to respond to some mundane questions: Do they
come out every night? From 1 a.m. or from 2 a.m.? What about the purple
spots on the ground that look like old juiciest mulberries getting moldy?
Accuracy (be it the accuracy of a measurement or the accuracy of a scientific
claim) comes always in tandem with other standards, such as consistency,
6. Here a relativist may come to the fore and make this kind of rejoinder. (One is
reminded of the familiar story about the Azande and their witchcraft and how stan-
dards of adequacy vary from one epistemic community to another. See Kusch 2002 for
a helpful discussion.) Adiscussion of relativism would lead me into territory farther
afield from the topic of my essay here, and as such Iwill not pursueit.
B r i n g i n g R e a l R e a l i s m B ac k H o m e [111]
performance adequacy in its own scientific perspective (and even more so
when assessed from other perspectives). In other words, it is not the case
that p is false and nonetheless meets standards of performance adequacy
in a given perspective. Ascientific claim of this kind would not satisfy the
criterion of success from within.
A different kind of worry may be raised at this point. Isnt the truth of
the propositional content p enough to secure success from within? Arent
the standards of performance adequacy themselves idle wheels, not neces-
sary to secure success? Here a different critic is envisaged, who may retort
that a real realists working posits ultimately underpin the truth of p, and
my definition of success from within collapses onto the real realists success
from above at a closer inspection. The critic may insist that Fresnels theory
worked and proved successful not because it met standards of performance
adequacy in Fresnels time (e.g., it was fruitful in predicting novel phenom-
ena; it seemed accurate in explaining polarization by reflection; and so
forth) but because Fresnels light wave referred to electromagnetic waves
of high frequency. Or better, Fresnels theory met those standards because
its working posits (i.e., electromagnetic waves) were true. And to empha-
size the idleness of the standards of performance adequacy themselves,
one could easily invoke consistency with the ether theory as an example.
(Yes, Fresnels theory was consistent with popular ether theories at the
time, yet consistency in and of itself does not cut any ice for the success of
Fresnels theory.)
In reply one may consider what would happen to a lone researcher who
gets it right without yet meeting the standards of performance adequacy
of her community at the time (perhaps because such community has not
quite gotten to the stage of precisifying standards able to capture the truth
of what the lone researcher has just discovered). Should we conclude that
the researcher has been successful? Here Icannot help but share Richard
Boyds negative answer to this question,7 although I give a perspectival
gloss to what he portrays as the social dimension of scientific inquiry. That
p is true is not sufficient by itself to ontologically ground success, unless p
also meets standards of performance adequacy at SP1 when assessed from
another scientific perspective.
Consider the astronomer V.M. Slipher, who, at the Lowell Observatory
in Arizona throughout 191217, was able to measure with precision the
radial velocity of galaxies and to empirically establish that galaxies were
7. Boyd (2010, 21718) describes the hypothetical scenario of the lone researcher
who gets it right, but she does not make any contribution to the reliability of our sci-
entific practice unless her success is also recognized as such by a community.
8. Here Idraw on John Peacocks account of this episode in Massimi and Peacock
(2014).
B r i n g i n g R e a l R e a l i s m B ac k H o m e [113]
E NVOI
6.
Fifteen years after Real Realism we are all heirs of Galileos strategy. We
learned from Kitchers real realism how to tell truth from falsehood, how
to discern bits that work from idle wheels, and most of all how to believe in
the reliability of the deliverances of our instruments. Empiricists and con-
structivists of all stripes owe us an argument for maintaining a justifiable
degree of skepticism about science and its success. More to the point, they
owe us an argument for justifiably retreating into metaphysical never-
never-lands on the face of so many homely arguments for being realists
about the things with which we interact all thetime.
For myself and for my generation real realism has enticed us to explore
new avenues and encouraged us to appraise success and failure across the
history of science in a careful way. Maybe success from above should leave
room for success from within. We stand on Galileos shoulders by acknowl-
edging our continuity with the past and our ability to assess past scientific
claims by their own lights and from our current vantage point, a vantage
point that is neither metaphysically nor epistemically privileged. That is
how, in my view, a perspectival slant can help us bring real realism back to
the Kantian home, to which it naturally belongs.
ACKNOWLEDGMENTS
Reply toMassimi
PHILIP KI TCHER
B r i n g i n g R e a l R e a l i s m B ac k H o m e [115]
I reply by trying to specify the exact claims at issue. In the Aristotelian
tradition the terms we translate as free fall pick out the motions of bodies
after their release as they tend toward the earths surface. Some contexts
will fix the reference of those terms by importing theoretical ideas from
Aristotle; for example, free fall will be taken to be a motion toward a natural
place. On other occasions, however, thinkers in the Aristotelian tradition
refrain from any such theoretical imposition:free fall is just the downward
motion of a body that has been released. Call these the strong and the
weak senses, respectively. Consider now the statement Bodies in free fall
accelerate as they approach the earths surface. When Aristotelians make
claims we properly translate by this statement (and, of course, they dont
use accelerate!), everything turns on the sense we assign to free fall. In the
strong sense, if the logical form is taken to be a universal generalization,
the claim is vacuously true, since nothing satisfies the antecedent. In the
weak sense, however, the claim is a true empirical generalization, one that
might be used to make some modest predictive successes. On this read-
ing Aristotelians are claiming (correctly) that, when released, heavy bodies
gain velocity in their motion toward the earth. Its thus possible for real
realism, in its original form, to endorse parts of the Aristotelian view as
correct and as predictively successful in consequence of their correctness.
The approach just sketched appears to make progress with a problem
Massimi raises for me, but it remains historically crude, unable to make
sense of the subtle conceptual shifts that lead from Hipparchus through
Buridan to Galileo (early and mature). But before Ioutline some measures
for refining the historical treatment, introducing a form of perspectivism
akin to Massimis, its important to see how the problem of understand-
ing the conceptual shifts is different from that of responding to Laudans
actual skeptical argument.
Historians, and historically sensitive philosophers, often writhe when
they encounter judgments about the truth of particular claims made in
the scientific past, especially when that past is quite distant. By what right
do we adopt a view from nowhere, judging the correspondence of (say)
Aristotles words with nature? The answer is that real realists dont adopt
a view from nowhere but a view from right here, and they do so because its
forced on them by the skeptical challenge. Laudan and the antirealists who
follow him introduce that view. Heres the challenge:You realists believe
that success indicates truth. But there are many past successful scientific
theories that you regard, by your own lights, as false, so success cant reli-
ably indicate truth. Evaluating bits and pieces of past science as correct
from the perspective of contemporary science is an intrinsic part of any
response to Laudans celebrated argument.
B r i n g i n g R e a l R e a l i s m B ac k H o m e [117]
remarkably blunt. Following Kuhn (especially 2000), we might declare the
central claims of Aristotelians to be literally untranslatable. Id prefer to
put the Kuhnian point differently:we can arrive at approximative transla-
tions, highly context-dependent and requiring preliminary glosses to show
the ways Aristotelian and modern terms cut across one another. Neither
the Kuhnian version nor my preferred alternative vitiates the basic point
that Aristotles successors achieved an insight we can best capture with the
formulation Bodies in free fall accelerate as they approach the earths sur-
faceand its precisely by attributing that insight to them that Laudan-
style skepticism is answered (in the particular instance).
Real realism needs extension, in my view, because it should appreci-
ate the limitations of the translations advanced in combating skepticism.
Massimi is right to suppose that, through a sequence of theoretical devel-
opments of a large perspective (say, Aristotelianism), the more immediate
descendants are more able to reconstruct and explain the ideas of their pre-
decessors than those who come later:once the working posits and idle wheels
have been identified (by us!), Buridan is a better interpreter of the terms
used by Hipparchus in characterizing them than is Galileo, and Galileo, in
turn, does better than we can. If we can attribute a core insight about accel-
eration toward the earth (as Ithink we can), we must also recognize that
the Aristotelian terms in which that insight is expressed are alien to us
and that they are less strange to those who are closer to the Aristotelian
worldview.
Recognition invites the idea of a sequence of perspectives, distinguished
by ways of conceptualizing the phenomena (in this instance, phenomena
of motion), in which close successors are better able to capture the claims
of their predecessors. To rebut Laudans skepticism its enough to show
that where theres success theres an underlying use of correct ideas, often
expressed in what much later scientists see as highly peculiar ways. A much
deeper understanding of the phenomena of success, and how its won,
requires the historian-philosopher-of-science to reconstruct the perspec-
tives of the past, tracing their continuities with the present.
I read Massimi as aiming to avoid any privileging of the contemporary
standpoint, the view from here. For me, the principal motivation for per-
spectivism stems from the shortcomings of the translations used in recon-
structing the successes of past science. In the rest of this response Ill trace
a route to my preferred version of perspectivism.
Massimi draws from Kant (as Ionce did). My sources these days are the
classical pragmatists. On my interpretation of Peirce, James, and Dewey,
none of them rejects my favorite (post-Tarskian) version of correspon-
dence truth for scientific statements (see Kitcher 2012c, ch. 5). Yet James
B r i n g i n g R e a l R e a l i s m B ac k H o m e [119]
The neopragmatist perspectivism Ihave sketched attempts to take up that
challenge.
In the end, Ithink, a perspectivalist real realism is doubly motivated
and we dont have to choose which rationale is more important. Massimi
views Laudans use of evaluation from above as betraying one of his cen-
tral insights; thus she introduces a more probing account of success and
elaborates real realism with respect to it. Ihave been more troubled by a
Kuhnian challenge. But we come out in much the same place. This is not
so much because of the resurgence of my lapsed Kantianism as through
celebration of a characterization Kuhn came to relish:he was redoing Kant
with movable categories. So too, Ibelieve, were the classical pragmatists.
Real realism should continue the enterprise.
Unificationism, Explanatory
Internalism, and Autonomy
JAMES WOODWARD
1.INTRODUCTION
1. Talk of levels has come in for a good deal of well-deserved criticism recently. My
view is that there is a relatively innocuous way of understanding this notion:think of
it as a way of capturing the idea that certain factors (within some range of variation) do
not make a difference to other factors or relationships. When this is the case the latter
can be regarded as at a different level than the former. It is this understanding that
Iadopt in this essay. Levels sometimes but by no means always track differences in the
spatial or temporal scale at which processes occur; seebelow.
U n i f i c at i o n i s m , E x p l a n at or y I n t e r n a l i s m , a n d Au t o n o m y [123]
or disunified from the point of view of molecular biology are treated in a
much more unified fashion in classical genetics, with the same argument
patterns, formulated in the vocabulary of that theory, being used repeat-
edly to derive a range of different results. This unified pattern would be lost
if we relied solely on derivations from molecular biological premises.
U n i f i c at i o n i s m , E x p l a n at or y I n t e r n a l i s m , a n d Au t o n o m y [125]
plays a central role. (We might say that the guiding focus on EU1 is relevance
or dependence, and that of EU2 irrelevance or independence.) In one very
common kind of case, EU2s explain or demonstrate or at least make use of
or exploit the independence of various upper-level relationships that figure
in the special sciences from lower-level microdetails about their realizers;
they show or make it understandable why those upper-level relationships
turn out to be stable or invariant across various changes or variations in
other sorts of factors, including those involving microdetails. To employ
an illustration discussed in more detail below, renormalization techniques
explain (in the sense of EU2) why materials of many different sorts, dif-
fering in microphysical details, exhibit similar generic behavior near their
critical points. As another illustration, the method of arbitrary functions
and its elaborations explain why gambling devices of different design and
material composition exhibit similar behavior with respect to the relative
frequencies they exhibit. Kitchers claims about the irrelevance of (many
of) the underlying molecular details to the generalizations of classical
genetics (independent assortment of nonhomologous chromosomes, etc.)
can, Ibelieve, be naturally assimilated to cases of this sort. Because of this
focus on the irrelevance of microdetails, EU2 projects are often bound up
with antireductionist themes about the relative autonomy or independence
of the relationships that are the subject matter of the special sciences. Such
independence can enable or make possible theorizing that seems correctly
describable as having a unificatory aspect or feel to it, since it involves
generalization across or abstraction from irrelevant microdetails. However,
the focus of this sort of unificatory achievement seems different in impor-
tant respects from what is achieved in EU1 projects and, I will suggest,
involves features that are perhaps not so well captured by Kitchers offi-
cial theory of unification, although they are fairly well captured by various
other, more informal observations ofhis.
3.2
U n i f i c at i o n i s m , E x p l a n at or y I n t e r n a l i s m , a n d Au t o n o m y [127]
3.3
3. Or, more weakly, we need not cite these non-difference-making details, and we
should not represent them as difference-making when they arenot.
U n i f i c at i o n i s m , E x p l a n at or y I n t e r n a l i s m , a n d Au t o n o m y [129]
capacitance across the neural membrane, the existence of physically sepa-
rated voltage-dependent ionic currents across the membrane with differ-
ent time courses, and so on. Any neuron with this circuitry conforming to
the differential equations characterizing the H-H model will generate an
action potential under the appropriate conditions, independently of such
matters as the particular ions making up the ionic currents or the particu-
lar molecular mechanism involved in the transport of those ions. These
latter factors are not difference-makers (given that the neuron has the
generic features described above) if what we want to explain is the over-
all shape of the action potential. On the other hand, if what we want to
explain is the opening and closing of particular ion channels and the fac-
tors affecting the transport of ions through them, such molecular details
are relevant difference-makers.
3.4
U n i f i c at i o n i s m , E x p l a n at or y I n t e r n a l i s m , a n d Au t o n o m y [131]
explanation should be externalist and criticizes both Hempel and Kitcher
for providing purely internalist models of explanation.
When initially encountered, the internalist/externalist (or epistemic/
ontic) contrast can seem puzzling or at least not entirely perspicuous, since
it is clear that supposedly internalist models like Hempels and Kitchers
also contain commitments that look ontic. For example, on Hempels ver-
sion of the DN model, the requirement that the explanans be true requires
the holding of various external, worldly facts, as Kim himself recognizes.
Moreover, on a natural construal of Hempels views, there must be a corre-
sponding external relation in the world in which the explanandum event is
subsumed under or instantiates the regularity described in the explanans.
Similarly in connection with Kitchers model, although it is true that
which argument patterns are instantiated by a derivation, how unifying
these are, and the stringency of the derivations are matters internal to an
investigators corpus of beliefs, Kitchers model also requires that the deriva-
tion itself appeal to true premises and that some appropriate subsumption
relation be present. It is hard to see how to make sense of these require-
ments without supposing that facts about the way the world is constrain
which are the most unifying and stringent derivations. I might undertake
to construct a theory that unifies true propositions from the theory of juve-
nile delinquency, the astrophysics of the early universe, and the molecular
genetics of C. elegans, but given the regularities that actually obtain in the
world, and the fact that the explanatory generalizations to which my pur-
ported unification appeals need to reflect these regularities, the resulting
theory is unlikely to score very high along the dimensions of successful uni-
fication emphasized in Kitchers model. Similarly one would think it is facts
about the world and the nature of the gravitational force that make it pos-
sible to construct a theory that unifies, according to the criteria described
by Kitcher, the motion of terrestrial and celestial bodies. To the extent that
this is so, why shouldnt we think, contra Kim, of Kitchers theory as having
an externalist as well as an internalist component?
Although these observations seem correct, as far as they go, there is an
important insight behind the distinction that Kim and Salmon are attempt-
ing to draw and that has important implications for Kitchers proposals
about unification. One way of bringing this out is to ask the following
question about a model of explanation:When it comes to characterizing
the explanatory relation(s) R, which (if either) is primary and which is
derivative(i) internal (e.g., deductive) relationships among propo-
sitions or (ii) external or worldly relations? Over and above any truth
requirement we impose on the explanans, can we characterize the explana-
tory relationship R (just) in terms of (i), or does the characterization of R
U n i f i c at i o n i s m , E x p l a n at or y I n t e r n a l i s m , a n d Au t o n o m y [133]
The internal relations between X and Y (whatever they may be) do not auto-
matically provide a correct answer to this question.
As an illustration that is particularly relevant to both Hempel and
Kitcher, suppose that we are presented with a derivation of some explanan-
dum from premises specified in a candidate explanans, where the deriva-
tion has a DN structure (the derivation is deductively valid, the premises
are true, at least one is a law essential to the derivation, etc.) We may then
ask, within an interventionist framework, whether the (nonnomic) fac-
tors cited in the candidate explanans are such that there are interventions
on those factors that would change the explanandum phenomenon in the
way described by the nomic premises in the derivation; if so, the deriva-
tion can be thought of as tracking or representing the difference-making/
dependency relations between the factors cited in the explanans and the
explanandum phenomenon. However, whether the derivation satisfies this
interventionist requirement depends on what nature is like, and this is not
settled just by whether the derivation meets the official DN requirements.
For example, a derivation running from the height h of a flagpole and the
angle of the sun on the horizon to the length s of the shadow it casts
identifies factors such that interventions changing the value of those fac-
tors will change the value of s; in this sense the derivation can be thought
of as providing information about how the value of s depends on the value
of h and and as demonstrating that (and how) h and are difference-
makers for s. However, a parallel claim is not true regarding a derivation
of the value of h from s and a point that can be established by, for
example, varying (or observing its variation) and seeing whether there
are changes in the value of h.4 Within an interventionist framework this
4. There is a good deal more to be said about the elucidation and identification of
explanatory asymmetries. One consideration is that when one gets the direction of
explanation wrong, this is often reflected in the apparent presence of unexplained
coincidences or correlations that are not present when one gets the direction right.
(To put the same idea in a slightly different way, one gets violations of the requirement
that, in the absence of some special reason for supposing otherwise, the independent
or cause variables in a purported explanatory relationship should be capable of varying
independently of each other and should not exhibit any particular stable correlation.)
In the example above, if one alters the value of (e.g., by tilting the angle between the
pole and the ground) or even just observes the naturally occurring variation over the
course of the day, one will observe a corresponding change in the value of s that occurs
in such a way that the value of h appears to be constant; that is, the values of s and
are correlated, adjusting in just the way that is required to maintain the same value
for h. Furthermore the envisioned explanation provides no explanation of this cor-
relation. By contrast, when the direction of explanation is from h and to s, no such
mysterious correlation is present:the independent variable h is a constant and hence
is uncorrelated with any variations in . Related procedures are used in statistics and
machine learning to identify causal direction in nonexperimental contexts.
U n i f i c at i o n i s m , E x p l a n at or y I n t e r n a l i s m , a n d Au t o n o m y [135]
dependency relations in many areas of science, both in the social and
behavioral sciences and in the biological sciences. In such graphs an arrow
drawn from one variable to another (XY) represents that Y depends in
some way on X, but without specifying the exact functional form or param-
eterization of the dependence. For certain explananda this may be all that
is needed for successful explanation.
U n i f i c at i o n i s m , E x p l a n at or y I n t e r n a l i s m , a n d Au t o n o m y [137]
For example, this feature is also present in Carl Cravers (2014) recent
defense of an ontic approach to explanation, even though he does not adopt
Salmons specific view of what the ontic involves. For Craver the relevant
ontic facts in the case of biological explanations are apparently (or at least
usually) characterized in molecular/chemical terms. As an illustration,
Cravers version of an explanation of the generation of the action poten-
tial (the spiking activity of a single neuron), which is in the ontic mode, is
the following: The flux of sodium (Na+) and potassium (K+) ions across
the neuronal membrane explains the action potential (31). This is favor-
ably contrasted with models of the generation of the action potential such
as the H-H model, which abstracts away from the molecular details of the
processes by which ions are transported across the neural membrane and
which instead exhibits how the action potential depends on macroscopic
variables characterizing the whole neuron such as the capacitance across
the neural membrane, the total ionic currents, the membrane potential,
and so on. Craver regards the H-H model as at best an explanation sketch
or as merely phenomenological (and hence defective qua explanation)
because of its neglect of molecular detail; he associates the fact that the
generic shape of the action potential is a solution to the H-H equations for
certain parameterizations of those equations with the claim that the model
functions in DN-like fashion to show the action potential is nomically
expectable under certain conditions, but takes this to be a consideration of
merely epistemic rather than ontic significance, since nomic expectability
is an epistemic rather than an ontic notion.
Rather than illustrating the advantages of an ontic approach to explana-
tion, Cravers discussion illustrates the disadvantages of neglecting the epis-
temic or internal dimension of explanation. One relevant consideration is
that different types of neurons exhibit different firing patterns in response
to synaptic input; some exhibit the characteristic firing pattern associated
with generation of the sort of action potential described by the H-H model,
and some do not. Whether a neuron exhibits this firing pattern does not
depend just on whether there are fluxes of Na+ and K+ ions across the cell
membrane but rather on the way these are organized (that is, on the circuit
diagram for the whole neuron, including the physical separation of the Na+
and K+ channels and the fact that these operate according to different time
courses, the fact that the cell membrane is sufficiently insulated to act as a
capacitor, and so on).5 Aneuron that is not organized in this way will not
5. Kitchers discussions in his 1984 and 1999 papers also emphasize the importance
in explanation of considerations having to do with system-level spatial organization
that are often neglected in philosophical defenses of reductionism.
U n i f i c at i o n i s m , E x p l a n at or y I n t e r n a l i s m , a n d Au t o n o m y [139]
on it merely reflects our parochial intellectual limitations. On this view of
the matter the underlying reality is that the action potential is the product
of facts about the component molecules and atoms making up the neuron
and their relationships, and it is this underlying reality and nothing more
that is relevant for the purposes of explanation. What we can calculate or
measure is irrelevant.
In my view this objection draws on the mistaken idea that internalism
and externalism about explanation are mutually exclusive alternatives and
that the internal features of explanation are entirely disconnected from the
external onesideas rejected above. Although it is true that we lack the
computational ability to derive facts about the behavior of a macroscopic
sample of gas from facts about the trajectory of individual molecules, it
is also true that it is a fact about the world (and not a fact about our com-
putational limitations) that many aspects of the macroscopic behavior of
a dilute gas can be captured by means of relationships among a few mac-
roscopic parameters that are to a very large degree independent of (stable
across) variations in molecular details. Similarly for models of neuronal
behavior that abstract away from molecular details. Thus these models
capture ontic (or worldly or external) facts about macroscopic dependency
relations in these systems. The correct way to think about the relationship
between these ontic facts and epistemic considerations is that the ontic
facts provide us with computational and derivational opportunities that are
then reflected in epistemic or internal features of the models and repre-
sentations we construct.6 In other words, it is because of the existence of
certain stable upper-level patterns (ontic facts) that we are sometimes able
to construct tractable models (reflecting internalist considerations) of the
behavior of the systems that interest us. But this does not mean that these
stable upper-level patterns somehow spring into existence just as a result
of our interests or our computational limitations or as a result of our cogni-
tive organizing activities, or that they are mere projections of our inter-
est in finding tractable models with epistemically pleasing features. Rather
we find or discover preexisting relationships in the world that fit with and
enable the application of our limited cognitive and calculational abilities in
the construction of explanations.
On this view of the matter there are lots of difference-making and
dependency relations in nature; these occur at different levels or scales
U n i f i c at i o n i s m , E x p l a n at or y I n t e r n a l i s m , a n d Au t o n o m y [141]
magnets near their critical points, both characterized by the same criti-
cal exponent b. In the case of fluids, for example, behavior near the criti-
cal point can be characterized in terms of an order parameter S given by
the difference in densities between the liquid and vapor forms of the fluid
S=liq vap. As the temperature T of the system approaches the critical
temperature Tc, S is found to depend upon a power of the reduced tem-
perature t= TTc/T:
S~|t|b
where b is the critical exponent referred to above. Remarkably the same
value of b characterizes not just different fluids but also the behavior of
magnets in the transition from ferromagnetic to paramagnetic phases.
Suppose one is interested in explaining why some particular kind of fluid
has the critical point that it does. Since different kinds of fluids have differ-
ent critical points, the value of Tc for any particular fluid will indeed depend
on microphysical details about its material composition.7 However, if one
is instead interested in explaining the universal behavior just described
(the phenomenon or generic fact that S ~ |t|b with fixed b for many differ-
ent materials), then information about the differing microphysical details
of different fluids is irrelevant:within the framework for thinking about
explanation defended above these details are non-difference-making fac-
tors. In other words, the universality of this behavior shows us that its
explanation must be found elsewhere than in details about the differences
in material composition of different fluids. Instead the explanation for this
universal behavior is provided by renormalization group techniques, which
in effect trace the behavior to very generic qualitative features (e.g., certain
symmetries) that are shared by the Hamiltonians governing the interac-
tions occurring in each of the systems, despite the fact these Hamiltonians
differ in detail for each system.
In this case we have a kind of unification since we are shown why a
variety of very different systems exhibit a common or unified pattern of
behavior near their critical points. Isuggest, however, that the kind of uni-
fication achieved seems somewhat different from the sort of unification
(EU1) that is achieved when a number of (apparently) different phenomena
are attributed to the same general type of causal factor. In the case of EU1
we begin with a variety of apparently different phenomena (the orbits of
7. This illustrates the notion that the autonomy of upper-level behavior in such sys-
tems is only partialholding with respect to some explananda but not others.
U n i f i c at i o n i s m , E x p l a n at or y I n t e r n a l i s m , a n d Au t o n o m y [143]
Depending on the system under investigation, there are many differ-
ent reasons (in addition to those operative in the cases mentioned above)
why various factors (including facts about microstructure) may be irrel-
evant to overall patterns in their behavior. In a large and important range
of cases the irrelevance of certain factors or processes for certain depen-
dency relationships follows from considerations having to do with the
differences among the spatial or temporal or energy scales that are rel-
evant to the behavior of those factors. For example, a process or influence
may either occur so quickly (in comparison with the dependency relation-
ships in which we are interested) or so slowly that we may safely regard it
as irrelevant. Or the influence may operate at length or energy scales that
make it irrelevant to the phenomena we are trying to explain, as when the
details of the behavior of the strong and weak force (which are very short-
ranged) are justifiably ignored in explaining chemical behavior. In particle
physics, processes operative at very high-energy scales are thought to be
irrelevant to many processes operative at lower energy scalesirrelevant
in the sense that many different alternative high-energy theories are con-
sistent with the same low-energy behavior, so that variations in these
make no difference for low-energy behavior. (Various decoupling theo-
rems provide results about the extent of this independence.) This fact
makes particle physics, as currently practiced, possible, enabling the con-
struction of so-called effective theories since really high-energy behavior
is (currently) unobservable. (If finding an adequate low-energy theory
required identifying which high-energy theory is correct, physics would
be stuck.)
Similar considerations (with separations of scale motivating claims
of independence and irrelevance) are very likely operative in connection
with many biological phenomena, although there has been less systematic
exploration of such cases than in physics. For example, because biologi-
cal processes occur on quite different time scales, it is sometimes possi-
ble to treat processes that are slow relative to the process one wants to
understand as approximately constant, hence warranting the assumption
that there are no actual variations in the slow process that are relevant
to the faster process. Similarly it is sometimes reasonable to assume that
certain processes occur very quickly and reach a steady-state equilibrium
relative to some process of interest; again this justifies treating the former
as approximately constant (Voit 2013, 10). Alon (2007, 1011) provides
illustrations:inputs change the activities of gene transcription factors on
a subsecond scale; in contrast binding of the active transcription factor to
its DNA site reaches equilibrium in seconds. Transcription and translation
U n i f i c at i o n i s m , E x p l a n at or y I n t e r n a l i s m , a n d Au t o n o m y [145]
from showing that Mendels laws figure in the derivation of a variety of dif-
ferent phenomena, which is an EU1 project.
I conclude with a related observation: I suggested earlier that reduc-
tive explanations seem to be naturally viewed as cases of EU1 or as in part
motivated by the aspiration to construct EU1s. For example, the reduction
of the thermodynamics of dilute gases achieved by statistical mechanics
makes use of the laws of Newtonian mechanics, which of course figure in
the EU1 of many other phenomena. Similarly many of the physical and
chemical generalizations employed to explain aspects of the behavior of
biological systems can also be used to provide EU1 explanations of the
behavior of inorganic systems. Thus if we should prefer those EU1s that
best satisfy Kitchers criteria for explanatory unification, it is not obvious
that we can resist the contention that we should prefer reductive explana-
tions that appeal to statistical mechanics over those that appeal to phe-
nomenological thermodynamics or explanations of biological phenomena
that appeal to more fundamental principles of physics and chemistry over
those that do not. For this reason appeals to EU1 do not seem to be an
entirely convincing way of defending antireductionist theses about biology
or the other special sciences. Isuggest that a better strategy is to appeal to
the sorts of considerations that underlie EU2s:the behavior of many of the
systems that are the subjects of the special sciences simply do not depend
on the factors that figure in underlying microtheories of those systems. To
the extent this is the case, we have all of the justification we need for treat-
ing the sciences of those systems as relatively autonomous.
Reply toWoodward
PHILIP KI TCHER
U n i f i c at i o n i s m , E x p l a n at or y I n t e r n a l i s m , a n d Au t o n o m y [147]
How did philosophy of science get sucked into thinking that there must
be some single type of explanation (explaining why) governed by a single
pattern of successful explanation? The answer lies deep in the Humean past
of logical empiricism. One great attraction of Hempels covering-law model
lay in its avoidance of any suspicious reliance on a prior notion of causation.
Indeed logical empiricists often wanted to reconstruct causal talk by iden-
tifying the causes with the factors identified in successful explanations. By
the 1970s it was evident that the Hempelian attempt to characterize scien-
tific explanation had broken down. Those of us who continued to harbor
Humean scruples cast about for new ways of characterizing explanation
independently of causationsometimes explicitly thinking of causation
as the projection of explanatory structure onto the world (Kitcher 1986).
Unification seemed to me an especially attractive possibility for three main
reasons:it promised to provide an account of the theoretical explanation
(systematization) of laws; it resonated with the ways many prominent sci-
entists defended their explanatory proposals; and it offered an embryonic
account of the explanatory relations within mathematics, a topic on which
Ihad made some timid forays (Kitcher 1975; for bolder work in this area,
see Steiner 1978, and now Lange forthcoming).
I tried to explicate unification in terms of the repeated use of patterns
of argument to generate a broad set of conclusions. There is something to
this idea, and Iwould still claim that it offers a useful way of reconstruct-
ing some parts of theoretical science, offering a hybrid between the so-
called semantic conception of theories and the idea of a family of kindred
methods of problem solution that lies at the heart of Kuhns much-abused
notion of a paradigm. But my efforts to specify the notion of unification so
as to solve problems of causal asymmetry led to difficult (probably intrac-
table) issues of defining technical notions. Fortunately conversations with
Nancy Cartwright dislodged my Humean scruples, and Ibegan to see the
notion of unification for what its worth:a useful device for some exercises
in philosophical reconstruction of scientific theories (e.g., Darwins) and a
virtue possessed by some, but by no means all, explanations.
So far a confessional preamble to taking up Woodwards major points.
Woodward distinguishes two notions of unification. The first, the concept
Iattempted to characterize, sees unification as consisting in showing how
apparently different phenomena depend on a small number of explanatory
factors. The second, which Woodward takes to figure in some of my antire-
ductionist arguments, consists in establishing that some facts are inde-
pendent of certain others. So, for example, in defending the autonomy of
classical genetics, Iattempt to show that the transmission of genes at loci
U n i f i c at i o n i s m , E x p l a n at or y I n t e r n a l i s m , a n d Au t o n o m y [149]
traditions are local, specific to the major branches of the sciences or to sub-
fields within them, or even to subfields of subfields. Local unification is
compatible with a lot of difference and disorder across the space of areas of
inquiry. My explication of Kuhns important point, and my endorsement of
local unification, can coexist with the Cartwright-Dupr thesis of disunity.
Unification is a regulative ideal:enjoy it where you can find it, but dont
suppose that it can be achieved on any global scale. Woodward sides with
me against Cartwright and Dupr, but Isee no need to take sides.
(2) Woodward cites my discussion of the supposed reduction of clas-
sical genetics to molecular biology as exemplifying his second type of
unification. He is quite right to recognize that central to my antireduction-
ist thesis is a claim that the details of the molecular structures and rear-
rangements are irrelevant to the patterns of transmission, specifically to
the independence of assortment of alleles on different chromosome pairs.
Iwant to suggest, however, that unification in the first sensethe sense
Ihave attempted to explicateplays an important supporting role in the
judgment of irrelevance. Precisely because we have a unifying account of
the transmission of genes on nonhomologous chromosomes, we see that
the underlying molecular details dont matter.
Imagine our predicament if we had no such explanation, either in this
case or in the comparable case of the sex ratio at birth. Under these circum-
stances wed be in the predicament often affecting medical researchers who
study complex diseases: in this instance, the etiology involves this com-
bination of factors, in that something quite different, and so on and on.
Without some explanation to bind them together, wed view the apparent
regularity as a giant coincidence, something that comes about on the basis
of quite diverse causal antecedents and that might well break down as we
sample further. Unless, of course, like Dr.Arbuthnot, the first observer of
the preponderance of male births, we were ready to chalk the whole thing
up to divine providence.
So, I suggest, its the unifying power of seeing gene transmission as
a type of pairing-selection process or of understanding birth sex ratios
from the perspective of Fishers evolutionary argument that generates the
judgment that the lower-level factsthe gory detailsare irrelevant.
Unification of the first kind warrants our ascribing unification of the sec-
ond kind. (See (4) for Woodwards potential response to this.)
(3) As Ioriginally proposed it, Itake the unification approach to explana-
tion to be internalist in exactly the same way Hempels covering-law models
were. Hempel requires true covering laws; Idemand that the instantiations
of the unifying patterns be true statements. Equally Hempel avoids any
ontic commitment to causal relations by supposing that explanatory
U n i f i c at i o n i s m , E x p l a n at or y I n t e r n a l i s m , a n d Au t o n o m y [151]
encouraged me to think that unification is important in explanation) but
also to the precepts issued by philosophers:Hume, recall, appends a list of
rules by which we should judge causation, and, in my reading at least, they
endorse unification as a regulativeideal.
CONCLUSION
It may seem that Ihave treated Jims probing essay in a highly selective
way, picking up a question here and another there. But in the spirit of my
opening report on how my views have evolved, Iwant to propose that he
and Iare both now engaged in a different type of philosophical project with
respect to explanation. The name of the game is no longer to enunciate
the Final and Complete Analysis of Explanation in All Forms. Rather the
philosophical task is to develop and refine some tools for specific recon-
structions and analyses of bits and pieces of explanation across the range
of diverse sciences. As I read Jims rich discussion, hes recognizing that
both difference-making and unification might be valuable tools (concepts),
good for working on different problems, or even in combination on some
occasions. Hes showing some ways those tools might be put to use. Ive
tried to sketch a little in the samevein.
1. Classical genetics does not contain the kind of general laws required by
Nagels (1979) canonical account of intertheoretical reduction.
2. The principal vocabulary of classical genetics cannot be translated into
the vocabulary of lower-level sciences; nor can the vocabularies be con-
nected in any other suitable way (that is, by bridge principles).
3. Even if the reduction were possible it would not be enlightening, because
once you have the cytological explanation of genetic phenomena, the
molecular story adds nothing of further interest.
This essay takes issue with the third of these arguments, contending that a
robust explanatory reductionism can coexist with the sort of explanatory
1.ANTIREDUCTIONISM FROMAUTONOMY
Kitchers argument from autonomy in 1953 and All That turns on the
explanation of an enhanced version of Mendels secondlaw:
2. On the prehistory of the explanation in the work of Darwin and later writers, see
Edwards (1998).
Sp e c i a l - S c i e n c e Au t o n o m y a n d t h e Di v i s i o n of L a b or [155]
According to Fishers story, the sex ratio has become fixed at 1:1 because
the even ratio is a stable and unique equilibrium. And equilibration occurs
at 1:1 because, in a population with more females than males, individuals
with a propensity to produce more males than females will have a higher
expected number of grandchildren and vice versa. Why a higher expected
number of grandchildren? Your expected number of grandchildren is pro-
portional to your expected number of children and your childrens expected
number of matings. Since matings require exactly one male and one female,
a males expected number of matings will increase, relative to a females, as
the proportion of males in a population decreases.
If explaining an explanations explainers deepens the original explanation
if explanatory relevance is transitivewhy not go further? Why not detail
the biological mechanics in virtue of which, for example, successful repro-
duction always and only involves a single male and a single female?
Here Kitcher might have said: As with molecules in the Mendelian case,
I draw the line at these gritty goings-on. As long as you see that procre-
ation, if not copulation, is strictly one on one, it is irrelevant whether the
protagonists are made of meat or Swiss cheese.
In fact he is somewhat more circumspect, more choosy about the ingre-
dients of the explanatory sandwich. Certain details of implementation are
worth investigating, he remarks. Presumably he allows that even quite
low-level details might appear in an explanatory extension of Fishers
model. But they do not, merely because they extend the model, count as
explanatorily relevant to the things that the model explains, such as the
sex ratio, and if they are sufficiently low level they certainly do not count
as relevant.
Indeed in the sex-ratio case and almost everywhere else, Kitcher appears
to hold the following view:moving down the levels of potential explanation
from ecology to physiology to cytology to chemistry to fundamental phys-
ics, there is some point beyond which further unpacking of mechanisms
becomes entirely irrelevant. Thus, for example, deriving the one-on-one
nature of procreation from quantum mechanics adds nothing whatsoever
to our understanding of Arbuthnots observation. Thanks to some aspect
of the nature of scientific explanation, the facts of implementation cease to
explain when the scale reaches so fine a level of description.
What explanatory principle is it that undercuts the transitivity of
explanatory relevance, severing the link in certain cases between the
explainers of explainers and the explanandum?
In 1953 Kitcher accounts for failures of transitivity by proposing that,
when transitivity falls through, it is because categories essential to explain-
ing high-level phenomena cannot be ascribed explanatory relevance by
Sp e c i a l - S c i e n c e Au t o n o m y a n d t h e Di v i s i o n of L a b or [157]
that the correct story about an explainers implementation does not always
contribute to the explanation. Let me simply take it as given that Kitchers
claims about relevance are correct. It is a fact, Iwill suppose, that scientists
of the high level regard much information about implementation as irrel-
evant to their researchthe sociological fact of explanatory autonomy.
Assuming that those scientists are not wholly mistaken, a philosophical
fact follows immediately: details of implementation often are explanato-
rily irrelevant, and below a certain level of description, in many or perhaps
even all cases, every detail of implementation is irrelevant.
That irrelevance looks to be flatly incompatible with explanatory reduc-
tionism. Or as Fodor (1974, 11213) memorably writes:
Reductivism... flies in the face of the facts about the scientific institution:the
existence of a vast and interleaved conglomerate of special scientific disciplines
which often appear to proceed with only the most token acknowledgment of
the constraint that their theories must turn out to be physics in the longrun.
2.EXPLANATORY REDUCTIONISM
The physical world, there is ever more reason to think, is the only world we
have. Everything is made of physical stuff, and everything that happens,
happens because of the way physical laws push physical stuff around. This
is the doctrine of physicalism. It is deniable, but ever ascendant.
From physicalism it follows that any state of affairs or pattern of behav-
ior we find in the world, no matter how high level or abstract, can be derived
from fundamental physical facts and laws. It also seems plausible (though
it does not strictly follow) that any state of affairs or behavior at a given
level can be derived from facts and regularities about entities at the next
level down. Thus economic regularities can be derived from psychological
facts and regularities, psychological regularities from physiological facts
and regularities, and so on through cytology, molecular biology, and chem-
istry to fundamental physics. (The relevant lower-level facts will concern to
a great degree the arrangement of and relationships between lower-level
entities:we do not get psychology from the study of neurons in isolation
4. Kitchers skepticism about bridge principles for general categories such as gene is,
as he himself notes, compatible with the derivability of all observed behaviors of genes,
as the behaviors may be derived piecewise, that is independently for particulargenes.
5.If particular matters of fact appear in the explanation, it might be further
improved by tracing back in time the genesis of those facts. But this proposal (Strevens
2008, 4.31) has nothing to do with reductionism.
Sp e c i a l - S c i e n c e Au t o n o m y a n d t h e Di v i s i o n of L a b or [159]
Give this thesis a name:explanatory reductionism. An explanatory reduc-
tionist, then, holds that an explanation is always improved by giving a
lower-level explanation of its partsof the initial conditions, regularities,
and structural facts that it citesand that such further explanations are
available for every nonfundamentalpart.
How is this compatible with the sociological fact of autonomywith
the fact that working scientists consider many and in some cases all lower-
level details to be irrelevant? That looks like a simple question; the answer,
clearly, is that the fact of autonomy refutes explanatory reductionism.
But no: it is a complex question, and compatibility is possible after
all. The key is to identify more than one sense of explanatory relevance.
Explanatory reductionism is true of one sense, autonomy of theother.
To make this dichotomizing plausible, in the next section Iwill intro-
duce a reductionist account of explanationmy own kairetic theory
and flesh out the notion of explanatory relevance at its heart. Agreat deal
of lower-level detail turns out to be irrelevant to high-level phenomena.
But not all detailas must be the case, since the kairetic theory is a form
of explanatory reductionism. The discussion thus opens the door to, with-
out entirely achieving, the synthesis of reductionism with autonomy. That
goal requires the postulation of the second kind of explanatory relevance
in section4.
7. As this list makes plain, there are many ways to give a metaphysical theory of
causal influence. There is no need, in what follows, to choose amongthem.
Sp e c i a l - S c i e n c e Au t o n o m y a n d t h e Di v i s i o n of L a b or [161]
course of the disease in each victim, its transmission to others. Notionally
the process is described at a maximal level of detail:the position of every
bacterium and the disposition of every drop of water. Now take away what
you can without invalidating the descriptions entailment of the explanan-
dum, that is, of the fact of the epidemic. The precise positions of individual
micro-organisms do not matter at all; that information can be deleted,
leaving only a specification of the approximate density of the organisms
in the water retrieved from the pump. The time of day (I assume) that
the water is pumped is also irrelevant; what matters is only that a certain
amount was consumed. Likewise the course of death need not be charted
in excruciating detail; the rough facts about the degree of dehydration and
its inevitable physiological effects is enough to entail an upward step in the
statistics of mortality. Throughout this process of information removal,
nothing is added; rather, more detailed specifications of the causal web are
replaced with strictly less detailed specifications. What you are left with is a
description of the same causal web with which you began, but an extremely
abstract description: bacteria in considerable quantities leaked into the
water supply; the water was consumed by a significant number of people;
given the prevailing conditions, also specified at a high level of generality,
they went on to contract cholera.
More or less the same recipe applies to the explanation of regularities.
Why does Marss orbit around the sun conform approximately to Keplers
laws? Begin with a complete causal model for Marss orbita complete
specification of the causal influences on the planets trajectory over the
course of a Martian year, along with the relevant physical laws. Your model
will predict every minor twist and turn in Marss movement. But with or
without these perturbationswith or without the other stars, the planets,
the interstitial rubblethe model predicts Keplerian behavior. The kairetic
criterion therefore orders the perturbers removal, or more exactly, it tells
us to replace the painstaking specification of the distribution of mass with
something as abstract as possible having the same net implications for the
explanandumin this case a specification that the total gravitational force
due to objects other than the sun did not exceed some (small) upper bound.
What remain are the difference-makers: the physics of gravitation; the rela-
tive size, position, and velocity of Mars and the sun; and the aforemen-
tioned upper limit.
Note that to determine explanatory relevance it is not necessary to con-
struct complete causal models. We can see that certain things are not going
to count as difference-makers without going through the rigmarole, so it is
possible to have knowledge of difference-makers while having only a very
rough knowledge of the underlying causal web. The kairetic recipe is the
8. The question of exactly what details can be removed from a description of the
gravity mechanism is determined by the kairetic accounts cohesion constraint.
Rather than attempting to describe the workings of the constraint in this essay, Isim-
ply spell out its reductionist consequenceswhich reductionism is what matters for
the purposes of the discussion.
Sp e c i a l - S c i e n c e Au t o n o m y a n d t h e Di v i s i o n of L a b or [163]
meiosis, describing the fundamental physical basis of the molecular inter-
actions that make the whole machine work the way it does. It will, that is,
give precisely the details, and only the details, you would expect if you were
to ask a molecular biologist to explain the relevant features of the cytology
and he were in turn to ask a physicist to explain the relevant features of the
molecular biology.
A reductionist theory of explanation can, as the kairetic account
shows, prescribe very abstract explanations, but they must be at the
same time physical explanations. There, it seems, is the rub:the fact of
explanatory autonomy is simply the fact of the irrelevance of these phys-
ical explanations. Their high degree of abstraction is no palliative. The
kairetic account says that the physical structure of the spindle must be
specified, albeit not in great detail. Autonomy says it does not. Thus the
kairetic account, like all reductionist accounts of explanation, is falseso
you might conclude.
What next? One strategy is to give something like the kairetic account
a stronger criterion for difference-makingstrong enough to imply the
absolute irrelevance of all facts below a certain level (Franklin-Hall forth-
coming). Another is to find a different kind of explanatory irrelevance,
orthogonal to the irrelevance diagnosed by the kairetic criterion, and to
show that autonomy is all about that other kind of relevance. That is the
route I will take.
Sp e c i a l - S c i e n c e Au t o n o m y a n d t h e Di v i s i o n of L a b or [165]
organizers of inquiry into the world. Whereas including objectively irrel-
evant factors in an explanation is an intellectual error, including contextu-
ally irrelevant factors is a social or practical error. Nevertheless the success
of an explanatory investigation may hinge just as much on the norms of
contextual relevance as on the norms of objective relevance. Investigators
must take both equally seriously. (I should add that the norms for both
kinds of relevance govern what goes into an explanation, but they do not
govern what the explanatory investigators may contemplate. Asuccessful
explainer will of necessity have to spend rather a lot of time thinking about
non-difference-makers, just so as to be able to recognize them assuch.)
Sp e c i a l - S c i e n c e Au t o n o m y a n d t h e Di v i s i o n of L a b or [167]
certain mathematical properties when viewed in the abstract; further,
these high-level similarities have important ecological consequences. It
would be a waste of many scientist-hours to have the fish ecologists and
the mammal ecologists independently derive the consequences; better to
have a single group working on the high-level properties while the fish and
mammal specialists confine their attention to those lower-level phenom-
ena where scales and fur comeapart.
The explanatory pieces so efficiently produced are made for a higher pur-
pose, to be sewn together into complete explanations, that is, explanations
in which all objective difference-makers, whether described at the level of
physics, cytology, psychology, or whatever, are brought together into a sin-
gle explanatory model. It is one of the great glories of modern science that
we have, in many cases, the necessary materials for something approaching
this full understandingthat we have, in other words, the ability to sketch
answers to the chain of explanatory questions leading all the way down
to the fundamental level. But of course no one person is custodian of this
understanding. It is spread across the sciences, with (in the predator-prey
case) the population ecologists leading off, the cognitive ethologists then
taking up the baton, and so on, all the way to the particle physicists.
It would be madness to organize explanatory inquiry in any other way.
The creation of explanations is therefore a fragmented process; fragmenta-
tion, however, is merely a means to a unitary end that can be achieved only
by science as awhole.
The conception of explanation I have proposed marries reductionism
and autonomy. On the one hand, autonomy: the work of explanatory
inquiry is divided among many domains, each of which is not merely per-
mitted but required to black-box the explanatory models generated by
other domains. Within a domain an explanatory model is complete if it
omits only details that are either objectively or contextually irrelevant.
On the other hand, reductionism: Explanatory models within a domain
are a means to a greater end, namely explanations that replace all black
boxes with substantive models accounting for the explanandum in physical
terms, that is, relating what goes on to the fundamental-level causal web,
as the kairetic account requires. Such an explanation is complete if it omits
only details that are objectively irrelevant.
Two complementary senses of explanatory completeness figure in this
scheme of things:contextual completeness, that is, completeness relative
to the standards by which explanatory labor is divided among domains, and
what you might call absolute completeness. The ultimate end of explana-
tory inquiry is the reductionist goal of completeness in the absolute sense;
however, the means to this end, given the division of explanatory labor, is
There is a systematicity to the facts about objective relevance, that is, to the
facts about difference-making. It is not that stardust makes a difference to
the orbits of some planets but not to others, or that the vibrational modes
of a telomere make a difference to meiosis in some types of cell but not
in others:the non-difference-making, the irrelevance, is across the board.
High-level explainers may consequently ignore certain kinds of detail as a
matter of general policy; they need not treat each case on a custombasis.
Approaching the question of orbits, the celestial mechanic has good
reason in advance to abstract away from planetary constitution and the
existence of interplanetary pebbles and dust. For explanatory purposes she
need not see her system as made up of molecules at all; it is enough to see
it as made up of planetsdiscrete spherical objects of great mass and inde-
terminate composition. The facts about difference-making, then, provide
the explainer, by way of abstraction, with a high-level taxonomy of the sys-
tem in question from which certain aspects of the underlying fundamental
physics have disappeared altogether.
Likewise the cytologist is apt to think in terms of telomeres rather than
in terms of DNA sequences, letalone in terms of the individual atoms that
make up such sequences, each with its own particular position, velocity,
modes of vibration, and so on, and the population ecologist will naturally
think in terms of populations or standard types rather than actual organ-
isms in all their individuality, letalone in terms of the organisms molecu-
lar makeup.
Contextual explanatory relevance is also systematic: there are long-
lasting, across-the-board rules determining what is and is not contextually
relevant within an explanatory domain.
That may sound dubious. What is contextually relevant is determined
by the way the explanatory enterprise is parceled out among the members
of the scientific community. These allocations are nothing more than epis-
temic heuristics; why expect such matters of practical policy to be uniform
over time, from place to place, or across a discipline?
Sp e c i a l - S c i e n c e Au t o n o m y a n d t h e Di v i s i o n of L a b or [169]
If research strategies were simply a matter of fashion, systematicity
would be surprising. But given their goal of investigative efficiency, they
are strongly constrained by the world out there: there are certain divi-
sions of explanatory labor that are very efficient, and many that are grossly
inefficient.
Consider again the investigation of the internal combustion engine.
One way to divide the work of understanding the engine is for one team
to take the left half, one team the right half. The left-hand team might
find themselves, then, trying to understand the workings of the left-hand
side of a piston while black-boxing the pistons right-hand side. This is not
impossible in principle, but the black box in question, which must specify
all behavior of the right-hand side that makes a (relevant) difference to the
behavior of the left-hand side, will have to contain an extremely detailed
description of the right-hand sideso detailed that it will hardly be a black
box atall.
More generally, black-boxing is useful only when the system to be com-
partmentalized or stratified is somewhat modular, in the sense that it can
be divided into units or strata, each of which makes a difference to the oth-
ers in ways that can be specified compactly and tractably.
The sense of difference-making relevant to determining the proper level
of detail in a black boxs functional specification is not quite identical to
the objective difference-making discussed in the previous section but turns
on the same core idea and is equally objective. Here, in outline, is a kai-
retic characterization of the notion (though there are no doubt other viable
approaches to characterizing functional difference-making).10
Take some proposed division of your system into putative black boxes,
either at the same level of description (compartmentalization), or at dif-
ferent levels (stratification), or both. The aim is to determine, for any such
division, which aspects of the proposed black boxes are difference-makers
for other boxes and which are not.
To this end, associate with each proposed black box a complete func-
tional specification of the boxs inputs and outputsnot just the appar-
ently relevant inputs and outputs but every way the box interacts with its
surroundings. This specification will therefore be a comprehensive map-
ping from environment to behavior. The functional specifications, together
with whatever aspects of your explanatory model are not black-boxed,
10. Perhaps the most important difference, on the kairetic approach, between the
kind of difference-making that applies to causal factors and the kind of difference-
making that applies to functional specifications is that the latter allows for difference-
making properties that are radically multiply realizable.
Sp e c i a l - S c i e n c e Au t o n o m y a n d t h e Di v i s i o n of L a b or [171]
of explanation explains why the sciences form a mosaic of explanatory sub-
cultures, autonomous units each only peripherally and sporadically con-
cerned with the work going on nextdoor.
The reductionist can, Ihave shown, explain the fact of autonomy. But is his
explanation as good as the antireductionists? Two antireductionist argu-
ments may suggestnot.
The status of the first claim is at best somewhat murky; the second, by con-
trast, seems quite reasonable. Its rightness is palpable because the given
that locution specifies that the objectively relevant parts of the molecular
implementation are being frameworked, hence excluded from the explana-
tory picture. The same locution has no effect when dealing with objective
irrelevance:
Given that does not improve the irrelevance claim at all; indeed it gives
the impression of sloppy writing. Because is far better. Put because in
the claim about independent assortment, however, and youget:
Sp e c i a l - S c i e n c e Au t o n o m y a n d t h e Di v i s i o n of L a b or [173]
That seems malformed, indeed perverse, whether or not you are an explan-
atory reductionist.
There are two kinds of irrelevance claim, then: the kind that sounds
good with a given that rationale and bad with a because rationale and
the kind for which the situation is reversed (unless given that is heard as
meaning because). Now a sweeping claim:Take all of the antireduction-
ists claims of irrelevance for lower-level details that the kairetic theory
counts as difference-makers. Apply the given that/because test. Their
rationales will sound better with given that. Do the same with lower-
level details that the kairetic theory counts as non-difference-makers,
hence as objectively irrelevant, and the rationales will sound better with
because. That is just what this essays distinction between objective and
contextual irrelevance predicts: in the former case the lower-level facts
do something explanatorily important, but their contribution is frame-
worked; in the latter case the lower-level facts are explanatorily otiose.
The antireductionist who collapses the two kinds of relevance can make
no senseofit.
Sp e c i a l - S c i e n c e Au t o n o m y a n d t h e Di v i s i o n of L a b or [175]
ACKNOWLEDGMENTS
Reply toStrevens
PHILIP KI TCHER
Sp e c i a l - S c i e n c e Au t o n o m y a n d t h e Di v i s i o n of L a b or [177]
meeting raised so many questions that I was distracted. I was thinking
about them all the way home, and simply walked past the grocery store.
Now there are occasions on which a simple explanation of this sort can
be deepened by introducing biological causes. Perhaps (although I hope
fervently that this is not so) there have been recent changes in my brain
that manifest themselves in increased forgetfulness. Or perhaps some neu-
robiological details would explain my lack of attention to the milk short-
age:Ifailed to remember later because my earlier registration of the need
for milk was superficial. Suppose, however, that factors of this sort are
entirely absent. My forgetting is just what Itook it to be:a product of my
tendency to become absorbed in the interesting points raised in a discus-
sion and consequent absentmindedness. Under these circumstances, Isug-
gest, the descent into neurobiological detail sheds no further light on why
Ibehaved as Idid. We know all there is to know about why Iwalked past the
store rather than turning in and fulfilling my promise. (Here, Ithink, the
abstract point about contrast classes is made concrete.)
Move now to a more important type of case, one in which the explanation
traces an etiology:Why did the English win the Battle of Agincourt, even
though they were exhausted, hungry, and seriously outnumbered? Military
historians typically dont answer the question by citing Henry Vs inspiring
words on St. Crispins Eve. They point to environmental conditions (the mud),
to brilliant tactics (the placement of stakes and bowmen), and to the skill of
the English archers. Of course the etiology could be easily extended:Why was
the ground muddy? How did the stakes get formed and placed? How did the
archers come to acquire their skills? We could amplify the causal story in any
number of directions and even lead it back through previous days, months,
years, centuries, millennia, epochs, all the way to the Big Bang. (T)supposes
that any and all of these extensions deepen the original explanation. Of
course Strevens will insist that his kairetic account doesnt suppose that the
extensions he envisages involve all the gory details; he emphasizes that his
preferred explanations abstract and filter. How that is to be done in examples
of this type isnt entirely clear to me. But as I envisage tracing Agincourt
into the remote past, Ifind it very hard to understand how any appeals to
causal preconditions or to basic physical mechanisms add to the explana-
tions military historians standardly give. Even if Iimagine something more
local and more promisingan account of where the wood used for the stakes
came from or a meteorological history that explains why the ground was so
muddyno advance in understanding seems to accrue. Once we see how a
small group of skilled archers could create havoc among the French cavalry,
the question of why a hopelessly outnumbered group of weary soldiers could
prove victorious is completely resolved.
Sp e c i a l - S c i e n c e Au t o n o m y a n d t h e Di v i s i o n of L a b or [179]
and drawing one chromosome from each pair to each pole. Cytology offers
an answer that abstracts from many molecular details, and extending the
original account in terms of pairing and segregation by introducing the
cytological details of spindle formation and chromosome migration really
does deepen the answer to the original question. This specific instance of
(T)is correct. Furthermore its easy to see why. The original why question
generates a more sophisticated why question, and one that retains the
same contrastclass.
Recognizing that feature of the situation helps us see why (T)is often
violated and why Strevenss vision of a long chain of ever-deeper explana-
tions leading back to fundamental physics is misguided. You have the cyto-
logical story about spindles and migration, and you see why (in the usual
case) the segregation is random. What questions do you ask next? There
are many. You might wonder about the etiology of the spindle or about the
selective history that has given rise to this intracellular mechanism. That
direction of extension would lead to the types of causal explanations weve
seen in previous examples, for which (T)plainly fails. Strevens surely wants
to probe the molecular bases of the mechanism invoked in the cytological
storybut what exactly is the explanation-seeking question to which the
more fundamental analysis provides the answer? Im inclined to think
that it isnt a why question at all, but something more like How does it
work? or What is going on at the level of the constituents? Even if you
try to frame a why question, it seems highly dubious that it will retain the
contrast class. We already know from the cytological story what is to be known
about why the segregation is random rather than biased to some particular asso-
ciation of nonhomologous chromosomes. Recall that the kairetic approach was
supposed to abstract from the details of particular instances and bring out
the crucial common feature. Thats already been done, and its been done by
cytology. Theres nothing more to deliver along theselines.
How does this relate to the argument originally advanced in my 1984
paper? Strevens chides me (gently) for a diagnosis in terms of natural
kinds, pointing out that this was hardly a concept beloved by those who
defended the unity of science hypothesis. That is literally correct. Its
worth pointing out, however, that something in the vicinity was a stan-
dard part of logical empiricist thinking from the late 1940s on, after
Nelson Goodman forced philosophers of science to come to terms with
the distinction between predicates that are projectible and those that are
not. So I recapitulate my diagnosis here, with a different emphasis: The
reductionists problem is that higher-level explanations turn on identify-
ing kinds that cut across the kinds featuring at lower levels. When the
advocate of kairetic explanation does his abstraction and filtering, he is,
Sp e c i a l - S c i e n c e Au t o n o m y a n d t h e Di v i s i o n of L a b or [181]
CHAPTER7
T o wa r d a P ol i t i c a l P h i l o s op h y of S c i e n c e [183]
inevitably arise. One of the things that makes truth significant is that we
can do things with it:cure disease, grow more food, understand our place
in nature. How are these to be evaluated or compared? How and when are
scientific truths to be applied to individual or political decisions? Several
such questions will arise as this essay progresses.
T o wa r d a P ol i t i c a l P h i l o s op h y of S c i e n c e [185]
He discusses The Bell Curve, the notorious book by Richard Herrnstein and
Charles Murray (1994) in which they argue, first, that economic class gen-
erally reflects talent: people are poor because they are relatively dumb; and,
more notoriously, that the great overrepresentation of African Americans
among the poor reflects the fact that African Americans are on average less
intelligent than Americans of European descent. Suppose this is true. What
would be the benefits of knowing it? As Kitcher argues, there are few obvi-
ous benefits and some obvious harms. Centrally the position of a widely
disadvantaged group is likely to be substantially worsened as those who
have continued a centuries-old tradition of discrimination against African
Americans, for example in employment, feel justified in their discrimina-
tion and are (even) less motivated to end the practice. Worsening the posi-
tion of the already badly off is widely agreed to be a very bad thing.
There will no doubt be many who will respond to this argument by
insisting that we should want the truth whether or not it hurts. After all,
the discovery claimed by Herrnstein and Murray really may show that
alleged discrimination in employment isnt what it seems. Perhaps it is just
a reflection of the systematically lower qualifications of African Americans.
Just as women are (appropriately, it may well be claimed) underrepre-
sented in occupations requiring upper body strength, so African Americans
are appropriately underrepresented in jobs that require exceptional
intelligence.
At this point we need to be a little more critical of the assumption
that the research in question is likely to generate truth. In the first place,
research of this kind assumes that there is something being measured,
intelligence, that is somehow a purely biological property, independent
of upbringing and education. This is highly contentious. If, on the other
hand, we recognize that measured intelligence reflects the outcome of a
developmental process influenced at least as much by education as by any
natural endowment, then we see that the outcomes of intelligence tests
are likely to be a symptom of inequality rather than a measure of the cause
of inequality. Moreover this misinterpretation, if taken seriously, is likely
to justify the continuation of the unequal treatment that in fact causes
the perceived differences. Thus this research may not only be taken to jus-
tify unequal treatment, but it may help to perpetuate the phenomena that
form the basis of its misguided interpretation.
It is, in addition, highly debatable whether a further premise of the most
notorious aspect of the research in question is justified, namely that there
are two kinds of peopleAmericans of African descent and Americans of
European descentthat may turn out to differ systematically in their prop-
erties. As is well-known, genetic diversity is much greater within standardly
T o wa r d a P ol i t i c a l P h i l o s op h y of S c i e n c e [187]
Nazi research on issues such as hypothermia, while sound in principle, is
so morally repugnant in its methods that it has been intensely debated
whether it is even morally acceptable to make use of its outcomes (Moe
1984). Perhaps exploding atomic bombs in earthquake faults would be a
good way of learning about tectonics, but few people would advocate pur-
suing this line of inquiry.
The preceding remarks illustrate a central theme in Kitchers work:the
traditional idea of science as value-free is indefensible. Values are unavoid-
ably implicated not only in decisions about what topics we decide to inves-
tigate but in the concepts in terms of which we formulate the questions
we try to answer. These questions, in turn, cannot be answered apart from
decisions about the goals we would like our science to serve. This is not, as
is still sometimes supposed, an argument that science is subjective or that
the acceptance of its results is a matter of taste. It is an argument that we
cannot understand science properly without attending to these fundamen-
tal normative aspects.3
It is, at any rate, uncontroversial that there should be limits on what
scientific research should be undertaken at all, letalone publicly funded,
and Kitchers contribution in his discussion of the example of racial dif-
ference is valuable especially for exploring the wide range of important
respects in which this research may be highly undesirable even if from a
scientific point of view, in terms of its likelihood to discover truths, it were
perfectlysound.
The question then inevitably arises, how we should decide what research
should be undertaken. Kitchers answer, very roughly speaking, is that
such decisions should be made democratically. The democracy he has in
mind, however, is not the vulgar democracy of popular referenda but
a more Millian conception that recognizes the importance and value of
expertise. Ishall consider some aspects of this solution in the next section.
Here I note only that Kitcher does not advocate the simple and obvious
solution of banning research that is deemed undesirable, though presum-
ably the processes that he advocates for decision making would at least
make public funding of, for example, the research just discussed on racial
difference very unlikely.
I will confess, in passing, that Iam somewhat tempted to a more coercive
view. Democratic decision making, vulgar or sophisticated, is likely mainly
to affect questions of public funding. The Millian perspective that Kitcher
3. The role of values in science has been quite widely discussed in recent philosophy
of science. See, for example, Douglas 2009; Kincaid etal.2007.
T o wa r d a P ol i t i c a l P h i l o s op h y of S c i e n c e [189]
scientific information so as to take sensible precautionary health measures
without coercion. To do this they must have either the ability to assess the
value of scientific research or have a high level of trust in scientific experts.
But the first option seems unrealistic in the foreseeable future, and the
second seems both frequently absent and anyhow problematic in various
ways. This brings us to the heart of the problem Kitcher (2011b) addresses
in most detail in the second of the books under discussion here, Science in
a Democratic Society:What is the proper relation between democracy and
expertise?
Kitcher approaches the problem through what he calls the division of
cognitive labour. Various people, including Immanuel Kant, have been
described as the last person to know everything worth knowing, but there
is no doubt that that is a feat far beyond the reach of anyone currently liv-
ing. PubMed, an index of biomedical publications, contains almost 24mil-
lion citations at the time of writing, and a new one is added about every
minute. Even if 95percent of these have nothing very interesting to say,
this still leaves a million or so worth reading, and this is just one major area
of scientific knowledge. Hundreds of other databases can be found listing
tens or hundreds of thousands of resources on topics from Japanese his-
tory to gardening, from astrophysics to philosophy. (PhilPapers now lists
over one million books and articles.) Of course only a fraction of all this
should probably count as worth knowing, but then the problem is to find
out what fraction. The only solution to this problem is a division of cogni-
tive labor: for many different areas of knowledge there are some people
who know a good deal about what is known or credibly believed; these are
the experts. If we need to know something about an area on which we are
not an expert, rather than dive into this ocean of more or less reliable ver-
biage, we find an expert.
So far, so good. Now return to the central questions for political phi-
losophy of science:How do we decide what science should be done, or at
least funded, and how do we apply science to real practical problems? Two
problems arise. First, there is wide consensus that decisions on public
policy should be to some extent democratic. Placing decisions wholly in
the hands of an elite caste, whether they be politicians, philosopher kings,
priests, or scientific experts, notoriously leads to abuse and oppression.
On the other hand, democracy is likely sometimes to lead to policies that
scientific evidence suggests will be disastrous. Second, there are areas in
which, arguably, no expertise exists. Ihave already suggested that there are
no scientific experts on which science should be supported, as all scientists,
by virtue of the cognitive specialization that is their job, are bound to be
biased on this issue. Kitcher (2011a, 286)adds an additional interesting if
T o wa r d a P ol i t i c a l P h i l o s op h y of S c i e n c e [191]
and elsewhere in Europe on genetically modified (GM) foods. Emerging
from a remarkable explosion of knowledge of and technical capacity in
interaction with genomes, these were developed as products intended to
meet human needs for food.4 Within the scientific community they were
widely agreed to be an excellent pathway to produce many desirable char-
acteristics of cropsreduced competition from weeds, resistance to insect
attack, extra nutrients, and so onand hence to a more efficient and pro-
ductive agriculture. Though a fair amount of due diligence was seen as nec-
essary, as in any major modification of the human food chain, scientists
generally saw little reason to anticipate serious risks to health and claimed
possible benefits for the environment, including reduced need for herbi-
cides and pesticides.
Public reaction was another matter. Under the inspired label of
Frankenfoods, GM crops were widely portrayed as a God-like interference
with the natural order, likely to involve unacceptable threats to human
health and the environment. This perspective seemed quite unaffected
by declarations by scientific experts that such fears were groundless, and
political pressure driven by the strength of public hostility led to the large-
scale abandonment of the technology. In some sense this must presumably
be seen as a success for democracy:the public on balance rejected this tech-
nology, and the democratic government responded to their preferences.
This appears also, however, to be a departure from well-ordered science.
Assuming the experts are rightand there seems little compelling reason
to doubt it in this casean opportunity for a valuable technology, not to
mention a leading position in the development of the technology as an
internationally marketable commodity, appeared to have been passed up.
Why did the UK public, and publics in much of the rest of Europe, take
so strongly against this technology? The story is, unsurprisingly, very
complex, involving a variety of interests and arguments.5 There is a his-
tory that goes some way to explain the background of public suspicion of
scientists, for example the then recent fiasco in the United Kingdom over
the management of the bovine spongiform encephalopathy (mad cow
disease) outbreak, in which scientists had informed the government that
there was no risk to public health, something that turned out to be false.
There was widespread and understandable suspicion of the motives of the
4. They were also, of course, technologies intended for the enrichment of corpora-
tions and their shareholders, a factor that no doubt played some part in generating the
public suspicion directed towardthem.
5.Much more detailed discussion of some of these arguments can be found in
Barnes and Dupr (2008, chs. 6 and7).
T o wa r d a P ol i t i c a l P h i l o s op h y of S c i e n c e [193]
likely that it would have been found in the wild, though killing the organ-
isms with which it associatesall or most plantsprobably would not be
a selectively advantageous strategy. Moreover the extrapolation from the
ability of this modified bacterium to kill a plant in a confined environment
in a laboratory to its likely spread across the planet is, to put it politely, bio-
logically unlikely. But, and this is my main point, to the reader with limited
scientific knowledge and limited access to scientific resources (should he or
she even think of looking for them) the conclusion that GM technology had
almost wiped out life on Earth might seem perfectly reasonable.
Parallels with the case of the MMR vaccine are obvious. The notorious
paper by Andrew Wakefield et al. (1998) has been fully discredited in the
scientific community, disowned by Wakefields collaborators, and with-
drawn by the Lancet. Dr. Wakefield has been struck off the medical reg-
ister by the British Medical Council. Yet again, consulting Google on this
topic it is clear that in the wider public a debate still rages. Many suggest
that Wakefield was smeared by the medical establishment, certainly one
possible interpretation of what has happened to him. A particularly telling
comment in the Internet discussion is the following: Why dont you let
parents just make their own decisions? Do your research and make what-
ever decisions you think are best for your OWN children. Isnt that the very
essence of parenting? This seems to reflect the ideological reality in which
many parents in fact decide not to vaccinate their children and the conse-
quent rising incidence of disease.
KITCHERS SOLUTIONS
T o wa r d a P ol i t i c a l P h i l o s op h y of S c i e n c e [195]
concerned to resolve social discord. It is not easy to see how this alignment
is to be achieved.
A somewhat cynical view is that the notion of ideal deliberators often
seems like little more than a philosophical delivery van for Kitchers policy
proposals (Brown 2013, 395). Slightly less cynically, it is difficult to see
what would make a deliberator ideal if, as Kitcher insists, there are no ethi-
cal experts.6 Not cynically at all, while Ihave a lot of sympathy for the model
of an ideal conversation as an intellectual tool, it seems to me somewhat
tangential to what is arguably the most interesting feature of Kitchers
work in this area, the aim of understanding and addressing the discordance
between two social technologies that Kitcher, and Iexpect most of his read-
ers, admire:science and democracy. For this is a technological problem, a
problem, that is to say, in social technology. Here it seems we should be
concerned with a social process, perhaps Kitchers actual proposed conver-
sations, rather than a theoretical, meta-ethical account of what would be
the normatively desirable outcome of a well-ordered science.
This may seem wrong for the following reason. Surely we need some
explanation of why we thought science was disordered in the first place,
and doesnt this require an account of value against which the present situ-
ation can be judged wanting? Ithink this worry does point to a real tension
in Kitchers writings on this topic. His concerns do certainly begin with
the firm conviction that things are amiss, for example the unwillingness
of democratic states to adopt serious measures to combat climate change.
And I dont doubt that there are good reasons for this. Indeed it might
well be sufficient to formulate these reasons in terms of evidence and
self-evidently appalling consequences regardless of the views that would
be reached by ideal deliberators. But of course both evidence and conse-
quences do involve normative assumptions.
At any rate, if one is serious about democracy, even a sophisticated
democracy free of all the familiar forms of vulgarity, one cannot prejudge
the question of what any suitable democratic process will decide about any
actual policy issue. Surely it cannot be a condition on an adequate account
of democracy that it always reach the correct decision as judged by some
external normative standard. Kitcher is an optimist, and he clearly believes
that a proper, nonvulgar democratic system will produce decisions more
or less of the sort that sensible people like he and Iwould prefer. But it
is conceivable, for example, that a democratic society might reach a fully
6. This is because Kitcher (2011b, 4950) sees the ethical project as fully egalitarian,
in which democratic processes are designed to produce consensus. It is the process, not
some privileged access to the proper result, that is important.
T o wa r d a P ol i t i c a l P h i l o s op h y of S c i e n c e [197]
All this is not to deny that groups of this kind can be a good idea. Indeed
they are demonstrably so. One highly effective quango, the UK Human
Fertilisation and Embryology Authority (HFEA), which regulates research
involving human gametes and embryos, is a very widely respected group, in
many respects very much the kind of entity Kitcher advocates.8 This body
includes scientists, doctors, women who have experienced fertility treat-
ments, a bishop, a lawyer, and even a philosopher. Unfortunately its deci-
sions do not appear to change the minds of those who oppose them. For
example, when the HFEA first licensed therapeutic human cloning in 2004,
there was outrage from a range of pro-life groups. Perhaps most telling
was the comment attributed to Josephine Quintavalle, of the pro-life
group Comment on Reproductive Ethics:It is very worrying indeed. We
have decisions of this magnitude being taken by an unelected government
quango.9 The absence of vulgar democracy may thus be used as a weapon.
I have argued that there may be a problem in harmonizing the ideal
and practical conceptions of conversation. This is most obvious in rela-
tion to the question of the scope of the problems that are to be addressed.
Recall that the kind of social technology that is the model for addressing
ill-ordered science, the ethical project, is taken to have originated in discus-
sions within small proto-human tribes. The problem of climate change, in
contrast, is global. The possibility of small groups sitting around a table to
hammer out a relatively local problem is appealing; perhaps even the adult
male citizenry of a Greek polis might gather in the agora to similar effect.
Representation of all the peoples of Earth, and all the different perspec-
tives within each nation, is a different matter and perhaps accessible only
to the ideal conversation. One problem with the ideal conversation when
it is more than some kind of rationalization of a practical process is that it
seems suspiciously like a job for an expert ethicist, of whom, Kitcher has
told us, there arenone.
The difficulties with a conversation over climate change do not end
here. Those most affected, in all but the most pessimistic scenarios, have
yet to be born. Perhaps there are many possible people who will not be
born unless we do something serious to address this problem. Who will
speak for the unborn or the possibly never to be born? Kitcher thinks that
among peoples central life goals are the well-being of their children and
CONCLUSION
I have been somewhat skeptical about the proposals Kitcher sketches for
reconciling science and democracy. He is, Ihave said, an optimist, and in
this domain Iam more pessimistic. Kitcher thinks that scientists discredit
themselves by making excessive claims for their expertise and by acting on
values other than those that should legitimately underlie their professional
work. Iagree. If institutions could be constructed that would expose these
flaws and open their work to rational and civilized discussion, these flaws
could be removed and well-ordered science would regain the deserved trust
of the citizenry.
I fear that the sources of these defects go deeper and that they will be
even harder to remove than Kitcher supposes. Ihave described Kitchers
project as a political philosophy of science, and I wholeheartedly agree
that that is something we need. The politics, however, is largely limited
to the commitment to an admittedly sophisticated conception of democ-
racy. This is, no doubt, a good thing to be committed to. But many of the
problems that Kitcher is concerned with arise not merely from failures of
democracy but also from the intrinsic problems with the liberal, or increas-
ingly neoliberal, framework within which most current democracies exist.
In a social system that foregrounds competition between individuals it is
hard to imagine scientists who dont have their own agendas, even if these
are often no worse than the quest for personal success by doing good sci-
ence. More problematically, a system that encourages the accumulation of
wealth in large competitive corporations, and encourages these corpora-
tions to fund scientific research, will inevitably produce research infected
with the values of the corporate funders. These are issues, I fear, that
can be addressed only at a more systemic level than even the most well-
constructed institutional add-ons for enlightened public debate. They are,
that is to say, political rather than ethical problems.
Having expressed these doubts, Imust nonetheless reiterate my endorse-
ment of the importance of Kitchers project and the gratitude we should
feel for the work he has done to open up the philosophy of science to these
absolutely fundamental questions. The two books Ihave been considering
T o wa r d a P ol i t i c a l P h i l o s op h y of S c i e n c e [199]
on well-ordered science, together with his account of the ethical project,
constitute a systematic attempt to address the political question of the role
of science in society, and his account is full of valuable insights that should
remain part of this debate.
My suspicion, however, is that the attempt to construct a democratic
science may ultimately be impossible without more integration into the
problem of constructing a democratic society. In 2012 the worlds one hun-
dred richest people became $241 billion richer. They are now worth $1.9
trillion:just a little less than the entire output of the United Kingdom.10
The problem that Kitcher is discussing is democracy of voice, not equality
of resources. But as resources become ever more unequal, democracy of
voice becomes ever more unrealistic. It would be nice to see a democratic
socialist account of well-ordered science, though perhaps, given the dis-
tance we are from democratic socialism, Kitchers account, despite inevi-
table weaknesses, will be more useful.
ACKNOWLEDGMENTS
Reply toDupr
PHILIP KI TCHER
For the past thirty-five years John Dupr has been one of my closest philo-
sophical interlocutors, someone from whom Ihave learned more than from
almost anyone else. His lucid intelligence, his wide-ranging knowledge, and
his deep humanity are apparent in the essay he has written herealong
with the wit so familiar to his friends. As has so often happened in recent
decades, his reservations about some of my ideas crystallize concerns Iam
already beginning to formulate, helping me to become more explicit about
10. See Bloomberg News, Worlds 100 Richest People Got $241 Billion Richer in
2012, Los Angeles Times, January 3, 2013, http://articles.latimes.com/2013/jan/
03/business/la-fi-billionares-gain-20130103, accessed May 20, 2014. The figures are
derived from the Bloomberg BillionairesIndex.
T o wa r d a P ol i t i c a l P h i l o s op h y of S c i e n c e [201]
it before, and havent even made the distinction about underlying atti-
tudes, Iam not a utopian but a pragmatist. Pragmatists, Dewey prominent
among them, dont have to abandon ideals. They can see those ideals not
as characterizing a goal-state to be achieved but as marking a direction in
which to move. Ideals are to point us to ends-in-view (Dewey [1925] 1981,
ch. 3). As we move toward the current end-in-view we take stock of our
situation and continue to modify our aims. The principal contribution of
ideals is thus to help us in diagnosing the deficiencies of our current state,
alerting us to what is most problematic about it, so that we can start an
attempt to address the difficulties we face. Well-ordered science should be
viewed as just this type of diagnostictool.
My claims about ethical expertise need a similar sort of clarification. The
ideal of a conversation that involves representatives of all perspectives, fully
informed by the best available knowledge and mutually engaged (deter-
mined to take other perspectives seriously and to work for a solution all can
accept), is not something we can ever expect to realize. Nevertheless it draws
attention to my thesis that ethical authority is collective, and to the features
we should try to approximate in our collective decision making. None of us
has ultimate authority in the ethical conversation, but that doesnt mean
we are all on a par as potential deliberators. Some people are more igno-
rant than others, or more dogmatic than others, or uninterested in trying
to accommodate rival points of view. When such people participate in joint
discussions about what to do, things are not likely to go well. On the other
hand, there arefortunatelysome who have broad and deep knowledge,
who are genuinely concerned to listen to alternative perspectives, who try
to find ways of addressing the concerns of people who occupy those perspec-
tives, and who also have a talent for facilitating discussion. The late John
Rawls was an obvious exemplar. Our deliberations about vexed questions
would surely go better if he, and others like him, were in the room. Yet even
Rawls, wise as he was, should not have the last word on ethical matters, for,
as Ihave emphasized, ethical judgment is collective.
Pragmatism favors an approach to progress that abandons teleol-
ogy. From Dewey ([1909] 2007) on, pragmatists have found inspiration
in Darwin, thinking of progress from rather than progress to. (This theme
is also sounded in the last chapter of Kuhn 1962 and elaborated in Kuhn
2000.) Of course there are contexts in which a teleological notion of prog-
ress is apt. When we are traveling we typically measure our progress by
the decreasing distance to our destination. In many areas, however, a very
different conception is in order. Medicine does not progress by diminish-
ing the gap between human lives and some ideal state of health. What
would that ideal be? Rather it progresses by solving particular problems,
T o wa r d a P ol i t i c a l P h i l o s op h y of S c i e n c e [203]
their choices will advance (or retard) their central ends. When elections
are dominated by monied interests, when time at the public microphone
can be purchased, the voices of experts are drowned out by those of slick
spokesmen hired to trumpet the corporate view, and the Millian ideal of
free speech disintegrates. Those people who are centrally interested in
leaving a habitable planet for their children and grandchildren, and more
generally for those who will come after them, are confused or conned into
voting against that central interest. Democracy degenerates into statistical
plutocracy, a form of government in which the wealthy few invest so as
to increase the chances that voters will elect representatives who support
the well-heeled minority. In effect the electorate becomes a set of cogs in
a probabilistic machine, designed so the right results will flow from elec-
toral choices. Unfortunately the plutocracy is still statistical. Sometimes
the machine is imperfect. (A black man, concerned at least with climate
change, is elected.) But no doubt the plutocrats look forward to the day
when the bugs are all fixed and voters obediently traipse to the polls to
endorse the policies on which the wealthy rulers have agreed.
This, however, cannot be the whole story. Dupr is skeptical of my
thought that citizens are centrally concerned with future generations. Ive
come to see his skepticism as probably warranted. Yet again its worth ask-
ing why peoplequite likely a significant fraction of themwould violate
the conditions of ideal deliberation, not attending to the perspectives of
those who are distant in time and space. The closing pages of Duprs essay
contain the materials for an answer. Contemporary capitalism, with its
relentless emphasis on unbridled competition, generates another patholog-
ical condition. Many people are in no position to fashion their life plans
or central desires in the ways philosophers view as ideal. They lack educa-
tional opportunities that would open up a wide array of careers for them,
their chances of achieving genuine community with others are severely
limited, and, most obviously, they are faced with constant challenges to the
most basic tasks of securing their lives and those of their immediate fam-
ily. The pronounced inequalities, across the globe and within even the most
affluent democracies, constitute a condition under which its eminently
likely that the subtraction of unnecessary material goods will be viewed as
an unacceptable sacrifice:once a gas-guzzler has been acquired, its steer-
ing wheel will have to be pried from the owners cold, dead fingers. Myopia
prevails. There is no serious sense that life becomes valuable through con-
tributions to a larger human project.
Duprs conclusion expresses his own conviction that this social
pathology must be remedied. Well-ordered science, with its attendant
views about ethical authority, generates the same diagnosis. It marks out
T o wa r d a P ol i t i c a l P h i l o s op h y of S c i e n c e [205]
CHAPTER8
K i t c h e r o n S c i e n c e , De m o c r a c y, a n d H u m a n F l o u r i s h i n g [207]
of cultural and political contexts. Since many debates about the advisability
of bringing science and democracy into a different alignment hinge on the
risks of thereby corrupting both science and democracy, it is useful to have
a trove of past examples and counterexamples against which to test claims
made by all sides. (The same is of course true, mutatis mutandis, for the
history of the polities and values with which science is to be brought into
alignment, but other, better qualified scholars will have to pronounce upon
these.) Third, Kitchers own work in the philosophy of science, beginning
with his earliest publications on mathematics, has been deeply informed
by historynot just the usual toy examples retailed at secondhand but
detailed, thorough analyses based on immersion in the primary sources.
This is a philosopher who thinks with history. In the case of his recent work
on science and the public good, this historicism has seeped into the concep-
tualization of the problem:he conceives of both science and democracy as
under construction, not as finished achievements but as projects in the
making. On Kitchers account these institutions are not so much like bro-
ken artifacts in need of repair as evolving organisms with contingent pasts
and open-ended futures.
In what follows I bring all three historical perspectives to bear on what
I take to be the most fully developed version of Kitchers (2011b) views
on science and human flourishing to date. I begin with a brief summary
of what seem to me the principal points in Kitchers argument and then
examine how his analysis might promote human flourishing, mostly from
the standpoint of a historian. My conclusion returns to Kitchers vision of
the project.
K i t c h e r o n S c i e n c e , De m o c r a c y, a n d H u m a n F l o u r i s h i n g [209]
Note only how radically Kitchers vision of democracy diverges from cur-
rent reality and also from more familiar liberal ideals:sympathetic mirror-
ing is worlds away from John Stuart Mills ([1859] 1975)jostling, agonistic
free marketplace of ideas, in which rival positions clash in gladiatorial
combat, each armed to the teeth with its strongest arguments. It is equally
at odds with Isaiah Berlins ([1950] 1979, 14760) conviction that there
are irreconcilable differences among communal visions of the good, each
the product of a distinct history. Kitcher shares Mills commitment to fair-
ness (everyone has a voice on matters upon which no one is an expert)
and Berlins respect for pluralism (there are multiple visions of the good
life). But Kitchers subtle but strenuous requirements of mirroring and
mutual engagement demand that the conversationalists modify their
desires in light of those of all the others, striving toward a kind of prear-
ranged harmony through sympathy.
If the ideal conversation expands participation in some discussions, the
division of epistemic labor narrows it in others:
Consider the entire range of questions pertinent to public life, all matters about
what the society should aspire to and how it might realize whatever aims are
set. These topics are partitioned, divided into non-overlapping sets, and for each
set in the partition except one [ethical matters], a particular group of people is
designated as authoritative with respect to that set. For the remaining set, epis-
temic equality holds. (Kitcher 2011b,21)
K i t c h e r o n S c i e n c e , De m o c r a c y, a n d H u m a n F l o u r i s h i n g [211]
less fruitful of results and advances in understanding than blue sky or
curiosity-driven research.
However utopian the ideal of well-ordered science may seem, Kitcher
contends that it might still serve as a pole star by which to guide future
attempts to improve the forms of democracy and science we have inherited
from the vicissitudes of histories driven more by accident than by design
(just as, by analogy, an ideal of true equality might guide political and social
reform, even if it is an asymptote never reached). He asserts that even
partial attempts to realize the ideal would be superior to current arrange-
ments (e.g., majority rule in politics, increasing privatization in science)
and offers suggestions concerning specific dilemmas likely to arise (e.g.,
what principles might adjudicate competing claims as to which diseases
most urgently deserve research efforts).
Although a strict observance of the division of epistemic labor may seem
to restrict citizen participation in decisions about science-informed public
policy, Kitcher is particularly concerned with how the circle of those able at
least to appreciate the factual constraints on and probabilities of success of
alternative lines of scientific investigation might be widened to include more
than just specialists. Some citizens might be taken behind the scenes to
observe relevant research firsthand or tutored in the details of a burning
question; more scientists might dedicate more of their time to explaining sci-
ence clearly and engagingly to the general public; scientific education might
be improved for all. Similarly (but more vaguely) citizens might be educated
to be both aware of and sympathetic to the plight of others:One part of
the remedy would take seriously the idea that part of education consists in
the encouragement and expansion of altruistic tendencies (Kitcher 2011b,
130). Even achieving a first approximation of Kitchers ideal of a new social
contract between democracy and science would seem to require substantial
modifications of what citizens can be expected to know and to feel. It is, to
use Kitchers own word, a project, and a highly ambitious one atthat.
What reason do we have to think that such a project could succeed? And
even if we were persuaded of its bare possibility, at least in some approxima-
tion, what reason do we have to think that Kitchers ideals for science and
democracy would be enough of an improvement to warrant the Herculean
effort of striving to realize them? Kitchers answers to both of these chal-
lenges appeal to history, but history in different registers and enlisted to
serve differentends.
K i t c h e r o n S c i e n c e , De m o c r a c y, a n d H u m a n F l o u r i s h i n g [213]
science-based technology have repeatedly transformed society, for good or
ill, these commitments are at best antiquated and at worst dangerous.
For the historian Kitchers account raises two questions:Is the account
accurate where it matters to Kitchers theses? If it is in the main accurate, is
it a legitimate use of history? It would be pedantic to criticize what is meant
to be a sketch as if it were a treatise, but there is one consequential aspect
of Kitchers depiction that most historians of science would query:the ori-
gins and pervasiveness of scientific autonomy. In cultures that have insti-
tutionally supported systematic inquiry since the Renaissance, examples
of autonomy as an ideal are rare, and actual examples are rarer still. All
the luminaries of early modern science Copernicus, Kepler, Galileo,
Descartes, Boyle, Hooke, Newton, Leibniznot only vaunted the utility of
their scientific pursuits; they also practiced what they preached. Whether it
was Kepler (1615) estimating the volume of wine barrels for the city of Ulm,
Galileo marketing a proportional compass for military engineers (Valleriani
2010, 2740), or Leibniz corresponding with alchemists about artificial
phosphors (P. H. Smith 1994, 24855), these natural philosophers were
profoundly committed to science in the service of worldly goals. Thomas
Shadwell ([1676] 1966)and other Restoration wits might have poked fun
at the Royal Societys investigations of the blue of plums and luminescent
lamb shanks, but the Fellows pursued such topics with practical applica-
tions in mind. The Royal Societys French counterpart, the Acadmie royale
des sciences in Paris, was from its inception charged with the responsibility
of offering technical advice to the state on everything from large engineer-
ing projects to smallpox inoculation (Hahn 1971). If anything, the cult of
public utility only intensified during the eighteenth century. Even with the
rise of the prestige of pure science (reine Wissenschaft) in the latter half
of the nineteenth century in the context of the German research univer-
sity (Daniels 1967), German scientists of the first rank, such as Justus von
Liebig (Rossiter 1975)and Hermann von Helmholtz (Cahan 1989), prided
themselves on their contributions to the advancement of agriculture and
electromagnetic technology.
Far from being a holdover from an earlier era, the insistence on the
autonomy of science seems to have first emerged in twentieth-century
democracies, usually as part of a defense of public funding for pure or
basic or fundamental science without immediate prospect of applica-
tion (but usually with the promissory note of future practical benefits).
Although such pronouncements date at least as far back as the 1920s
(Pielke 2012), Kitcher is surely correct to identify Vannevar Bushs ([1945]
1980)ScienceThe Endless Frontier as the locus classicus of the view that the
autonomy of basic science guarantees the fruitfulness of applied science. As
K i t c h e r o n S c i e n c e , De m o c r a c y, a n d H u m a n F l o u r i s h i n g [215]
at least approximately realized and that both science and democracy would
be better off if they were? In the case of mutual engagement, Kitchers use
of history is both variegated and subtle. He musters examples from both
actual and speculative history to make his case that over the course of cen-
turies and even millennia, human beings have shown themselves capable of
widening their circle of sympathies to include ever more people entitled to
the full dignity of personhood. The gradual de jure abolition of slavery all
over the world (de facto abolition, alas, has yet to be achieved) is an impres-
sive example for Kitchers case:most citizens of modern polities recoil at
the intuitions that Aristotle and his ancient Athenian contemporaries con-
sidered self-evident on this subject, however much his writings on other
ethical and political topics may still resonate. With such examples in mind,
Kitcher (2011b, 48) formulates a different version of moral progress: not
progress toward any ethical truth but rather progress from, measured in
terms of our ability to solve problems.3 If humans are capable of so remark-
able an expansion of their sympathies in the teeth of their economic and
other egoistic interests, why is it not realistic to hope and strive, Kitcher
asks, for a still broader embrace of the predicaments, perspectives, and aspi-
rations of others, embracing even unborn generations and other species?
So far, so good: the cautious historian may point out the glacial pace
of such conceptual, affective, and institutional transformations and their
costs in blood, wealth, and stability; the cautious anthropologist may won-
der how far the radius of human sympathy can stretch without snapping
but neither would challenge the relevance and force of Kitchers appeal to
actual historical analogy. However, both may falter over his use of spec-
ulative history. A keystone of Kitchers notion of mutual engagement is
the ethical equality of the participants in the ideal conversation. You and
Imay differ concerning how well we understand the facts and probabili-
ties of the matter at hand, and the epistemic division of labor dictates that
Ishould defer to you if you are better informed. But there is no deference
concerning whose interests in the possible outcomes should prevail. Why
should this be so? There are many historical precedents of societies that
have institutionalized deference to various sources of authority:religious
saints, wise elders, powerful chieftains, rich bankers, hereditary princes,
charismatic politicians, violent warlords. Hierarchies are the rule, not the
exception in the history of human societies.
Of course Kitcher knows all this. However, he interprets historyor
rather prehistory, and therefore speculative historyas offering an earlier
our Paleolithic predecessors sat down to decide on the precepts for governing
their group life.... All adult members of the band are to be heard, and the wishes
of each must be considered. To diverge from egalitarianism of this sort would
risk the survival of the group, for all had to pull together on occasion, to meet
the challenges of the environment.
K i t c h e r o n S c i e n c e , De m o c r a c y, a n d H u m a n F l o u r i s h i n g [217]
nonegalitarian regimes. Indeed the weight of the statistical evidence would
seem to buttress the conclusion that hierarchy is more natural than equal-
ity to Homo sapiens. Kitcher (1985) has elsewhere written brilliantly of the
snares and pitfalls of such partial arguments concerning human nature and
does not renege on those principles hereagain, in my opinion, rightlyso.
To conclude: Kitcher invokes two kinds of argument from history in
order to fortify his case that present forms of both democracy and science
are neither inevitable nor incorrigible. One kind of argument uses history
to broaden the horizon of the possible:no necessary trajectory of events
has led us to our current arrangements; alternatives are both conceivable
and instantiated by history; history provides at least analogies to the sorts
of cognitive and affective transformations Kitcher holds to be preconditions
for realizing improved versions of democracy and science. This is history
in the service of the imagination, and its message is that there have been
(and could be) more things in heaven and earth than have been heard of in
our current philosophy. As Ihave argued, the details of this history matter
in ways consequential for Kitchers larger aims, but his general strategy of
using history to enlarge the realm of the thinkable and to fortify hope that
reform is possible seems to me to succeed admirably. Iam more skeptical
about the second kind of argument, which uses history to ground norma-
tive claimsespecially because it is deployed asymmetrically in the case of
science and politics. History can persuade us only that certain forms of sci-
ence and politics are possible (and this is well worth knowing, given the lazy
tendency of our presentist culture to naturalize its own arrangements). But
history alone cannot suffice to show that one alternative that has existed in
the past is better than another, nor to make us desire still other alternatives
enough to make significant sacrifices in order to realize them. Only a com-
pelling vision of a better way of life can mobilize those energies, and that is
why Kitchers notion of human flourishing lies at the heart of his proposals.
HUMAN FLOURISHING
What would it mean for both democracy and science to promote human
flourishing? Kitcher (2011b, 55) does not pretend to have an answer to
the age-old philosophical question about what constitutes the good life
(though he offers some observations, e.g., making a positive difference in
the lives of others), and it would be contrary to the spirit of his refusal to
admit ethical expertise to privilege the answers of philosophers over those
of other people. However, he does have views about the conditions most
likely to further the goal of providing, for the entire population, equal and
K i t c h e r o n S c i e n c e , De m o c r a c y, a n d H u m a n F l o u r i s h i n g [219]
Would such a process of mutual adjustment of desire promote human
flourishing in se? The nobility of the end ennobles the means, but is that
enough? The question is made more pressing because of the uncertainty
of the outcome at every stage:participants in the ideal conversation can
only try to find the best balance among the varying assessments (indefi-
nitely iterated) made by fellow participants (Kitcher 2011b, 52); even if
they succeed, they can equalize only chances (not certainties) of leading a
worthwhile life; the well-ordered science guided by the ideal conversation
can never promise to deliver the ideal results. It is a grand game, played in
earnest for the highest stakes, but a game all the same. Is the game worth
the candle?
I think that there are at least two reasons for doubt in the case of the
ideal conversation. The first is intrinsic to the process itself: even if the
extraordinary levels of cognitive and affective perspectival suppleness
Kitcher calls for are humanly possible, cultivating them may weaken or
even extinguish traits rightly deemed essential to human flourishing. How
would spontaneity fare in the deliberative vertigo of the hall of extended
mirroring? Would we not be permanently sicklied over with the pale cast
of thought?4 Deliberation famously cools the passions, but might not the
indefinitely iterated deliberation of extended mirroring freeze all desire?
The particular kind of deliberation required, infused by sympathy, poses
still more troubling threats to values of character and integrity. Firmness
of purpose and resolve, constancy of traits, peculiarities of personality and
taste might withstand deference to duty, however arduous the task of self-
mastery required. But could they withstand the modification of the very
structure of desire through sympathy with the desires of others, all others?
The hypersympathetic psyche might resemble the chameleon of Woody
Allens film Zelig (1983), who involuntarily took on the tincture of what-
ever personality he encountered. Such a pliant self might raise worries as
to whether it was a self atall.
The second reason for doubt concerns the objects rather than the act of
sympathy:the desires of others. In primitive mirroring these are egois-
tic desires; in extended mirroring, desires filtered to eliminate impos-
sible and incompatible options. But they remain individuals desiresnot
shared ideals of the collective good, not common visions of a better way
of life. Of course nothing in Kitchers scheme prevents individuals from
4.Elster (1984, 40) makes a similar point in the context of cost-benefit analy-
sis:Even if the deliberations do succeed in modifying the behavior of the character in
the desired way, the very act of deliberating can modify the character for the worse, and
in ways judged even more important, through the stultifying effects on spontaneity.
K i t c h e r o n S c i e n c e , De m o c r a c y, a n d H u m a n F l o u r i s h i n g [221]
is progressing toward some ultimate goal, be it perfect justice or absolute
truth; both are instead progressing from some anterior, unsatisfactory
state of affairs that must be remedied. How the remedy is to be sought
is as important as what the remedy is; the process of achieving the ideal
conversation and well-ordered science teaches us more about both than
any solution handed down ever could. Dissatisfaction is never-ending, and
therefore so is the search for remedies. This is a vision suffused with the
lessons of history, both the history of science and the history of democracy.
Both are works in progress and destined to remainso.
This is a vision more in tune with the self-image of science than with that
of democracy (at least American democracy). Because the latter is bound to
interpretations of its Constitution, consequential decisions about the let-
ter and spirit of democracy look to the past rather than the future or even
the present. Not all interpretations of the Constitution have been funda-
mentalist (e.g., attempts to divine and abide by the original intentions of
the eighteenth-century framers), but all necessarily must build analogi-
cal bridges between past principles and present case (a structure of argu-
ment reinforced by common law appeals to precedent). Even jurists who
disavow literalist interpretations of the Constitution bear the burden of
demonstrating the continuity of the present decision with past precepts
although the Supreme Court has proved itself capable of rare moments of
remarkable and consequential creativity.6
In contrast, modern science is largely oriented toward the future.7 This
future fixation has sometimes taken on a melancholy cast, when scien-
tists confront the likelihood that they and their work will be forgotten in
a generation and that much of what they believed to be true will inevi-
tably be revised by their successors (Daston 2001). But the open-ended
dynamism of science has also exerted a powerful magnetism, and not just
upon scientists. Since the early nineteenth century science has been held
up as a model of knowledge that liberates (Secord 2014) and of a polity
that reconciles progress and stability (Jewett 2012). Kitchers account of
well-ordered sciences contribution to human flourishing rightly concen-
trates on products: therapies for cruel diseases, hardy crops for hungry
people, accurate accounts of everything from the causes of species deple-
tion to the evolution of pathogens. However, there is some evidence that
6.For example, in the 2013 Supreme Court decision United States v. Windsor
(Becker2014).
7. However, some sciences, especially those that study phenomena that unfold on
a superhuman time scale, also look to the past to supply archives of essential data
(Daston2012).
Reply toDaston
PHILIP KI TCHER
K i t c h e r o n S c i e n c e , De m o c r a c y, a n d H u m a n F l o u r i s h i n g [223]
work is motivated by viewing history as providing clues to the character
of areas of human practice. Our understanding of mathematics, science,
ethics, and democracy is limited if we fail to probe the complex histories
out of which our current commitments in these domains have emerged.
For my philosophical purposes details of history sometimes matterand
sometimes they do not. Dastons thoroughly sympathetic exploration of
my views leads her to see where overhasty historical reading has interfered
with my philosophical purposes. She doesntcarp.
As she recognizes, my account of the autonomy of early modern sci-
ence (drawn, Ifear, from attending too much to the Thomases, Sprat and
Shadwell) is defective, and the deficiency affects my argument. The lumi-
naries she cites (Galileo, Boyle, and company) certainly engaged in proj-
ects they took to be socially useful. Isuspect that at least some of them
were not so committed to being free and unconfind that they would have
resisted directives from the broader society to engage with particular areas
of inquiry. So my history of a long-held dedication to autonomy needs revi-
sion. One consequence, drawn by Daston herself, is an added emphasis
on justifying claims to scientific autonomy in an age of radically increased
public funding. But, as she sees, the argument I offer against those who
defend autonomy by gesturing at famous instances of government inter-
ference with sciencethe Lysenko affair being the favorite exampleis
undermined once the history is corrected. My proposal to replace the
appeal to anecdote with a statistical survey of the effects of outside direc-
tion of research topics will not do if the current level of interference is his-
torically unprecedented. Fortunately the negative point about appeal to
anecdotes survives. Moreover even if a more accurate and richly variegated
history were to raise questions about the potentially harmful effects of
diminishing scientific autonomy, the appropriate stance (envisaged in my
reply to Nancy Cartwright and Alexandre Marcellesi) would be to engage
in cautious experimentation rather than insisting that a handful of loosely
analyzed examples demonstrate that scientists knowbest.
A deeper challenge focuses on the history Ioffer in an attempt to liber-
ate us from the constraints usually taken to limit the options in thinking
about ethics. As Daston rightly notes, my history involves reconstructing
events and processes that occurred long before the invention of writing
and thus is inevitably speculative. But claims about prehistoric human
life, even prehistoric human social life, dont all involve the same degree of
conjecture. The bones and artifacts left by our precursors enable archaeolo-
gists to defend relatively well-grounded hypotheses about the size of ances-
tral bands at various moments in the pastand to infer the less firmly
established but still not completely speculative conclusion that, until about
K i t c h e r o n S c i e n c e , De m o c r a c y, a n d H u m a n F l o u r i s h i n g [225]
coming to feel or perceive or judge (its not obvious what the right verb is)
things we have previously taken for granted to be no longer acceptable. The
historical narrative has exposed what we were doing, and our new aware-
ness is embodied in a feeling of profound discontent (shame or guilt) and
perhaps that a particular form of change is the way to set thingsright.
How does this apply to the narrative offered in The Ethical Project? The
mix of how possibly and how actually explanations generatesat
least in mea sense of the norms and structures of shared human lives
as emerging from a long series of attempts to wrestle with a deep prob-
lem in the human predicament. Isee members of our species as drawn to
live together with one another, without biologically based psychological
adaptations for making that shared life go smoothly. Our responsiveness
to others is limited, and ethical life grows from a technology for amplifying
that responsiveness. The pioneers who began the ethical project found the
obvious solution to that deep problem. They deliberated on terms of rough
equality, listening in the cool hour to the claims issued by all members of
the band. Out of their efforts has emerged a far more complex way of life,
one in which human beings experience thoughts and emotions that were
inaccessible to their ancestors. Along the way the old egalitarianism has
been severely compromised. Many voices, many claims are no longer heard.
Yet the old problem of limited responsiveness endures, manifesting itself
in new forms and causing continued suffering.
Perhaps Ishould have written differently, posing questions rather than
trying to defend theses:Here is a narrative of how we arrived at our cur-
rent ethical practices. Do you see how the same old problem reemerges in
new guises? Doesnt that problem deserve to be seen as a deep feature of
our predicament, because it endures and because its effects are so large and
damaging? Isnt the ancestral solution, of mutually engaged conversation,
the obvious way to address that problem? What else could we do today
besides trying to scale up the old solution, difficult though that will surely
be? When you reflect on the history, what possible ways of going on do
yousee?
The questions Ive posed could be viewed as opening a conversation
with Daston, Smith, Srinivasan, and others. Out of the conversation might
come something entirely new, a consensus on a different set of emotional
reactions and revisions of commitment. We would have learned from his-
tory, but not quite in the way I originally thought. History would have
sparked some individual reactions, and interactions with others would
have refinedthem.
But what are the principles that guide inferences from premises about
the history of ethical practice to conclusions about how we should modify
K i t c h e r o n S c i e n c e , De m o c r a c y, a n d H u m a n F l o u r i s h i n g [227]
out whether our maxim could be universalized or the utilitarian doomed
to impossibly complex calculations. We often proceed, justifiably, by taking
our habits (Deweys good word) for granted, interrogating them according
to our ethical standard only under conditions of live doubt (Peirces impor-
tant counter to Cartesianism). Neither Hamlet nor Zelig is our destiny.
I entirely agree with Daston on the importance of individual projects to
human flourishing. Following Kant, Humboldt, and Mill, Itake it to be cru-
cial that our life themes should be freely chosen, and an important mode of
ethical progress would consist in spreading more broadly and more securely
the conditions under which people can freely choose the shape they intend
their lives to have. Yet I add another condition, one that may renew an
Aristotelian theme. Valuable life themes must involve connections among
lives. Thus the individual must enjoy the possibility of living in commu-
nity with others, of participating in ventures that affect a larger group than
the single self. So Ishare Dastons sense that flourishing lives contribute
to something bigger than themselves (Kitcher 2013, 2014). Whether my
emphasis on this form of community extricates me from a narrow method-
ological individualism is a matter Ileave to my critics.
Among these larger ventures might be an attempt to achieve pure under-
standing of nature, without thought of solving practical human problems.
Moreover parts of that attempt might be endorsed by the participants in
an ideal conversation. Suppose, however, that were not so. At a particular
historical moment the scientific community forges ahead with its search
for deeper understanding of the universes first few microseconds, at a cost
to the health and the lives of a large group of people. Imagine that a com-
prehensive, informed, engaged conversation would resolve that the neglect
of those peoples welfare was an ethical mistake. In that context the scien-
tific communitys declaration that they were pursuing an important mode
of human flourishing would ring hollow. They would be like imaginary
British mathematicians who rejected the call to Bletchley in favor of con-
tinuing their work on Fermats Last Theorem. Flourishing is important, but
there are occasions on which your preferred mode of flourishing must be
given up for the larger good. With this judgment Ihope Raine wouldagree.
Deliberating Policy
Where Morals and MethodsMix
NANC Y C ART WRIGHT AND ALEXANDRE MARCELLESI
P hilip Kitcher has long been a model and an inspiration. He has not
taken the easy way:philosophy for philosophys sake, speaking to our
own history, pursuing topics and ideas that wear their philosophical worth
on their sleeve, ideas that advance the problems they are grappling with
in an easily discernible fashion. Instead Kitcher has pioneered what Otto
Neurath called philosophy for life. His philosophy serves the problems we
all, as members of society, grapple with, problems that a facile application
of what our philosophical history provides will not help, where new phi-
losophy must be built, molded to the problems of life it aims to serve, and
to do so it must be serious, deep, and detailed. This is a kind of philosophy
that few of us can emulate, but we can venerate it. That is our purpose in
contributing to this volume in honor of Kitcher.
Throughout his career Kitcher (2001b, 65) has taught us that moral
and social values [are] intrinsic to the practice of the sciences. We agree
wholeheartedly. And we think this lesson is of special importance when the
sciences in question are directly relevant to policy issues such as child wel-
fare. And like Kitcher, we also think that values should not enter scientific
practice willy-nilly. As we will illustrate, when morals and methods mix in a
way they should not, undesirable consequencesregarding the welfare of
victims of child abuse, for instancecan follow.
In 2004 in the London borough of Haringey, seventeen-month-old Peter
Connelly was found dead in his crib. The child had suffered fractured ribs
and a broken back after months of abuse at home. His mother, her partner,
and a lodger were jailed for his death. Peter had been seen by health and
social services professionals from Haringey Council sixty times in the eight
months before hedied.
There were two kinds of government responses to this that we shall dis-
cuss. First, Minister of Education Ed Balls sacked the director of Childrens
Services in Haringey, Sharon Shoesmith, with immediate effect in a live
press conference on television. Shoesmith (2011) defended herself and the
Haringey Services in an interview with BBC Radio 4:We should not be put
into blame; it does not result in anything productive and obscures the
bigger picture. The journalist interviewing her argued to the contrary. If
nobody accepts the blame, he asked, then how can we stop this happening
again?
A second response came from Prime Minister Tony Blair (2006) in a lec-
ture delivered on September 5, 2006. He argued that the government can
make children and young people safer by identifying at-risk families and
interveningearly:
Many of our families reported feeling bruised, and some were seriously trauma-
tised, by their encountershowever briefwith social workers, emerging with
De l ib e r at i n g P ol i c y [231]
is unlikely prisoners with Hares characteristics will be paroled, consider-
ing the risk to the publicand the political risk:if a person with known
psychopathic characteristics reoffends, the parole board is in trouble.
So there is little incentive for parole boards to release those with Hares
characteristics.
Yet many moral and religious points of view would hold that people
should not be kept in prison because of predictions about their future
behavior. Prison is for people who have committed a crime (in the past
tense). Those up for parole have already served the minimum sentence for
their crime. Yet for those prisoners with Hares characteristics, other con-
siderations favoring their release on parole are overridden by the results
associating those characteristics with a high recidivism rate. Furthermore,
while the test predicted reconviction rates fairly accurately under the con-
ditions his lab was using, Hare (1998) is concerned that the test is not
as accurate when it is not administered under the same conditions. For
instance, court records suggest that the results of the test are biased by
the person paying to have the test administered: tests administered on
behalf of the prosecution see scores, on average, 8 points higher than those
administered on behalf of the defense, where the higher the score on the
test, the higher the predicted reconviction rate. And there are no guaran-
tees that the test is being administered by a trained psychologist.
Here is another example of a morally questionable policy reported on
the front page of the NewYork Times on March 3, 2013. In an effort to cut
down on robbery, the New York City Police are aggressively intervening
in the lives of certain young people. As the police describe it, they aim to
make them [the young people] radioactive and thus isolate them from
their friends. Which young people? Those destined for trouble and most
likely to commit these crimes (Ruderman2013).
All of these policy responses are morally questionable, and all promise
success based on what we can claim toknow.
A good policy decision always requires a mix of considerations: Who
benefits? Who suffers? Who pays? How much? What are possible good side
effects? Bad ones? Will the effects last? And so on. Central among these are
issues of effectiveness and issues of legitimacy:Will the policy achieve the
desired ends? Is it morally, politically, culturally acceptable? Few policies
will be all to the good for all concerned; few have only a moral upside and
no moral downside. So it is inevitable that a balance be struck. In particular
it can be perfectly acceptable to adopt a policy that is morally questionable
or that has morally negative aspects if we can be sure it will achieve good
ends, so long as the balance is reasonable and we operate within a range of
what is at least morally permissible.
1. Whether the policy will be effective, that is, will produce the expected
effects.
2. Whether the policy is morally, socially, politically, and culturally
acceptable.
But these requirements dont always mix in the way they should, because
we often focus on (1)and discount the importance of (2). Aplausible expla-
nation for this is that we think we have methods, for instance random-
ized controlled trials (RCTs), that will provide us with objective and fairly
certain answers to the effectiveness question. By contrast, we do not have
methods that can give us objective and certain answers to the moral ques-
tion; here things are much muddier and open to debate.
This slidefrom
A. The methods we have for ascertaining (1)are better than the ones we have
for ascertaining(2)
to
is eased along by the prestige that EBP enjoys. The point of EBP is to ensure
that effectiveness predictions are based on sound evidence. But the expec-
tations EBP creates can lead us to give this evidence and the predictions it
De l ib e r at i n g P ol i c y [233]
supports greater weight than concerns of moral acceptability. Effectiveness
can become the cardinal value in policy deliberations, as we think we are
now seeing in UK political calls for coercive interventions into families that
seriously change the state-family relationship.
In the rest of this essay we want to hack away some mistaken philo-
sophical stances that can make effectiveness considerations loom larger
than they should. These involve a circle of mistaken ideas about objectivity,
certainty, and causality:
1. We bank on certainty.
2. We suppose objectivity is the path to certainty because we assume that
objectivity amounts to the elimination of the subject and of judgment,
via the use of methods that have manuals that fix correct procedures.
3. We assume that causality is linear and that it is God-given.
Work Student
feedback motivation
Consistent Student
lessons ability
Study Supportive
space family
De l ib e r at i n g P ol i c y [235]
PROGRAM GOAL ACTIVITIES SHORT-TERM OUTCOMES INTERMEDIATE OUTCOMES LONG-TERM OUTCOMES
Home visits weekly the first month following program Pregnant women display improved
enrollment, then every other week until birth of infant.
Nurse address:
Effects of smoking, alcohol and illicit drugs on fetal growth,
health behaviors.
cigarette smoking
pregnancy-induced hypertension
Nurse-Family
Partnership
and assist women in identifying goals and plans for
IMPROVE reducing cigarette smoking, etc.; use of community resources
pregnancy outcomes Nutritional and exercise requirements during pregnancy
by helping women and monitor and promote adequate weight gain; Helping First-Time Parents Succeed
improve prenatal Other risk factors for pre-term delivery/low birth weight
health Newborns are 37 weeks gestation
(e.g., genitourinary tract infections, pre-eclampsia);
Preparation for labor and delivery/childbirth education;
& weigh 2500 grams or more.
pre-term delivery among smokers
Nurse-Family Partnership
Basics of newborn care and newborn states;
Family planning/birth control following delivery of infant; birth weight among young teens
(<17 years)
Theory of Change Logic Model
Adequate use of office-based prenatal care; and
Referrals to other health and human services as needed. neurodevelopmental impairment
Figure9.3 Acausal diagram for the Nurse-Family Partnership. Each step in the diagram corresponds to a causal cake depicting the requisite support factors.
Source:http://www.nursefamilypartnership.org/assets/PDF/Communities/TOC-Logic-Model.
and causal cakes:what we picture as the cause typically cannot produce the
effect on its own but needs the help of a whole team of support factors. Thats
going to be true for each step in this causal sequence. Theres not just one
causal cake here but a different causal cake for eachstep.
If we want to identify the support team necessary for the initial causes
in the Nurse-Family Partnership program to produce the targeted out-
comes, we have to gather all the members of all the support teams from
each stage and graph them together in one huge causal cake. Recall that the
point about causal cakes is that all their ingredients have to be in place or
you dont get the effect. To the extent that the presence of any of the neces-
sary ingredients is uncertain, so too is that of the final outcome. But look
at our circle of problems. We bargain for certainty. The simple linear causal
model makes this seem to be a far better bargain than it generally is.So:
The linear model and the omission of support factors also predispose us to
focus efforts on eliminating harmful causes at the head of a sequence, such as
family drug and alcohol abuse, which can be a tall order. But it can be just as
effective to remove support factors anywhere along the causal path. Consider
the growing body of research on resilience factors. Resilience describes the
product of a combination of mechanisms for coping in the face of adversity.
Evidence from longitudinal studies suggests that many children can recover
from short-lived childhood adversity with little detectable impact in adult
life. Encouraging resilience is important because resilient children are better
equipped to resist stress and adversity, to cope with change and uncertainty,
and to recover faster and more completely from traumatic events.
Linear models also dont have cycles in them. But cycles can mat-
ter. Consider the UK governments 2011 Munro Review of Child Protection
(Munro 2011), which notes that policies, even good ones, can figure in neg-
ative cakes alongside positive ones. The negative cakes diminish the good
effects of the policy and can even, if they are strong enough, outweigh the
good effects. This is just the trouble that the Munro Review pinpoints for
one of the big UK child welfare policies.
De l ib e r at i n g P ol i c y [237]
The policy in question was intended to improve welfare outcomes
in children and young people (CYP in Figure 9.4) by providing stricter
guidelines for what social workers must do in dealing with children and
families and by better monitoring of what they are doing:by ensuring that
specific mandated facts about the family and the child are ascertained and
recorded properly and that all required meetings take place. But this policy,
the Munro Review argues, can have serious negative effects alongside the
intended positive ones. How so? Through various negative feedbackloops.
Have a look at a diagram from the Munro Review (Figure9.4). Two nega-
tive loops are pictured, R1 and R2. Both start in the same way. Increasing
the amount of prescription imposed on social workers, you can reduce their
sense of satisfaction and self-esteem. In R1, this increases staff sickness
and absence rates; in R2, it increases staff turnover rates. Both these effects
tend to result in an increase in average social worker caseload, which leads
to social workers spending less time with each of the children and young
people and their families. This in turn reduces the quality of the social work-
ers relationships with the children and their families, which then reduces
the quality of the outcomes. So the policy may produce bad unintended
consequences. Worse, these negative effects can become amplified via the
feedback loops. When the outcomes are regularly too unsatisfactory, this
reduces social workers sense of self-esteem and personal responsibility,
and the negative cycle is set in motionagain.
Besides the habit of taking causality as linear, we take it to be God-given.
But the kinds of causal principles we rely on for policy prediction are not
God-given. They depend on intricate underlying structures. And our focus
on objectivity and certainty takes our attention away from these underly-
ing structures. To make this point vivid we use an example not from social
policy but from Nancys own daily policies. Nancy often writes her lectures
on paper with a sharp pencil. She sharpens her pencils by putting a kite out
her study window. She can do that because her study was designed by Rube
Goldberg. You can view this Rube Goldberg machine in (Cartwright and
Hardie 2012, 77). Pulling the kite string via a double pulley opens a small
door in a cage containing moths, which are thereby released. The moths
eat a flannel shirt sitting on a scale, and so forth, till finally a woodpecker
becomes free to peck the pencil sharp. Putting a kite out the window is
a very effective policy for Nancy to get nice sharp pencils. Still we dont
advise you to fly a kite to sharpen your pencils. Kite flying undoubtedly
figures in the causal principles that govern pencil sharpening in Nancys
study. It would, for instance, pass any rigorous RCT. Put the kite out the
window on randomly chosen days and you will certainly get more sharp
pencils when you put it out than when you dont. But that principle is local;
An arrow linking variable A to variable B should be read as a change in the value of A produces a change in the value of B. The qualitative nature of
the link is indicated by a link polarity. These should be read as:
S: the variables move in the same direction ceteris paribus, so a change in variable A produces a change in variable B in the same
direction: if A goes up, B goes up.
O: the variables move in the opposite direction ceteris paribus, so a change in variable A produces a change in variable B in the
opposite direction: if A goes up, B goes down.
double bars on a link indicate a particularly long delay in the causal connection.
Note that the link polarity says nothing about the size, or quantity of the change. The indication of the effect is qualitative only. Moreover, there is no
presumption of a linear relationship between the two variables.
Figure9.4 The complex impact of increased prescription of social work practice. Source:Munro 2011,136.
it depends on the underlying structure of her study. The causal role played
by kite flying in Nancys study is not God-given:it depends on a complex
pattern of interactions in an intricate underlying structure.
Of course this is not a typical social policy case. Social policies suppose
principles like Burnout causes turnover in child welfare service work-
ers or Age does not cause turnover in child welfare service workers. Or
Apathetic-futile mothers are more likely to maltreat their children. These
are clearly not God-given either. And surely it is implausible to suppose
that getting a good social regularity, as these are purported to be, depends
less on the details of the underlying structure than on getting regularities
between pure physical quantities.
Note too that we are not supposing that there are no human universals,
that people in Bangladeshi villages are essentially different from those in
NewYork high-rises, nor across the three hundred language groups and more
than fifty nonindigenous communities of London. In fact our Rube Goldberg
example of a local causal principle works in just the opposite way:by relying
on other causal principles that hold widely. The pencil sharpener depends
on a number of fairly universal principlesfrom the laws of the lever and
the pulley to the familiar fact that moths eat flannelto ensure that the
arrangement and interaction of the components result in the causal prin-
ciple Nancy uses to sharpen her pencils. So the fact, if it is one, that there are
a large number of universal truths about human behaviors, emotions, and
reactions does not demonstrate that the kinds of causal principles we rely on
in typical social policies will be anything other than verylocal.
Our aspirations for certainty divert our attention to these kinds of local
causal principles since they are ones that we can nail down with objective
methods, like RCTs:Flying kites in Nancys study sharpens pencils. Or from
the website of the MIT-based Jameel Poverty Action Lab, after a study in cer-
tain Indian villages:Informing villagers of poor teaching in their villages and
raising awareness of accountability mechanisms had no impact on teacher
attendance.1 Our efforts are taken away from the more difficult study of the
underlying structures that make these causal principles possible.
When it comes not to prediction but to evaluationlooking back to see
what was responsible for an outcomethe focus on linear causal princi-
ples, with their objective certifying methods, leads to skewed views about
human error and individual responsibility. In policy evaluation, just as in
policy prediction, methods and morals often mix in a way they ought not
1. Abdul Latif Jameel Poverty Action Lab, Healthcare Provider Attendance, http://
www.povertyactionlab.org/policy-lessons/health/healthcare-provider-attendance
(accessed December 16,2105).
There are two ways of viewing human error:the person-centred approach and
the system approach. The [person-centered]... approach focuses on the psycho-
logical precursors of error, such as inattention, forgetfulness and carelessness.
Its associated countermeasures are aimed at individuals rather than situations
and these invariably fall within the control paradigm of management. Such
controls include disciplinary measures, writing more procedures to guide indi-
vidual behaviour, or blaming, naming and shaming.
This is just what we saw in the case of Sharon Shoesmith. As is also noted
in the pamphlet, however,
aside from treating errors as moral issues, [the person-centered approach] iso-
lates unsafe acts from their context, thus making it very hard to uncover and
eliminate recurrent error traps within the system....
The system approach, in contrast, takes a holistic stance on the issues of fail-
ure. It recognises that many of the problems facing organisations are complex,
ill-defined and result from the interaction of a number of factors. (2021)
The title of this report encapsulates its purpose. Human beings, in all lines of
work, make errors. Errors can be prevented by designing systems that make it
hard for people to do the wrong thing and easy for people to do the right thing.
Cars are designed so that drivers cannot start them while in reverse because that
De l ib e r at i n g P ol i c y [241]
prevents accidents. Work schedules for pilots are designed so they dont fly too
many consecutive hours without rest because alertness and performance are
compromised.
The report urges, The focus must shift from blaming individuals for past
errors to a focus on preventing future errors by designing safety into the
system (5). To put this in the terms we have been using, we should be
less concerned with the easier-to-certify causal sequences that start with
human error and end with disastrous consequences and far more con-
cerned with understandingand restructuring the underlying struc-
tures that make this kind of causal sequence likely.
As Munro (2005, 378)notes, When society is shocked and outraged by
a childs terrible tale of suffering, there seems a basic human desire to find
a culprit, someone to bear the guilt for the disaster and to be the target
of feelings of rage and frustration. This puts us squarely in the business
of finding these local linear causal principles, and, with Tony Blair, we can
feel morally and epistemically safe in doing sowe are not likely to cast
blame in the wrong placesbecause these are the kinds of claims about
which, with due care, our objective methods can deliver reasonable cer-
tainty. But the kinds of preventative measures this leads torecall the UK
Department of Health examples:disciplinary actions; writing more proce-
dures to guide individual behavior; or blaming, naming, and shamingare
often unlikely to stop these kinds of sequences occurring. As Munro urges,
Child protection is a systems problem (375). So too are a good many other
social problems, from poor child nutrition in Bangladesh and poor school
attendance by teachers in Indian villages to crime, education, health, and
climate change adaptation almost anywhere. Our thirst for certainty and
our admiration for methods that can be run by rules must not lead us to
buy cheap knowledge that cant serve ourneeds.
For a timely illustration of linear causal thinking in child protection,
consider the investigation into the death of Daniel Pelka, a four-year-
old who died at the hands of his mother and stepfather in March 2012
in Coventry, UK. His death provoked a massive outcry across the United
Kingdom in large part because, according to the Final Overview Report of
Serious Case Review re Daniel Pelka, commissioned by the Coventry Children
Safeguarding Board (CSCB 2013), the social workers (as well as teachers
and police officers) who had been in contact with Daniel and his family
in the months leading up to his death missed twenty-six opportunities to
help him and to act in a way that would have prevented hisdeath.
Because the Final Overview Report blames individuals and it blames
them for failing to behave in certain ways, the lessons it draws are about
In this case, professionals needed to think the unthinkable and to believe and
act upon what they saw in front of them, rather than accept parental versions
of what was happening at home without robust challenge. Much of the detail
which emerged from later witness statements and the criminal trial about the
level of abuse which Daniel suffered was completely unknown to the profession-
als who were in contact with the family at the time. (CSCB 2013,6)
De l ib e r at i n g P ol i c y [243]
means fewer false negatives (and correlatively, it seems safe to assume,
more false positives, with more families subjected to nasty investigations
and sometimes losing their children at a lower level of evidence). From the
evidence in the Final Overview Report itself, the Daniel Pelka case has the
earmarks of a systems issue:it is the result of many minor flaws in practice
plus the basic incompleteness of information rather than one devastating
individual act. Each failure of each individual was just one of many ingre-
dients in the causal cake that led to the failure to see how much danger
Daniel was in. So it is not obvious that improving the detectors of abuse in
that system is the best way to improve child welfare and reduce the prob-
ability of such abuse occurring again. Perhaps it would be better to redesign
the system so it is easier to do it right and harder to do itwrong.
It should also be noted that the lessons the Final Overview Report pro-
poses to turn into rules of best practice are inherently local in character,
so that following these rules might in fact be harmful. The Report seems
to have used the following process to arrive at a recommendation for best
practice:Professionals are to be blamed for failing to act in way W when
doing so would have prevented harmful outcome O in the Pelka case.
Therefore professionals should, as a rule and whenever relevantly similar
situations arise, act in way W.Note, moreover, that since the failure of the
professionals to act in way W is perceived to be a moral failure, the rule
stating that one should act in way W becomes a moral rule. The issue here is
that a rule the application of which might have prevented harmful outcome
O in the Pelka case might not contribute to preventing the occurrence of
harmful outcomes of the same kind in other contexts. In fact if applied
widely it might contribute to promoting the occurrence of such outcomes.
Consider as an illustration of this worry the case of lesson 15.8 (CSCB
2013, 72):Any facial injuries to a child must be viewed with concern, with
physical abuse needing to be actively considered as a possible cause, and clear
records, interventions or referrals made accordingly. As Munro remarks
(Pemberton 2013), bruises to the head arent that uncommon in four-year-
olds, which means that strictly following 15.8 might lead to a great number
of false positives, that is, of cases in which parents are wrongly suspected of
abuse. Further, wrongly stigmatizing parents as abusing their children may
decrease the welfare of children in the population at large, for instance if a
large number of false positives leads families to distrust social workers.
The potential harmful effects of lesson 15.11 (CSCB 2013, 73)are even
more striking:
For professionals from Childrens Social Care or the Police to defer to medical
staff for the provision of the primary evidence to confirm or otherwise whether
De l ib e r at i n g P ol i c y [245]
food distribution. In Tamil Nadu there was a generalizable cause of improved
infant health:the nutritional education of a person (i)who is responsible
for household food selection, and (ii) who is responsible for food distribu-
tion, and (iii) who holds the infants welfare paramount in carrying out these
responsibilities. The Tamil Nadu study asked about mothers. And in Tamil
Nadu mothers refers to a class of people with the requisite characteristics to
improve infant health via nutritional education. But not in Bangladesh. If we
wanted more generalizable results, we were asking the wrong question.
The second example is another case from UK child-welfare policy, not
one that we currently have strong evidence for but where there seems to be
cause to worry. In many cases a childs caregivers, though not legally com-
pelled, are encouraged to, even badgered into attending parenting classes.
This includes fathers. But what constitutes a father? Is father instanti-
ated by biological father, or male partner of the mother who lives in the
household with the child, or maybe male caregiver? Maybe the policy will
be effective if the male caregivers or men living with the mother are tar-
geted, but not biological fathers. If so, to focus on being a father would be
to move to too high a level of abstraction since only the more specific male
caregiver or male partner of mother who shares the childs household
enters into a reasonably reliable principle. On the other hand compelling
the male caregiver to attend classes can be too concrete. Different cultures
in the United Kingdom have widely different views about the roles fathers
should play in parenting. Compelling fathers to attend classes can fall under
the more abstract description ensuring caregivers are better informed
about ways to help the child, in which case it could be expected to be posi-
tively effective for improving the childs welfare. But it may also instantiate
the more abstract public humiliation, in which case it could act negatively.
And of course it can fall under both at once. In any case, if the two more
abstract features pull in opposite directions, there will be no reliable prin-
ciple to formulate at the more concrete level involving fathers. Nor is this
pull in opposite directions an unrealistic hypothesis. We know from empiri-
cal research that there are varying outcomes associated with compelling or
strongly encouraging parents to attend parenting classes and also that these
are correlated with varying motivations (see, e.g., Barlow etal.2006).
Let us begin now to tie matters together. We have been criticizing three
mistaken, mutually supporting philosophical stances that contribute to
bad policy decisions:
1. We bank on certainty.
2. We do not trust the kind of open, multimethod, theory-infested scien-
tific investigations it takes to uncover the structures of these underlying
2. With the help of Nate Rockwood. Both Nancy Cartwright and Alexandre Marcellesi
would like to thank the UK AHRC project Choices of Evidence: Tacit Philosophical
Assumptions in Debates on Evidence-Based Practice in Childrens Welfare Services
and Eileen Munro in particularfor support for the research and writing of thisessay.
De l ib e r at i n g P ol i c y [247]
Reply toCartwright and Marcellesi
PHILIP KI TCHER
I have been learning from Nancy Cartwrights work, and from conversa-
tions with her, for more than thirty years. Sometimes the readingsand
especially our face-to-face exchangeshave planted seeds that have con-
tinued to flower in my thinking. In the essay they have written for this
volume, Nancy and Alexandre Marcellesi continue that line of interaction,
asking me to consider aspects of the uses to which scientific research is put
on which Ihavent previously focused. This reply will take up the invitation.
In my 2001b book, and more extensively in my 2011b book, Im con-
cerned with Science as an institution that interacts with other institutions
in contemporary societies (those that characterize themselves as democ-
racies). The philosophical project I conceive is melioristic in character: it
surveys the interacting institutions in hopes of improving their joint func-
tioning. (Regrettably Ihavent always been as forthright as Imight have
been about the pragmatism of my approach, inspiring some commentators
to interpret me as being more utopian than Iaim to be; my reply to John
Dupr tries to say some things I should have said earlier.) The center of
my interest has been Science conceived as a system of public knowledge,
potentially available to radiate out into individual and social projects. Ive
been asking what might valuably be done to improve that system, in its
ways of setting the research agenda, in its modes of certifying potential
novel findings, and in the channels through which certified results are
distributed. In her recent work Cartwright and her coauthors have been
probing many instances of a specialand vitally importantclass:public
policy decisions to investigate questions that bear directly on human wel-
fare, to demand particular methods for certifying answers, and to set up
mechanisms for intervening on the basis of the certified results.
One of Cartwrights enduring contributions to our discipline, encapsu-
lated in her brilliant 1999 book, is the recognition that the sciences provide
a bundle of loosely connected domains, in which nature goesor more
usually, can be set up to goreliably and predictably. Cartwrights term
for these is nomological machines; Ill call them spheres of order. She
reminds us, rightly, that these spheres are local, that they cannot be easily
extended to operate smoothly on a wider scale, and that their functioning
depends on a variety of special conditions (of which we often only slowly
and painfully become aware). This aspect of her views elaborates ideas
present in Neurath (to whom she is explicitly indebted) and in Dewey.
Deweys ([1925] 1981, [1929] 1984)account of the role of the sciences in
De l ib e r at i n g P ol i c y [249]
researchers doing socially relevant science follow certain supposedly privi-
leged methodsfor example, doing RCTs on populations that lend them-
selves to studywhich they pass on to bureaucrats who design broad-scale
policies without considering either the possible unintended effects of the
extrapolation or the ethical issues that inevitablyarise.
So we have a second- order misadventure, generating first- order
attempts at social engineering that repeatedly go awry. Viewing the situ-
ation in those terms poses a third-order question, one for philosophers
of science: How would it be possible to make the second-order domain,
in which researchers and policymakers interact, function better? Ive no
doubt that the status quo might be improved if those involved in shap-
ing the research agenda and pursuing inquiries of the pertinent types were
thoroughly cognizant of the Munro (2011) report, as well as Cartwright
(1999), Cartwright and Hardie (2012), Douglas (2009), and the essay
Cartwright and Marcellesi have written for this volume. Yet Idont think
this is sufficient. In the remainder of this reply Ill draw on my reading of
Cartwrights work, as well as my own ideal of well-ordered science, to make
an attempt at pragmatic meliorism.
The sacking of Sharon Shoesmith illustrates the procedures policymak-
ers tend to follow when terrible effects come to light. Reluctant to ques-
tion the system, officials look for a culpable actor whose behavior caused
the damage. Cartwright and Marcellesi want policymaking to rely less on
taken-for-granted (objectively established) results. Yet if the source of the
trouble is an unwarranted extrapolation from cases where order is achieved
by reliance on unrecognized local conditionsthe kind of fundamental-
ism Cartwright (1999) exposeshow is a better policy to be crafted? Lets
suppose that responses to breakdowns in existing policy are less peremp-
tory than they were in the case of Peter Connelly, that theres a serious and
thorough inquiry to understand the actions of the relevant actors (and the
burdens policy places on them and the limitations of their possible impact
on the circumstances with which they must deal). Lets even suppose that
the inquiry involves participation of people with different viewpoints,
social workers and parents, as well as academic investigators and bureau-
crats. Those who take part have digested the insights of Cartwright and her
coauthors, and they conclude that there is indeed a systematic problem.
Shoesmith and her staff are exculpated.
What next? It might turn out that there are ideas in existing psycho-
logical or sociological studies suggesting what could have gone wrong. In
that case there would be an indicated direction for amendment. Even so,
those appreciative of the Cartwright-Marcellesi critique would quickly
recognize the high probability of those studies also being dependent on
De l ib e r at i n g P ol i c y [251]
people. (The Rawlsian device of the original position seems to generate the
same result, allowing an institution for undertaking social experiments.)
Yet Id also expect them to want to amend our actual institution
which is, effectively, as Cartwright and Marcellesi make clear, one that
introduces attempts at order without advance knowledge of the range
of their effectsand not only by supplying the existing decision makers
with information about the complexities of causation. Proposals for social
experiments ought to involve those most likely to be affected by the ways
the planned sphere of order is to be constructed. Representatives of fami-
lies from different socioeconomic classes and with different structures
should be involved in the conversation about child welfare, as should the
social workers who will be asked to implement the envisaged policies.
Experimental review should be frequent and attentive to the possibilities
of various kinds of damaging consequences, not merely to the problem
for which the policy seeks a remedy.
Nor is this enough. Many of the examples Cartwright and Marcellesi
present consider policy interventions introduced in isolation from
related problems. This is most evident in relation to crime, where profil-
ing is defended by appeal to statistics, without any consideration for
background circumstances that might lie behind the numbers cited. To
inaugurate or continue a policy of stopping young men of a particular
race, without doing additional experiments to see what might be done
to increase the opportunities and prospects of members of the targeted
group, is to conduct an illegitimate experiment, one that an informed,
mutually engaged, comprehensive deliberation would almost certainly
reject.
If, as Ibelieve, we are committed to experimentation aimed at creating
valuable order in social domains, we are also committed to monitoring the
experiments and to recognizing the injustice of conducting certain kinds
of trials in isolation from complementary efforts. To reform the research-
policy institution whose current methodological defects Cartwright and
Marcellesi expose requires systematic attention to the background social
context, recognition that pervasive features of the society compromise
the ethical standing of many widely accepted social experiments. Only in
the context of a cluster of policy attempts may an apparently attractive
individual proposal count as just. If it were subjected to the kind of con-
versation well-ordered science demands, policymaking would be directed
to a package of trials that would better promote justice. Methods and
morals really do mix. We need both the causal insights Cartwright and
her coauthors offer and a more comprehensive conversation.
I n the first part of The Ethical Project Philip Kitcher (2011a) provides a nat-
uralistic account of the function of ethical practice. Though he conceives
of this as an unashamedly empirical taskhe denies that the function
of ethical practice is to get at the ethical truth or to get people to behave
in accordance with the ethical truthit is not one that he thinks has a
debunking upshot. Instead he thinks that his account of the function of
ethical practice, which is based on his reading of the practices origins in
the behavior of our hominid ancestors, provides us with all of the materials
required for a pragmatic naturalist account of ethical truth. More familiar
a priori attempts to ground ethical truth in the spirit of Kant are given
short shrift, as are the supposedly more empirically adequate noncognitiv-
ist accounts of ethical judgment which hold that such judgments express
emotions, commitments, intentions, desires, or noncognitive attitudes of
some otherkind.
Though there is a great deal to admire in The Ethical Project, especially for
someone like me, who was so antecedently ignorant of the relevant science
and history, Iam afraid that Iremain unconvinced. My discussion is divided
into three main sections. In the first I consider Kitchers account of the
function of ethics and his pragmatic naturalist account of ethical truth. If
ethical truth really were fixed by the function of ethical practice in the way
he suggests, then this would make the truth of ethical claims turn on issues
that seem to me, and Iwould have thought to Kitcher too, to be quite irrel-
evant. It would also mean that many of our ethical beliefs, including some
of those required to maintain ethical practice so understood, are false. This
doesnt entail that his account of ethical truth is mistaken, but it does sug-
gest that something might have gone wrong, and Isuggest what that some-
thing might be. In the second section Iargue that Kitchers account of the
function of ethical practice does not commit him to a pragmatic naturalist
account of ethical truth. Noncognitivism remains an alternative, notwith-
standing his reservations. Since it does a much better of job of shoring up
our ethical commitments than his own pragmatic naturalist account, Isug-
gest that Kitcher would be better off embracing some form of noncognitiv-
ism. In the third and final section Ibriefly consider Kitchers objections to
Kantian approaches to ethical truth, and Iexplain why at least one version
of this approach, the one that Imyself prefer, is not vulnerable to his objec-
tions. This too remains an option for someone who accepts his account of
the function of ethical practice.
Peace and mutual tolerance are typically hard-won. Precisely because of this,
observations of chimpanzee societies disclose periods of intense social interac-
tion, lengthy bouts of grooming undertaken to reassure friends who have been
disappointed by recent behavior. At times of great tension within a group, chim-
panzees can spend up to six hours a day huddled together, vastly longer than
any hygienic purpose demands. (Kitcher 2011a,73)
What Kitcher says here sounds like a credible bit of ethical reasoning, but
we arent officially supposed to be in the business of evaluating the thought
that it is good to relieve pain, wherever it occurs, in terms of its intrinsic
plausibility. That would be for progress to be ethically characterized. Instead
we are supposed to evaluate it in terms of its power to serve the function
of ethics, where that function is to fix the content of rules whose adoption
would remedy altruism failures that would interfere with cooperation.
Seen in this light it seems quite implausible to suppose that the rule we
have to adopt, for ethical practice to serve its function, is the injunction
to relieve pain, wherever it occurs. It is quite implausible because we are
evidently able to ignore vast quantities of pain, both human and nonhu-
man, without that having any affect at all on the levels of cooperation we
enjoy. A failure to extend altruistic concern to nonhuman animals quite
generally would thus seem to leave cooperation largely intact. Moreover to
the extent that a failure to extend our altruistic concern to nonhuman ani-
mals in some limited way would have an effect on cooperation, the explana-
tion of this would have to be that a failure to extend such concern would
interfere with cooperation, not the truth of the claim that it is good to
relieve pain wherever it occurs. In the first instance what is important is
the cooperation of other humans, not the alleviation of nonhuman animal
pain. Indeed Itake it that the claim that it is good to relieve pain, wher-
ever it occurs, thus turns out to be false, on Kitchers way of understanding
This may or may not be a sound account of how the feminist movement origi-
nated, but the point is that Kitcher interprets the question Why do we believe
that women and men have equal rights? as a request for causal explanation,
not as a demand for justification. Feminism is justified not because women want
to be treated like men and because it so happens that granting them equality
doesnt cause too much damage to family life:it is justified because of the equal-
ity of men andwomen.
How should Kantians respond? The first thing Kantians should say is,
Ihope, apparent from the earlier discussion.
According to Kitcher, the Kantians response to the skeptic reduces
to the claim that there is something wrong with him, and this means
that the Kantian and the pragmatic naturalist are on an equal footing as
regards the skeptic. They agree that there is something wrong with him,
but disagree about what that thing is:the Kantian says that the skeptic is
practically irrational, whereas the pragmatic naturalist says that he has an
inability to appreciate how central the ethical project is to human life. The
issue should therefore be decided by figuring out whose account of whats
wrong with the skeptic is better supported by the evidence. But as the
earlier discussion made clear, there isnt obviously anything wrong with
someones having the inability that the pragmatic naturalist identifies. To
be sure, the ethical project, as Kitcher describes it, is central to human life
to the extent that it sets the terms for stable cooperation. But given that
stable cooperation might itself be ethically problematicremember again
our discussion of what might underwrite the stable cooperation we find
between men and womenthat doesnt count in favor of anyones appreci-
ating that kind of cooperation, in the sense of their being disposed to bring
it about. We should therefore be skeptical of the ethical project, insofar as
we agree that it has the function Kitcher identifies.
This leaves us with Kitchers disparaging remarks about Kantian
approaches at the very beginning of the passage. Kitchers skeptic says,
You can call the procedures you use to generate the rules you favor pure
practical reason, if you like, and suppose those who dont go along with
them are involved in some sort of contradiction, but the mere label doesnt
frighten me, and the effects you envisage dont appear particularly dread-
ful. That isnt just the skeptic talkingthats Kitcher talking. As Iunder-
stand it, he is saying that the Kantian doesnt have a plausible story to tell
about what the practical irrationality of the skeptic consists in. The Kantian
1. That agents qua agents have the function of being desire realizers and
knowledge acquirers (analytic truth about agents).
2. That it follows from (1)that ideal agentsthat is, agents who perform
their function optimallymust have and exercise the capacity to realize
their desires no matter what their content and know what the world is
like, no matter what it is like, at least insofar as the worlds being the
way that it is bears on their realizing their desires (the modal conception
of the ideality of a functionalkind).
3. That it follows from the potential for conflict in the optimal possession
and exercise of these two capacitiesthink of the conflict present in an
otherwise ideal agent who desires to believe that p whether or not it is
truethat ideal agents must have certain dominant conflict-resolving
desires: specifically ideal agents must have a dominant desire not to
interfere with their exercise of their capacities to know their world or
realize their desires in it (on condition that those desires wouldnt lead
them to interfere), whether in the present or the future, and they must
also desire to do what they can to help ensure that they have the capaci-
ties to know their world and realize their desires in it, whether in the
present or the future (novel premise whose truth depends on (2), the
fact that agents are temporally extended, and a painstaking examina-
tion and rejection of alternative ways in which the conflict identified
might be resolved).
4. That there is no consistent way for an ideal agent to restrict his desires
to help and not interfere to the current and future stages of himself, and
hence that ideal agents have expanded versions of these two dominant
desires ranging over all agents:that is, they desire to help everyone and
not to interfere with anyone (a premise inspired by Nagel [1970] and
Parfit [1984] that teases out the implications for the theory of practi-
cal rationality of the fact that each agent is [potentially] just one agent
amongmany).
5. That there is an analytic tie between facts about an agents reasons for
action and facts about which of that agents options his ideal counter-
part has desires for (analytic truth about reasons for action).
Reply toSmith
PHILIP KI TCHER
W h at t o D o W h i l e R e l i g i o n s E volv e b e f or e O u r V e r y E y e s [275]
preempt the controversy by taking one side. This is the tactic that dare not
speak its name; it belongs to a family of tactics that spread in a continuum
from indefensible (lying for Christ is an example) to indispensable (respect-
ing the privacy and feelings of others to the greatest extent possible). Some
relatively benign points on this continuum can be justly called diplomacy or
tact, and other points, over near the other extreme, should be called what
they are: intellectual dishonesty at its most culpable. One of the clearest
results of my ongoing research with Linda LaScola on the plight of nonbe-
lieving clergy (Dennett and LaScola 2010, 2013, 2015) is the recognition
that the slippery slope between these extremes has no boundaries to which
one may cling in a principled way. What begins as tact can sour into dissem-
bling that shatters the integrity and blights the lives of many clergy. I have
coined a deliberately mildcomically mildterm for the middle ground,
faith fibbing, which is not so harsh that people find it impossible to confess
to, but not so benign that we can all publicly advocate it as a policy without
undermining it in the process. No one would occupy a pulpit adorned with
the declaration Faith Fibbing Practiced Here.
I went to some lengths in Breaking the Spell (Dennett 2006) to distin-
guish two spells one might consider breaking: the taboo against looking
too closely at religion, holding it up to the same harsh light of rational
probing to which we subject all other important phenomena, and the spell
of religion itself. In my book I declared my intention to break the first spell
and my agnosticism about the wisdom of breaking the secondciting the
very considerations that Kitcher advances more positively. Kitcher ignores
my distinction but in fact is in nearly perfect harmony with my positions
on them. His essay is an example of breaking the first spell: he writes with
unflinching candor about the shaky status of any religion adopted on what
he calls the belief model and uses that spell-broken perspective to look hard
at the prospects for keeping the second spell unbroken by relying on what
he calls the orientation model, supposing that this is perhaps the only sur-
viving mode of religion that can provide the benefits he wants to preserve,
which may just be a necessity of meaningful life for many people. As I noted
in my book, there is a reasonable fear that breaking the first spell will inevi-
tably break the second as well, which fear is the obligatorily tacit standard
justification for not breaking the first. Kitcher vividly illustrates that prob-
lem in his essay, trying to split the difference between being patronizing
on the one hand and uneasy complicity with unacceptable nonsense on the
other.
The point of Kitchers introduction of the orientation model is to give
him a way of reversingmost of the timethe otherwise standard depen-
dence of serious commitments and aspirations on grounded beliefs. The
W h at t o D o W h i l e R e l i g i o n s E volv e b e f or e O u r V e r y E y e s [277]
discouraged among religious spokespeople. There are some famously arro-
gant Nobel Laureates in the scientific community, but Ihave never encoun-
tered one who can hold a candle to the overweening overconfidence, the
smug certainty of the typical self-righteous defender of religion. Their dis-
dain for evidence-seeking and careful argumentation is often breathtak-
ing, amounting to self-disqualification for the role of rational discussant.
Those who view it as positively immoral to entertain alternatives to, or
even objections to, their faith place themselves outside the marketplace
of ideas, incompetent to participate in the serious political conversations
that we ought to be engaging in today. And one of the effects of fostering
epistemic modesty in all matters religious, as Kitcher recommends, is pro-
tecting the social niche in which these influential subverters of reason can
operate largely unchallenged.
I have been asking defenders of sophisticated theology for a recom-
mended reading list, for works they are prepared to defend as intellectually
bracing and honest, and have yet to have my challenge met. Iam tempted
to draw the conclusion that, on closer examination, they recognize that
they have indeed been adopting a double standard and letting pass as deep
thought work that is actually just obscure, and apparently often deliber-
ately obscure. These works do serve a useful purpose for the adopters of
epistemological modesty who can reason as follows:These professors are pro-
fessional thinkers about religion; they are still in the church, so they must have
gone way beyond me in thinking these issues through. I dont get it, but they
do, so I should accept their authority. In other words, these works provide
examples of high-flown rumination that one can confess to finding some-
what incomprehensible but nevertheless deem inspiring and authoritative.
Go read the meticulous arguments of this thinker; they should sweep away your
doubts. (And if they dont, it must be your fault.)
Kitcher (2011c, 9)is well aware of the risks entailed by adoption of the
orientation model, and Im happy to say that his defense of the three con-
veniently smudged alternatives firmly draws the line at letting any of these
options abrogate a commitment to reason when deciding ethical mat-
ters:Someone who makes decisions affecting the lives of others is ethically
required to rely on those propositions best supported by the evidence.
This requirement runs into awkward interference from the orientation
model, however, as Kitcher reveals in discussing one of the central problems
with the Abrahamic religions (Judaism, Christianity, and Islam):Abraham!
Abraham is called upon by God to perform an unthinkable act (to commit an
unspeakable crime, bluntly put). Is this any kind of inspiring example? In a
felicitous phrase Kitcher (2011c, 8)notes that there ought to be no teleo-
logical suspension of the ethical, but then just what is the positive role of
If you start with the thought that the predominance of religion in human societ-
ies is to be explained by a cognitive deficiency, you will tend [my italics] to see
your campaign for the eradication of myths in terms of a return to intellectual
health. . . . By contrast, if you suppose that the social factors towards which
Ihave gestured have played a non-trivial role in the spread of the worlds reli-
gions, you will wonder [my italics] if there are psychological and social needs that
the simple abandonment of religion will leave unfulfilled.
W h at t o D o W h i l e R e l i g i o n s E volv e b e f or e O u r V e r y E y e s [279]
I discuss this in Breaking the Spell (Dennett 2006, 28692), where Inote
that religion has the unparalleled capacity to give people a chance to be, in
Kitchers good phrase, important participants in the world they were born
into. But as Igo on to discuss there, nobody has yet estimated what price
we should be prepared to payin xenophobia, violence, the glorification of
unreason, the spreading of patent falsehoodfor that wonderful sense of
importance religion gives to many people who would otherwise lead lives
without drama, without a point. Kitcher wants to preserve religions (at
least for the foreseeable future, I gather), but I think it would be better
to work constructively on secular institutions that can provide alternative
structures of meaning for everyone, hastening the day of religions demise.
Still we might accomplish this most practically by encouraging existing reli-
gious institutions to evolve into former religions. Some have already
done so, but they are not yet competing very well in the marketplace of
allegiances. That may soon change.
The transparency of information engendered by electronic media has
dramatically changed the epistemological environmentthe environment
of knowledge, belief, error, illusion, confidencethat we all inhabit. It
threatens the security and stability of all institutions that depend on con-
fidence and trust, which includes such disparate entities as newspapers,
banks, hospitals, religions, and universities. If a reliable source of news
loses its reputation for telling the truth, it may be out of business, no mat-
ter how scrupulously it checks the facts it publishes. So a new arms race is
now ensuing, dealing in the manipulation of reputations for truth-telling,
and its campaigns can be detected on all sides. Al Jazeera has an excellent
and deserved reputation for truthful reporting in most of the world out-
side the United States. Will its recent acquisition of Al Gores news website
finally secure its respectability in this country or damage Gores reputa-
tion? Time magazine continues its print edition in the United States, while
Newsweek abandoned its US edition in 2012, choosing to link its fate to
the Daily Beast website. In recent decades both magazines have experi-
mented with the ploy of bolstering sales by running favorable cover stories
on religious topics:the Shroud of Turin, the Dead Sea Scrolls, the Gnostic
Gospels, new interpretations of putative relics and archaeological discover-
ies. They clearly have seen the security of religious institutions as a possible
lifeboat to cling to, but it is not obvious that their design choice was wise,
in retrospect.
Oxford zoologist Andrew Parker (2003) has proposed in his book, In
the Blink of an Eye that the famous Cambrian Explosion of novel life forms
that occurred about 530 million years ago was triggered by a change in
the chemistry of the atmosphere or the seas or both, which increased the
W h at t o D o W h i l e R e l i g i o n s E volv e b e f or e O u r V e r y E y e s [281]
of teenagers will be Bible-believing Christians as adults (Goodstein 2006;
see also Spencer 2009). Not without good reason do defenders of religion
inflate the numbers of their adherents and concealas best they canany
negative trends in attendance. For the same reasons, of course, colleges
and universities neglect to report declining enrollments and applications
while trumpeting any gains but among religions there are few gains to her-
ald. Muslims and Mormons have increasing populations, due mainly to dif-
ferential birth rates, not conversions, but no religion at all is the fastest
growing category, worldwide.
As these and a host of other such facts become more widely knownand
known to be knownit will be hard for religious spokespeople to maintain
their traditional tone of authority. Religions that flourished in the murky
epistemic waters of earlier millennia are going to find themselves increas-
ingly vulnerable to impertinent questions about their practices, their fund-
ing, their creeds. Recent history has shown us that cover-ups that used to
succeed in the past now have a way of imploding spectacularly. Will the com-
mitment model that Kitcher recommends be able to deflect the scrutiny that
would undermine it? Who knows what the near future will bring? Religions
have changed more in the past century than in the past millennium, and
perhaps they will change more in the next decade than in that past century.
Kitcher and Iagree on so much. We agree that public reason must be
thoroughly secular (Kitcher 2011c, 12). We agree that the belief model of
religion is indefensible. We agree that the first spell must be brokenwe
have both broken it. We differ, apparently, only in our assessment of how
to ease the people of the twenty-first century into a more reasonable and
socially benign form of organization. Its like the problem of how to remove
an adhesive bandage: slowly, gently lifting, pulling, pausingor a swift,
well-timed yank. Ifavor the quick shock and its overnot so bad, and now
lets get on with our lives. Kitcher favors the glacial approach, and whatever
there is to be said for it when calculating the costs and benefits, it is impor-
tant not to overlook the suffering of those who get caught in the pulpit,
slowly accumulating a deplorable history of dissembling and obfuscation
in the name of tact. We all indulge in those little white lies, and it would be
heartless to forswear them, but how much is too much? Ithink we should
both admit that we havent yet figured thatout.1
1. Portions of this essay are revised from Kitcher versus Dennett:Is New Atheism
Counterproductive?, Why Evolution Is True, October 7, 2010, https://whyevolu-
tionistrue.wordpress.com/ 2 010/ 10/ 07/ k itcher- versus- d ennett- i s- new- atheism-
counterproductive-2/, and from Dennett and LaScola 2013 and2015.
For four decades now Ive been a great admirer of Dan Dennetts philosoph-
ical work. His concerns that philosophy should be broadly accessible and
that it should focus on questions that interest more than a small coterie of
professionals are sources of inspiration. Dans ability to combine stylistic
elan with clarity and precision sets a model for us all. I have often been
encouraged by the fact that he and Ihave stood shoulder to shoulder on
many issues. As Dan points out, we agree on somuch.
With respect to religion our disagreements resemble those of the
Bolsheviks and the Mensheviksall the more intense, perhaps, because
the differences might appear, to an external observer, to be so small. In the
end, however, Ithink they are more important than simple divergences on
strategy. We are divided not only about the value of religion but also on
some basic questions about belief.
Dennetts characteristically lucid and forceful essay engages with the
argument of my Militant Modern Atheism (Kitcher 2010; also Kitcher
2012c, ch. 12). Dennett responds to my contrast between the belief and
orientation models by arguing that, once the orientation model is made
explicit, those whose religious commitments conform to it are placed in
an uncomfortable position. They have to decide exactly where they stand
on matters of doctrine: are they mythically self-conscious, doctrinally
indefinite, or doctrinal[ly] entangle[d]? Unless they fall into the last cat-
egory, in which case they are lapsing into the belief model, they are forced
to realize that they are kid[ding] themselves. When religion backs away
from full-blooded commitment to doctrines, the devout have to fess up to
faith fibbing.
To my mind this response is revealing, for it shows how Dennett thinks
about belief. His writings about religion tacitly adopt a pair of theses,
shared by the less sophisticated members of the Four Horsemen (I owe
the whole quartet an apology for having mistakenly suggested that they
proposed thisname):
W h at t o D o W h i l e R e l i g i o n s E volv e b e f or e O u r V e r y E y e s [283]
2. Believing truly is an ultimate value, one that can override other values.
Replacing false belief by true belief brings gains that outweigh whatever
losses are involved in the replacement.
In my later work on religion (Kitcher 2014), Ive called both theses into
question, insisting on the need to recognize that some valuable linguistic
practices stand in need of interpretation and that arriving at a state of clear
and definite true belief may not be worth the sacrifices required to achieve
it. My aim in what follows is to elaborate these points.
Dennett imagines religious believers who are dumbfounded by the
request to situate themselves with respect to the orientation model.
Perhaps he and I move in different circles, but I know many thoughtful
people who would have no trouble with the question. Some, like the late
Robert Bellah, would identify themselves as mythically self-conscious.
They recognize that their creedal professions arent intended to record
matters of historical fact, and they are happy to talk of their doctrines as
myths (Bellah 2011)and even to claim that clashing myths can count
as true. Chapter3 of Kitcher (2014) offers a philosophical reconstruction
of this usage, distinguishing alternative approaches to truth in different
domains (see my replies to Gideon Rosen and to Michael Smith). To put
the point very simply, religious statements are viewed as metaphorical, as
gestures toward a different aspect of realitythe transcendentand valu-
able because of their fruits forlife.
Others, probably the majority of those who profess liberal (or, as
Iwould characterize it, refined) religion, would answer Dennetts question
differently. They would point out, from the beginning, that the statements
in question are complex, in need of interpretation. Some of them might
invoke particular interpretations they favor, interpretations that separate
the content of belief from any literal commitment to supernatural goings-
on. All would deny that the interpretations they can give, or even the inter-
pretations that have been provided in the history of their religion, exhaust
the significance of the doctrinal statements. Many would likely concur with
the attitude recommended by Bellah, seeing all the worlds major religions
(and perhaps the less prevalent ones as well) as gesturing toward an aspect
of reality that literal language cannot capture, an aspect that deserves the
emotional responses William James ([1902] 1982, lecture 2)saw as central
to religion (solemnity and awe, combined with a joyous acceptance). Faced
with the suggestion that they are kidding themselves or faith fibbing,
they would reject the description (possibly indignantly); resorting to meta-
phor or allegory, they would explain, is not the same as lying, particularly
when there is no other way to express important insights.
W h at t o D o W h i l e R e l i g i o n s E volv e b e f or e O u r V e r y E y e s [285]
I turn now to Thesis 2 and its unfortunate consequences. Iagree with
Dennett and Dawkins that it is often good for people to be enlightened, for
their false beliefs to be replaced by true ones. From my Science, Truth, and
Democracy (Kitcher 2001b) on, however, Ive been arguing that not every
truth is worth having. The aim of our epistemic ventures is significant truth.
Moreover the significant truths are those that contribute to human well-
being. Practical significance is the most obvious species: we value some
truths because they help doctors ameliorate diseases or protect vulnerable
people from threats to their security. Yet we shouldnt overlook the value
of truths that simply advance our understanding of the world. Dawkins is
often eloquent on the joy that comes from viewing nature clearly and accu-
rately. He rightly laments the blindness that prevents many people from
experiencing the uplift enlightenment brings.
Nevertheless once you make the shift from the value of truth (period)
to the value of significant truth, Thesis 2 becomes problematic. Because
significance depends on the conditions of human lives, the predicaments
different individuals and communities face, significant truth is thrown into
a broader mix of values, so that its benefits may be swamped by the losses
entailed in achieving it. No doubt if we lived in some rather different world,
it would be good to learn about the degrees to which genetic differences
affect the talents on which our current societies place most emphasis. But
if our only method for acquiring the knowledge would be to breed identical
twins, rearing them apart in controlled environments, we rightly forgo the
benefits of understanding the precise genetic contribution. And even if we
have alternative routes to knowledge, we might be obligated not to take
them if we were convinced that the likely damage from future uses of the
information would be toogreat.
My resistance to sweeping away all religion as noxious rubbish stems
from combining the pragmatist sense of truth as one value among others
(the point of the previous paragraph) with reflections on the lives Ihave
known. For anyone like me, for whom a youth in the lower reaches of the
British class system brought acquaintance with many people whose lives
were made bearable by their local church (liberal, by todays standards) and
the sense of community it provided, the losses of religious involvement can
easily outweigh the gains of enlightenment:Ihave known too many people
who have said, sincerely and accurately, Without my faith Ijust couldnt go
on. Their declaration stems, of course, from failures of the ambient societ-
ies. They are probably at least as common among those who live just north
of me in Harlem and who turn to one of the local churches or synagogues
or mosques for comfort, for support, for community, and sometimes for
a joint campaign for social justice. Ideally we could attend to the causes
W h at t o D o W h i l e R e l i g i o n s E volv e b e f or e O u r V e r y E y e s [287]
REFERENCES
AIMS. 2009. Memorandum from the Association for Improvements in the Maternity
Services (AIMS) to the House of Commons Select Committee for Children,
Schools and Families. http://www.aims.org.uk/Submissions/socialWorker
Training.htm.
Alon, Uri. 2007. An Introduction to Systems Biology:Design Principles of Biological Circuits.
Boca Raton, FL:Taylor and Francis.
Alston, William. 1958. Ontological Commitments. Philosophical Studies 9:817.
Amundson, Ron, and George Lauder. 1994. Function without Purpose. Biology and
Philosophy 9 (4):44369.
Aronson, Elliot, Timothy Wilson, and Robin Akert. 1999. Social Psychology. 3rd ed.
NewYork:Longman.
Bacon, Francis. (1620) 1960. The New Organon. Ed. Fulton H.Anderson. Upper Saddle
River, NJ:PrenticeHall.
Balaguer, Mark. 1998. Platonism and Anti-Platonism in Mathematics. Oxford: Oxford
UniversityPress.
Barlow, Allison, Elena Varipatis-Baker, Kristen Speakman, Golda Ginsburg, Ingrid
Friberg, Novalene Goklish, Brandii Cowboy, Pauline Fields, Ranelda Hastings,
William Pan, Raymond Reid, Mathuram Santosham, and John Walkup. 2006.
Home-Visiting Intervention to Improve Child Care among American Indian
Adolescent Mothers:ARandomized Trial. Archives of Pediatrics and Adolescent
Medicine 160 (11):11017.
Barnes, Barry, and John Dupr. 2008. Genomes and What to Make of Them.
Chicago:University of ChicagoPress.
Batterman, Robert. 2001. The Devil in the Details:Asymptotic Reasoning in Explanation,
Reduction and Emergence. Oxford:Oxford UniversityPress.
Becker, Jo. 2014. Forcing the Spring: Inside the Fight for Marriage Equality.
NewYork:Penguin.
Bellah, Robert. 2011. Religion in Human Evolution. Cambridge, MA: Harvard
UniversityPress.
Benacerraf, Paul. 1965. What Numbers Could Not Be. Philosophical Review 74
(1):4773.
. 1973. Mathematical Truth. Journal of Philosophy 70:66179.
Benacerraf, Paul, and Hilary Putnam. 1964. Philosophy of Mathematics: Selected
Readings. Englewood Cliffs, NJ:PrenticeHall.
Benson, Etienne. 2012. Endangered Science:The Regulation of Research by the U.S.
Marine Mammal Protection Act and Endangered Species Act, 19711978.
Historical Studies in the Natural Sciences 42:3061.
Berlin, Isaiah. (1950) 1979. Concepts and Categories. NewYork:Penguin.
Bigelow, John, and Robert Pargetter. 1987. Functions. Journal of Philosophy
54:18196.
Blair, Tony. 2006. Our Nations FutureSocial Exclusion. Our Nations Future lec-
ture series, September 5. http://webarchive.nationalarchives.gov.uk/200705
06094157/http://pm.gov.uk/output/Page10037.asp.
Boorse, Christopher. 1976. Wright on Functions. Philosophical Review 85:7086.
Boyd, Richard. 1991. Realism, Anti-Foundationalism, and the Enthusiasm for Natural
Kinds. Philosophical Studies 61:12748.
. 2010. Realism, Natural Kinds, and Philosophical Method. In The Semantics
and Metaphysics of Natural Kinds, ed. Helen Beebee and Nigel Sabbarton-Leary,
21234. London: Routledge.
Boyd, Robert, and Peter Richerson. 1985. Culture and the Evolutionary Process.
Chicago:University of ChicagoPress.
. 2005. Not by Genes Alone. Chicago: University of Chicago Press.
Brake, Elizabeth, and Joseph Millum. 2013. Parenthood and Procreation. In The
Stanford Encyclopedia of Philosophy (Winter 2013), ed. Edward N.Zalta. http://
plato.stanford.edu/archives/win2013/entries/parenthood.
Brandon, Robert N. 2013. A General Case for Functional Pluralism. In Functions:
Selection and Mechanisms, ed. Philippe Huneman, 97104. Dordrecht: Springer.
Brown, Mark B. 2013. Review of Philip Kitcher, Science in a Democratic Society. Minerva
51:38997.
Burgess, John P. 1983. Why IAm Not a Nominalist. Notre Dame Journal of Formal
Logic 24 (1):93105.
Burgess, John, and Gideon Rosen. 1997. A Subject with No Object. Oxford: Oxford
UniversityPress.
Bush, Vannevar. (1945) 1980. ScienceThe Endless Frontier. NewYork:ArnoPress.
Cahan, David. 1989. An Institute for Empire:The Physikalisch-Technische Reichsanstalt,
18711918. Cambridge, UK:Cambridge UniversityPress.
Calcott, Brett. 2014. The Creation and Reuse of Information in Gene Regulatory
Networks. Philosophy of Science 81 (5):87990.
Carnap, Rudolph. 1950. Logical Foundations of Probability. Chicago: University of
ChicagoPress.
Cartwright, Nancy. 1983. How the Laws of Physics Lie. Oxford:ClarendonPress.
. 1999. The Dappled World. Cambridge, UK:Cambridge UniversityPress.
Cartwright, Nancy, and Jeremy Hardie. 2012. Evidence-Based Policy:APractical Guide to
Doing It Better. NewYork:Oxford UniversityPress.
Chihara, Charles. 1973. Ontology and the Vicious-Circle Principle. Ithaca, NY: Cornell
UniversityPress.
Christensen, Wayne. 1996. A Complex Systems Theory of Teleology. Biology and
Philosophy 11 (3):30119.
Citizen Science Alliance. 2014. Why Citizen Science? http://www.citizenscience
alliance.org/philosophy.html. Accessed April 26,2014.
Cooper, Harris. 1989. Homework. White Plains, NY:Longman.
Craver, Carl F. 2007. Explaining the Brain. Oxford:ClarendonPress.
. 2014. The Ontic Account of Scientific Explanation. In Explanation in the
Special Sciences, ed. Marie I.Kaiser, Oliver R.Scholz, Daniel Plenge, and Andreas
Httemann, 2752. Dordrecht:Springer.
[290]References
Crick, Francis H.C. 1958. On Protein Synthesis. Symposia of the Society for Experimental
Biology 12:13863.
CSCB. 2013. Final Overview Report of Serious Case Review re Daniel Pelka. Coventry,
UK:Coventry Safeguarding ChildrenBoard.
Cummins, Robert. 1975. Functional Analysis. Journal of Philosophy 72:74165.
. 2002. Neo-Teleology. In Functions:New Essays in the Philosophy of Psychology
and Biology, ed. Andr Ariew, Robert Cummins, and Mark Perlman, 15773.
Oxford:Oxford UniversityPress.
Curry, Haskell B. 1951. Outlines of a Formalist Philosophy of Mathematics.
Amsterdam:North-Holland.
DAgostino, Fred. 2013. Review of Philip Kitcher. Science in a Democratic Society.
Analysis 73:59394.
Daniels, George. 1967. The Pure- Science Ideal and Democratic Culture. Science
156:1699705.
Darwin, Charles. (1859) 1964. On the Origin of Species. Cambridge, MA: Harvard
UniversityPress.
. 1862. On the Various Contrivances by Which British and Foreign Orchids Are
Fertilised by Insects, and on the Good Effects of Intercrossing. London:John Murray.
Daston, Lorraine. 2001. The Historicity of Science. In Historicization-Historisierung,
vol. 5 of Aporemata:Kritische Studien zur Philologiegeschichte, ed. Glenn W.Most,
20121. Gttingen:Vandenhoeck & Ruprecht.
. 2012. The Sciences of the Archives. Osiris 27:15687.
Davidson, Nicholas O. 2002. The Challenge of Target Sequence Specificity in CU
RNA Editing. Journal of Clinical Investigation 109 (3):29194.
Dawkins, Richard. 2006. The God Delusion. NewYork:Norton.
Dennett, Daniel C. 2006. Breaking the Spell: Religion as a Natural Phenomenon.
NewYork:Viking.
Dennett, Daniel C., and Linda LaScola. 2010. Preachers Who Are Not Believers.
Evolutionary Psychology 8 (1):12250.
. 2013. Caught in the Pulpit:Leaving Belief Behind. Amazon.com.
. 2015. Caught in the Pulpit: Leaving Belief Behind. Expanded ed. Durham,
NC:Pitchstone.
Dennett, Daniel C., and Deb Roy. 2015. Our Transparent Future. Scientific American
312 (3): 6469.
Detlefsen, Michael. 2005. Formalism. In The Oxford Handbook of Philosophy of
Mathematics and Logic, ed. Stewart Shapiro, 236317. Oxford: Oxford University
Press.
Dewey, John. (1909) 2007. The Influence of Darwin on Philosophy. In The Influence
of Darwin on Philosophy and Other Essays, ed. Larry Hickman, 312. Carbondale:
University of Southern Illinois.
. (1925) 1981. Experience and Nature. Vol. 1 of Later Works, ed. Jo Ann Boydston.
Carbondale:University of Southern Illinois.
. (1929) 1984. The Quest for Certainty. Vol. 4 of Later Works, ed. Jo Ann Boydston.
Carbondale:University of Southern Illinois.
Douglas, Heather. 2009. Science, Policy, and the Value-Free Ideal. Pittsburgh:University
of PittsburghPress.
Dowe, Phil. 2000. Physical Causation. Cambridge, UK:Cambridge UniversityPress.
Dretske, Fred. 1972. Contrastive Statements. Philosophical Review 81:41137.
References [291]
Dummett, Michael. 1991. Frege:Philosophy of Mathematics. Cambridge, MA:Harvard
UniversityPress.
Dupr, John. 1993. The Disorder of Things. Cambridge, MA:Harvard UniversityPress.
Edwards, Anthony W.F. 1998. Natural Selection and the Sex Ratio:Fishers Sources.
American Naturalist 151:56469.
Edwards, Paul N. 2010. A Vast Machine:Computer Models, Climate Data, and the Politics
of Global Warming. Cambridge, MA:MITPress.
Elster, Jon. 1984. Ulysses and the Sirens: Studies in Rationality and Irrationality.
Cambridge, UK:Cambridge UniversityPress.
Field, Hartry. 1980. Science without Numbers. Princeton, NJ:Princeton UniversityPress.
. 1989. Realism, Mathematics and Modality. Oxford: Oxford University Press.
Fisher, Ronald A. 1930. The Genetical Theory of Natural Selection. Oxford: Oxford
University Press.
Fishkin, James. 2009. When the People Speak. NewYork:Oxford UniversityPress.
Fletcher, Liz. 2001. New Zealand GMO Debacle Undermines Green Lobby. Nature
Biotechnology 19:292.
Fodor, Jerry A. 1974. Special Sciences. Synthese 28:97115.
Forrest, Peter, and David M.Armstrong. 1987. The Nature of Number. Philosophical
Papers 16:16586.
Fox, Charles W., Monica S. Thakar, and Timothy A. Mousseau. 1997. Egg Size
Plasticity in a Seed Beetle: An Adaptive Maternal Effect. American Naturalist
149 (1):14963.
Franklin-Hall, Laura R. Forthcoming. The Causal Economy Account of Scientific
Explanation. In Philosophical Perspectives on Causal Reasoning in Biology, ed.
C. K. Waters and J. Woodward, vol. 21 of Minnesota Studies in the Philosophy of
Science. Minneapolis: University of Minnesota Press.
. 2015. Explaining Causal Selection with Explanatory Causal Economy:
Biology and Beyond. In Explanation in Biology: An Enquiry into the
Diversity of Explanatory Patterns in the Life Sciences, ed. P. A. Braillard and
C. Malaterre, 413438. Dordrecht: Springer.
Frege, Gottlob. (1903) 2013. Basic Laws of Arithmetic. Ed. Philip A. Ebert and Marcus
Rossberg. Oxford: Oxford University Press.
. (1918) 1984. Der Gedanke: Eine Logische Untersuchung. In Beitrge zur
Philosophie des deutschen Idealismus 1 (191819):5877. Translated as Thoughts,
in Collected Papers on Logic, Mathematics and Philosophy. Trans. Max Black,
V. H.Dudman, Peter Geach, Hans Kaal, E. H.W. Kluge, Brian McGuinness, and
R. H.Stoothoff. Ed. B. McGuinness. Oxford:Basil Blackwell.
French, Steven. 2012. The Resilience of Laws and the Ephemerality of Objects. In
Probabilities, Laws, and Structures, ed. Dennis Dieks, Wenceslao Gonzalez,
Stephan Hartmann, Michael Stoltzner, and Marcel Weber, 187200. Dordrecht:
Springer Science and BusinessMedia.
Friedman, Michael. 1974. Explanation and Scientific Understanding. Journal of
Philosophy 71:519.
Gibbard, Allan. 1990. Wise Choices, Apt Feelings. Cambridge, MA:Harvard University
Press.
Giere, Ronald. 2006. Scientific Perspectivism. Chicago:University of ChicagoPress.
. 2013. Kuhn as Perspectival Realist. Topoi 32:5357.
Gdel, Kurt. 1964. What Is Cantors Continuum Problem? In Philosophy of
Mathematics:Selected Readings, ed. Paul Benacerraf and Hilary Putnam, 25873.
Englewood Cliffs, NJ:PrenticeHall.
[292]References
Godfrey-Smith, Peter. 1993. Functions:Consensus without Unity. Pacific Philosophical
Quarterly 74:196208.
. 1994. A Modern History Theory of Functions. Nous 28:34462.
Goffman, Erving. 1959. The Presentation of Self in Everyday Life. NewYork:AnchorBooks.
Gonzalez, Wenceslao J., ed. 2011. Scientific Realism and Democratic Society: The
Philosophy of Philip Kitcher. Amsterdam:Rodopi.
Goodstein, Laurie. 2006. Evangelicals Fear the Loss of Their Teenagers. New York
Times, October6.
Gould, Stephen Jay, and Richard C. Lewontin. 1979. The Spandrels of San Marco
and the Panglossian Paradigm: A Critique of the Adaptationist Programme.
Proceedings of the Royal Society B 205:58198.
Greenberg, Daniel S. 2001. Science, Money, and Politics: Political Triumph and Ethical
Erosion. Chicago:University of ChicagoPress.
Griffiths, Paul E. 2006a. The Fearless Vampire Conservator: Philip Kitcher, Genetic
Determinism and the Informational Gene. In Genes in Development:Re-reading
the Molecular Paradigm, ed. Christoph Rehmann-Sutter and Eva M.Neumann-
Held, 17598. Durham, NC:Duke UniversityPress.
. 2006b. Function, Homology and Character Individuation. Philosophy of
Science 73 (1):125.
. 2009. In What Sense Does Nothing Make Sense Except in the Light of
Evolution? Acta Biotheoretica 57 (12):1132.
. 2013. Lehrmans Dictum: Information and Explanation in Developmental
Biology. Developmental Psychobiology 55 (1):2232.
Griffiths, Paul E., and Russell D.Gray. 1994. Developmental Systems and Evolutionary
Explanation. Journal of Philosophy 91 (6):277304.
Griffiths, Paul E., Edouard Machery, and Stefan Linquist. 2009. The Vernacular
Concept of Innateness. Mind and Language 24 (5):60530.
Griffiths, Paul E., Arnaud Pocheville, Brett Calcott, Karola Stotz, Hyunju Kim, and Rob
D.Knight. 2015. Measuring Causal Specificity. Philosophy of Science 82:52955.
Griffiths, Paul E., and Karola Stotz. 2013. Genetics and Philosophy: An Introduction.
NewYork:Cambridge UniversityPress.
Griffiths, Paul E., and James G.Tabery. 2013. Developmental Systems Theory:What
Does It Explain, and How Does It Explain It? Advances in Child Development and
Behavior 44:6594.
Haack, Susan. 1993. Evidence and Inquiry: Towards Reconstruction in Epistemology.
Oxford:Blackwell.
Hacking, Ian. 1999. The Social Construction of What? Cambridge, MA: Harvard
UniversityPress.
Hahn, Roger. 1971. The Anatomy of an Institution:The Paris Academy of Sciences, 1666
1803. Berkeley:University of CaliforniaPress.
Hamer, Dean. 2002. Rethinking Behavior Genetics. Science 298:7172.
Hamer, Dean H., S. Hu, V. L.Magnuson, N. Hu, and A. M.L. Pattatucci. 1993. A Linkage
between DNA Markers on the X Chromosome and Male Sexual Orientation.
Science Communication 261:32026.
Hare, Robert. 1991. The Hare Psychopathy ChecklistRevised. Toronto: Multi-Health
Systems.
. 1998. The Hare PCL-R:Some Issues concerning Its Use and Misuse. Legal and
Criminological Psychology 3 (1):99119.
Harris, Sam. 2004. The End of Faith: Religion, Terrorism and the Future of Reason.
NewYork:Norton.
References [293]
. 2006. Letter to a Christian Nation. NewYork:Knopf.
Hecht, Gabrielle. 2012. Being Nuclear:Africans and the Global Uranium Trade. Cambridge,
MA:MITPress.
Hempel, Carl G., and Paul Oppenheim. 1948. Studies in the Logic of Explanation.
Philosophy of Science 15:13575.
Herrnstein, Richard J., and Charles Murray. 1994. The Bell Curve:Intelligence and Class
Structure in American Life. NewYork:Simon and Schuster.
Hitchens, Christopher. 2007. God Is Not Great: How Religion Poisons Everything.
NewYork:Twelve.
Holling, Crawford S. 1959. The Components of Predation as Revealed by a Study of
Small-Mammal Predation of the European Pine Sawfly. Canadian Entomologist
91:293320.
Huneman, Philippe. 2013. Introduction to Functions: Selection and Mechanisms.
Dordrecht:Springer.
James, William. (1902) 1982. The Varieties of Religious Experience. Harmondsworth,
UK:Penguin.
. (1907) 1975. Pragmatism. Cambridge, MA:Harvard UniversityPress.
Jasanoff, Sheila. 2012. Science and Public Reason. London:Routledge.
Jewett, Andrew. 2012. Science, Democracy, and the American University:From the Civil
War to the Cold War. Cambridge, UK:Cambridge UniversityPress.
Kant, Immanuel. (178187) 1997. Critique of Pure Reason. Cambridge, UK:Cambridge
UniversityPress.
Kepler, Johannes. 1615. Nova stereometria doliorum vinariorum. Linz, Austria:
J. Plancus.
Kim, Jaegwon. 1988 [2010]. Explanatory Realism, Causal Realism, and Explanatory
Exclusion. In vol. 12 of Midwest Studies in Philosophy, 22539. Reprinted in
Kim, J. (2010). Essays in the Metaphysics of Mind, 14866. Oxford: Oxford
UniversityPress.
. 1994. Explanatory Knowledge and Metaphysical Dependence. Philosophical
Issues 5:5169.
Kincaid, Harold, John Dupr, and Alison Wylie, eds. 2007. Value-Free Science: Ideals and
Illusions. New York: Oxford University Press.
Kitcher, Philip. 1973. Fluxions, Limits, and Infinite Littlenesse. Isis 64:3349.
. 1975. Bolzanos Ideal of Algebraic Analysis. Studies in the History and
Philosophy of Science 6:22971.
. 1980. Arithmetic for the Millian. Philosophical Studies 37:21536.
. 1981. Explanatory Unification. Philosophy of Science 48:50731.
. 1982. Abusing Science:The Case against Creationism. Cambridge, MA:MITPress.
. 1983. The Nature of Mathematical Knowledge. NewYork:Oxford UniversityPress.
. 1984. 1953 and All That: A Tale of Two Sciences. Philosophical Review 93:
33573.
. 1985. Vaulting Ambition: Sociobiology and the Quest for Human Nature.
Cambridge, MA:MITPress.
. 1986. Projecting the Order of Nature. In Kants Philosophy of Physical Science,
ed. Robert Butts, 20135. University of Western Ontario Series in Philosophy of
Science. Dordrecht:D. Reidel.
. 1989. Explanatory Unification and the Causal Structure of the World. In
Scientific Explanation, ed. Philip Kitcher and Wesley Salmon, 410505, vol. 13
of Minnesota Studies in the Philosophy of Science. Minneapolis: University of
MinnesotaPress.
[294]References
. 1990. The Division of Cognitive Labor. Journal of Philosophy 87:521.
. 1993a. The Advancement of Science. NewYork:Oxford UniversityPress.
. 1993b. The Evolution of Human Altruism. Journal of Philosophy 90:497516.
. 1993c. Function and Design. Midwest Studies in Philosophy 18:37997.
. 1996. The Lives to Come: The Genetic Revolution and Human Possibilities.
NewYork:Simon and Schuster.
. 1998. A Plea for Science Studies. In A House Built on Sand: Exposing
Postmodernist Myths about Science, ed. Noretta Koertge, 3256. NewYork:Oxford
UniversityPress.
. 1999. The Hegemony of Molecular Biology. Biology and Philosophy 14:195210.
. 2001a. Real Realism:The Galilean Strategy. Philosophical Review 110:15197.
. 2001b. Science, Truth, and Democracy. NewYork:Oxford UniversityPress.
. 2003a. Battling the Undead: How (and How Not) to Resist Genetic
Determinism. In In Mendels Mirror: Philosophical Reflections on Biology, 283
300. Oxford:Oxford UniversityPress.
. 2003b. Developmental Decomposition and the Future of Human Behavioral
Ecology. In In Mendels Mirror: Philosophical Reflections on Biology, 30120.
Oxford:Oxford UniversityPress.
. 2007a. Joyces Kaleidoscope: An Invitation to Finnegans Wake. Oxford: Oxford
UniversityPress.
. 2007b. Living with Darwin: Evolution, Design, and the Future of Faith.
Oxford:Oxford UniversityPress.
. 2011a. The Ethical Project. Cambridge, MA:Harvard UniversityPress.
. 2011b. Science in a Democratic Society. Amherst, NY:PrometheusBooks.
. 2011c. Militant Modern Atheism. Journal of Applied Philosophy 28:113.
. 2012a. Afterthoughts:Reply to Comments. Analyse & Kritik 1:16789.
. 2012b. Mathematical Truth? In Preludes to Pragmatism, 166 91. New
York:Oxford UniversityPress.
. 2012c. Preludes to Pragmatism. NewYork:Oxford UniversityPress.
. 2013. Deaths in Venice:The Cases of Gustav von Aschenbach. NewYork:Columbia
UniversityPress.
. 2014. Life after Faith. New Haven, CT:Yale UniversityPress.
Kitcher, Philip, and Wesley Salmon. 1987. Van Fraassen on Explanation. Journal of
Philosophy 84: 31530.
Kohn, Linda, Janet Corrigan, and Molla Donaldson, eds. 2000. To Err Is Human:Building
a Safer Health System. Washington, DC:National AcademyPress.
Kuhn, Thomas. 1962. The Structure of Scientific Revolutions. Chicago: University of
ChicagoPress.
. 1990. Dubbing and Re-dubbing: The Vulnerability of Rigid Designation. In
Scientific Theories, ed. C. Wade Savage, 298318, vol. 14 of Minnesota Studies in
the Philosophy of Science. Minneapolis:University of MinnesotaPress.
. 2000. The Road Since Structure. Ed. James Conant and John Haugeland.
Chicago:University of ChicagoPress.
Kusch, Martin. 2002. Knowledge by Agreement. Oxford:Oxford UniversityPress.
Lange, Marc. Forthcoming. Because without Cause. NewYork:Oxford UniversityPress.
Laudan, Larry. 1981. A Confutation of Convergent Realism. Philosophy of Science
48:1949.
. 1984. Science and Values. Berkeley:University of CaliforniaPress.
Lewis, David. 1983. New Work for a Theory of Universals. Australasian Journal of
Philosophy 61:34377.
References [295]
. 2000. Causation as Influence. Journal of Philosophy 97:18297.
Lewontin, Richard C. 1983. The Organism as the Subject and Object of Evolution.
Scientia 118:6582.
Linquist, Stefan, Edouard Machery, Paul E.Griffiths, and Karola Stotz. 2011. Exploring
the Folkbiological Conception of Human Nature. Philosophical Transactions of
the Royal Society B 366:44453.
Lotka, Alfred J. 1956. Elements of Mathematical Biology. 2nd ed. New York: Dover.
Originally published as Elements of Physical Biology in1925.
MacFarlane, John. 2005. Making Sense of Relative Truth. Proceedings of the
Aristotelian Society 105:32139.
. 2009. Nonindexical Contextualism. Synthese 166:23150.
Maddy, Penelope. 1997. Naturalism in Mathematics. Oxford:Oxford UniversityPress.
Mantzavinos, Chrysostomos. 2012. The Ethical Project:ADialogue. Analyse & Kritik
1:2138.
Marques, Teresa. 2014. Relative Correctness. Philosophical Studies 167:36173.
Massimi, Michela. 2010. Galileos Mathematization of Nature at the Crossroad
between the Empiricist Tradition and the Kantian One. Perspectives on Science
18:15288.
. 2012. Scientific Perspectivism and Its Foes. Philosophica 84:2552.
. 2014. Natural Kinds and Naturalised Kantianism. Nos 48:41649.
. 2015a. Walking the Line:Kuhn between Realism and Relativism. In Kuhns
Structure of Scientific Revolutions: 50 Years On, ed. Alisa Bokulich and William
Devlin. Boston Studies in the Philosophy of Science. Springer. DOI: 10.1007/
978-3-319-13383-6_9.
. 2015b. Working in a New World: Kuhn, Constructivism and Mind-
Dependence. Studies in History and Philosophy of Science 50,8389.
Massimi, Michela, and John Peacock. 2014. The Origins of Our Universe: Laws,
Testability and Observability in Cosmology. In Philosophy and the Sciences for
Everyone, ed. Michela Massimi, 1432. London:Routledge.
Matan, Adee, and Susan Carey. 2001. Developmental Changes within the Core of
Artifact Concepts. Cognition 78 (1):126.
Maudlin, Tim. 1996. On the Unification of Physics. Journal of Philosophy 93:12944.
Maynard Smith, John. 2000. The Concept of Information in Biology. Philosophy of
Science 67 (2):17794.
Mayr, Ernst. 1961. Cause and Effect in Biology. Science 134 (3489):15016.
McLaughlin, Peter. 2001. What Functions Explain: Functional Explanation and Self-
Reproducing Systems. Cambridge, UK:Cambridge UniversityPress.
Mill, John Stuart. (1859) 1975. On Liberty. Ed. David Spitz. NewYork:Norton.
Millikan, Ruth Garrett. 1984. Language, Thought and Other Biological Categories.
Cambridge, MA:MITPress.
Moe, Kristine. 1984. Should the Nazi Research Data Be Cited? Hastings Center Report
14:57.
Morrison, Margaret. 2000. Unifying Scientific Theories: Physical Concepts and
Mathematical Structures. Cambridge, UK:Cambridge UniversityPress.
Munro, Eileen. 2005. Improving Practice: Child Protection as a Systems Problem.
Children and Youth Services Review 27:37591.
. 2011. The Munro Review of Child Protection: Final Report. London: UK
Department of Education.
Nagel, Ernest. 1979. The Structure of Science. Indianapolis:Hackett.
Nagel, Thomas. 1970. The Possibility of Altruism. Princeton, NJ: Princeton
UniversityPress.
[296]References
Nanney, D.L. 1958. Epigenetic Control Systems. Proceedings of the National Academy
of Sciences 44 (7):71217.
Neander, Karen. 1991. Functions as Selected Effects. Philosophy of Science 58:16984.
. 1995. Pruning the Tree of Life. British Journal for the Philosophy of Science 46
(1):5980.
. 2002. Types of Traits: The Importance of Functional Homologues. In
Functions: New Essays in the Philosophy of Biology and Psychology, ed. Andr
Ariew, Robert Cummins, and Mark Perlman, 390 415. New York: Oxford
UniversityPress.
. 2015. Functional Analysis and the Species Design. Synthese. DOI: 10.1007/s
11229-015-0940-9.
Neander, Karen, and Alex Rosenberg. 2012. Solving the Circularity Problem for
Functions:AResponse to Nanay. Journal of Philosophy 109 (10):61322.
Okasha, Samir. 2009. Causation in Biology. In The Oxford Handbook of Causation,
ed. Helen Beebee, Peter Menzies, and Christopher Hitchcock, 707 25.
NewYork:Oxford UniversityPress.
Olby, Robert Cecil. 2009. Francis Crick:Hunter of Lifes Secrets. NewYork:Cold Spring
Harbor LaboratoryPress.
Oppenheim, Paul, and Hilary Putnam. 1958. Unity of Science as a Working
Hypothesis. In Concepts, Theories, and the Mind-Body Problem, ed. H. Feigl,
M. Scriven, and G. Maxwell, 336, vol. 2 of Minnesota Studies in the Philosophy of
Science. Minneapolis:University of MinnesotaPress.
Oyama, Susan. 2000a. Causal Democracy and Causal Contributions in Developmental
Systems Theory. Philosophy of Science 67 (s1):S332S347.
. 2000b. The Ontogeny of Information: Developmental Systems and Evolution.
2nded., revised and expanded. Durham, NC:Duke UniversityPress.
Parfit, Derek. 1984. Reasons and Persons. Oxford:Oxford UniversityPress.
Parker, Andrew. 2003. In the Blink of an Eye:How Vision Sparked the Big Bang of Evolution.
Cambridge, MA:Perseus.
Pearl, Judea. 2009. Causality: Models, Reasoning, and Inference. 2nd ed. New York:
Cambridge UniversityPress.
Pemberton, Camilla. 2013. Eileen Munro on Daniel Pelka Case:I Cant Claim IWould
Have Done Better. The Children's Services Blog. Community Care. http://www.
communitycare.co.uk/blogs/childrens- services-blog/2013/09/eileen-munro-
admits-i-cant-say-i-would-have-done-better-in-protecting-daniel-pelka/.
Pielke, Roger, Jr. 2012. Basic Research as a Political Symbol. Minerva 50:33961.
Potochnik, Angela. 2010. Explanatory Independence and Epistemic
Interdependence:ACase Study of the Optimality Approach. British Journal for
the Philosophy of Science 61:21333.
Psillos, Stathis. 2000. The Present State of the Scientific Realism Debate. British
Journal for the Philosophy of Science 51:70528.
Putnam, Hilary. 1971. Philosophy of Logic. NewYork:Harper andRow.
. 1975. What Is Realism? Proceedings of the Aristotelian Society 76:17794.
Quine, W.V. 1961. Word and Object. Cambridge, MA:MITPress.
. 1969. Natural Kinds. In Ontological Relativity and Other Essays.
NewYork:Columbia UniversityPress.
Raddick, M.Jordan, Georgia Bracey, Pamela L.Gay, Chris J.Lintott, Carie Cardamone,
Phil Murray, Kevin Schawinski, Alexander S.Szalay, and Jan Vandenberg. 2013.
Galaxy Zoo: Motivations of Citizen Scientists. Cornell University Library.
http://arxiv.org/abs/1303.6886. Accessed April 26,2014.
Reichenbach, Hans. 1956. The Direction of Time. Berkeley:University of CaliforniaPress.
References [297]
Resnik, Michael. 1980. Frege and the Philosophy of Mathematics. New York: Cornell
UniversityPress.
Rosen, Gideon. 1999. Review of Maddy (1997). British Journal for the Philosophy of
Science 50:46774.
. 2014. Abstract Objects. In The Stanford Encyclopedia of Philosophy (Fall 2014),
ed. Edward N. Zalta. http://plato.stanford.edu/archives/fall2014/entries/
abstract-objects/.
Rosen, Gideon, and John P.Burgess. 2005. Nominalism Reconsidered. In The Oxford
Handbook of Philosophy of Mathematics and Logic, ed. Stewart Shapiro, 51535.
Oxford:Oxford UniversityPress.
Rosenberg, Alexander, and Daniel W. McShea. 2008. Philosophy of Biology: A
Contemporary Introduction. New York: Routledge.
Rosenberg, Alex, and Karen Neander. 2009. Are Homologies (Selected Effect or Causal
Role) Function Free? Philosophy of Science 76 (3):30734.
Rosenberg, Jay. 2002. Thinking about Knowing. Oxford:Oxford UniversityPress.
Rossiter, Margaret. 1975. The Emergence of Agricultural Science: Justus Liebig and the
Americans, 18401880. New Haven, CT:Yale UniversityPress.
Roughgarden, Joan. 1979. Theory of Population Genetics and Evolutionary Ecology.
NewYork:Macmillan.
Ruderman, Wendy. 2013. To Stem Juvenile Robberies, Police Trail Youths before
the Crime. New York Times, March 3. http://www.nytimes.com/2013/03/04/
nyregion/to- stem- juvenile-robberies-police-trail-youths-before-the-crime.
html?smid=pl-share.
Salmon, Wesley. 1984. Scientific Explanation and the Causal Structure of the World.
Princeton, NJ:Princeton UniversityPress.
. 1990. Scientific Explanation:Causation and Unification. Critica 22:323.
Sarkar, Sahotra. 2004. Genes Encode Information for Phenotypic Traits. In
Contemporary Debates in Philosophy of Science, ed. Christopher Hitchcock, 25974.
Malden, MA:Blackwell.
Schlosser, Gerhard. 1998. Self Re-production and Functionality:ASystems-Theoretical
Approach to Teleological Explanation. Synthese 116:30354.
Secord, James. 2014. Visions of Science:Books and Readers at the Dawn of the Victorian
Age. Oxford:Oxford UniversityPress.
Shadwell, Thomas. (1676) 1966. The Virtuoso. Ed. Marjorie Hope Nicolson and David
Stuart Rodes. London:Edward Arnold.
Shea, Nicholas. 2007. Consumers Need Information:Supplementing Teleosemantics
with an Input Condition. Philosophy and Phenomenological Research 75 (2):
40435.
. 2011. Developmental Systems Theory Formulated as a Claim about Inherited
Information. Philosophy of Science 78:6082.
. 2013. Inherited Representations Are Read in Development. British Journal for
the Philosophy of Science 64 (1):131.
Shiffrin, Seana V. 2000. Paternalism, Unconscionability Doctrine, and
Accommodation. Philosophy and Public Affairs 29 (3):20550.
Shoesmith, Sharon. 2011. Shoesmith:I Don't Do Blame. Interview. BBC Radio 4, May
28. http://news.bbc.co.uk/today/hi/today/newsid_9499000/9499424.stm.
Singh, Rama, Costas Krimbas, Diane Paul, and John Beatty, eds. 2001. Thinking
about Evolution: Historical, Philosophical, and Political Perspectives. Cambridge,
UK:Cambridge UniversityPress.
[298]References
Sjberg, Svein, and Camilla Schreiner. 2010. The ROSE Project:An Overview and Key
Findings. http://roseproject.no/network/countries/norway/eng/nor-Sjoberg-
Schreiner-overview-2010.pdf. Accessed April 26,2014.
Skyrms, Brian. 2010. Signals:Evolution, Learning, and Information. NewYork:Oxford
UniversityPress.
Smith, Adam. (1759) 1976. The Theory of Moral Sentiments. Ed. D. D. Raphael and
A. L.Macfie. Oxford:ClarendonPress.
Smith, Michael. 1994. The Moral Problem. Oxford:Wiley-Blackwell.
. 2011. Deontological Moral Obligations and Non-Welfarist Agent-Relative
Values. Ratio 24:35163.
. 2012. Agents and Patients, Or: What We Learn about Reasons for Action
by Reflecting on Our Choices in Process-of-Thought Cases. Proceedings of the
Aristotelian Society 122:30931.
. 2013. A Constitutivist Theory of Reasons:Its Promise and Parts. LEAP:Law,
Ethics, and Philosophy 1:930.
Smith, Pamela H. 1994. The Business of Alchemy:Science and Culture in the Holy Roman
Empire. Princeton, NJ:Princeton UniversityPress.
Sober, Elliott. 1984. The Nature of Selection:Evolutionary Theory in Philosophical Focus.
Cambridge, MA:MITPress.
. 1995. Natural Selection and Distributive Explanation:AReply to Neander.
British Journal for the Philosophy of Science 46 (3):38497.
Sosa, Ernest. 1991. Knowledge in Perspective:Selected Essays in Epistemology. Cambridge,
UK:Cambridge UniversityPress.
Spencer, Michael. 2009. The Coming Evangelical Collapse. Christian Science Monitor,
March10.
Srinivasan, Amia. 2012. In the Long Cool Hour. London Review of Books 34 (23):1718.
Stegmann, Ulrich. 2014. Causal Control and Genetic Causation. Nos 48 (3):45065.
Steiner, Mark. 1978. Mathematical Explanation. Philosophical Studies 34:13551.
Sterelny, Kim. 2012. Moralitys Dark Past. Analyse & Kritik 34: 95115.
Sterelny, Kim, Michael Dickison, and Kelly Smith. 1996. The Extended Replicator.
Biology and Philosophy 11 (3):377403.
Sterelny, Kim, and Paul E. Griffiths. 1999. Sex and Death: An Introduction to the
Philosophy of Biology. Chicago:University of ChicagoPress.
Sterelny, Kim, and Philip Kitcher. 1988. The Return of the Gene. Journal of Philosophy
85:33961.
Stotz, Karola. 2006. Molecular Epigenesis:Distributed Specificity as a Break in the
Central Dogma. History and Philosophy of the Life Sciences 28 (4):52744.
Street, Sharon. 2006. A Darwinian Dilemma for Realist Theories of Value. Philosophical
Studies 127:10966.
Strevens, Michael. 2008. Depth: An Account of Scientific Explanation. Cambridge,
MA:Harvard UniversityPress.
Swade, Doron. 1991. Charles Babbage and His Calculating Engines. London: Science
Museum.
Thomae, J. 1898. Elementare Theorie der analytischen Functionen einer complexen
Vernderlichen. 2nd ed. Halle, Germany:L. Nebert.
Thornhill, Randy. 2007. The Importance of Developmental Biology to Evolutionary
Biology and Vice Versa. In The Evolution of Mind: Fundamental Questions and
Controversies, ed. Steven W. Gangestad and Jeffry A. Simpson. New York:
Guilford.
References [299]
Tinbergen, Nikolaas. 1963. On the Aims and Methods of Ethology. Zietschrift Fr
Tierpsychologie 20:41033.
Tolan, Jeffrey S., and R. K. Finn. 1987. Fermentation of D-Xylose to Ethanol by
Genetically Modified Klebsiella planticola. Applied Environmental Microbiology
53:203944.
Tomasello, Michael. 2009. Why We Cooperate. Cambridge, MA:MITPress.
. 2016. A Natural History of Human Morality. Cambridge, MA: Harvard University
Press.
UK Department of Health. 2000. An Organisation with a Memory:Report of an Expert
Group on Learning from Adverse Events in the NHS. London:Stationery Office.
Valleriani, Matteo. 2010. Galileo Engineer. Dordrecht:Springer.
Van Fraassen, Bas. 1980. The Scientific Image. Oxford:ClarendonPress.
Voit, Eberhard. 2013. A First Course in Systems Biology. NewYork and London:Taylor
and Francis.
Wakefield, Andrew J., S. H. Murch, A. Anthony, J. Linnell, D. M. Casson, M. Malik,
M. Berelowitz, A. P. Dhillon, M. A. Thomson, P. Harvey, A. Valentine,
S. E.Davies, and J. A.Walker-Smith. 1998. Ileal-Lymphoid-Nodular Hyperplasia,
Non-specific Colitis, and Pervasive Developmental Disorder in Children. Lancet
351:63741. (Retracted in Lancet 2010, 375:445.)
Wallace, R.J. 1994. Responsibility and the Moral Sentiments. Cambridge, MA:Harvard
UniversityPress.
Wang, Z., and C. B.Burge. 2008. Splicing Regulation:From a Parts List of Regulatory
Elements to an Integrated Splicing Code. RNA 14 (5):80213.
Waters, Kenneth C. 2007. Causes That Make a Difference. Journal of Philosophy 104
(11):55179.
Weber, Marcel. 2005. Philosophy of Experimental Biology. Cambridge, UK: Cambridge
UniversityPress.
. 2006. The Central Dogma as a Thesis of Causal Specificity. History and
Philosophy of the Life Sciences 28 (4):595609.
Williams, Garrath. 2003. Blame and Responsibility. Ethical Theory and Moral Practice
6:42745.
Wilson, Mark. Forthcoming. Physics Avoidance and Other Essays. Unpublished
manuscript.
Wittgenstein, Ludwig. 1953. Philosophical Investigations. Ed. G. E. M. Anscombe and
R. Rhees. Trans. G. E.M. Anscombe. Oxford:Basil Blackwell.
Woodward, James. 2003. Making Things Happen. NewYork:Oxford UniversityPress.
. 2010. Causation in Biology:Stability, Specificity, and the Choice of Levels of
Explanation. Biology and Philosophy 25 (3):287318.
. 2011. Causes, Conditions, and the Pragmatics of Causal Explanation.
In Philosophy of Science Matters: The Philosophy of Peter Achinstein, 24757.
NewYork:Oxford UniversityPress.
Woolf, Harry. 1969. The Transits of Venus: A Study of Eighteenth-Century Science.
Princeton, NJ:Princeton UniversityPress.
Wouters, Arno. 2007. Design Explanation:Determining the Constraints on What Can
Be Alive. Erkenntnis 67 (1):6580.
Wright, Crispin. 1983. Freges Conception of Numbers as Objects. Scotts Philosophical
Monographs. Aberdeen:Aberdeen UniversityPress.
Wright, Larry. 1973. Functions. Philosophical Review 82:13968.
. 1976. Teleological Explanations. Berkeley:University of CaliforniaPress.
[300]References
INDEX
Abraham,27879 functionsof,52
absolute completeness,16869 precisification of,6667
Abusing Science:The Case against simplicity of,62n18
Creationism (Kitcher),1,2 atheists, 273. See also New Atheists
accuracy,11011 autonomy
action potential, explanation of,13840 antireductionism and,17275
Adams, John Couch,41 and division of scientific
adaptation, 4748, 8789,90 labor,16669
adaptationism, 68,70,97 explanatory, 910, 159,164
adaptive information,89,91 explanatory reductionism and,160
adaptiveness, 8788,90 genetics and,148
adaptive phenotypic plasticity,76 reductionism and, 15354, 160,168
Advancement of Science, The (Kitcher),3 science and, 11,21115
Albon, S.D.,69 special sciences and, 9, 121, 123,154
alleles unificationand,8
assorting of,17980 in upper-level relationships,12930
phenotypes and,7980 Avicenna (Ibn Sina),102
Alon, Uri,14445
alterations, adaptive and Babbage, Charles,215
maladaptive,48 Bacon, Francis,215
alternatives, selection and,60 Balls, Ed,230
altruism,3,191 Bateson, Patrick,95,96
failures in, 25460, 26669,271 Batterman, Robert,141
animals Bauplan,62,66
domestication of,256 behavior
ethical relations with,260 genetic bases for, 76,93,95
anthropocentrism, 139,141 universal, 14143,145
antinominalism,21n7 behavioral development, folk
antireductionism, arguments for,17275 biologyof,77
Arbuthnot, John, 150, 155,179 belief, as value,284
Archimedes, 103,104 belief model, for religion, 275,
Argand plane,22 27678,282
argument patterns, 122, 125, 131, 132, Bell Curve, The (Herrnstein and
135, 148,157 Murray),186
Aristotle, 1023, 1057, 11516,118 Bellah, Robert, 284,285
artifacts Benacerraf, Paul, 6, 26, 29,39,40
categorizing and labeling,59 Berlin, Isaiah,210
bijective mapping,81 Carnap, Rudolf,285
biological classification,124 Carnot, Nicolas Lonard Sadi,107
biological development, inherited Cartwright, Nancy, 1112, 148, 149, 150,
information and,86 176, 245,24852
biological information categorical imperative,271
as barrier,74 causal cakes, 23537,244
causal,78 causal democracy, 78, 7475,92,97
causal specificity and,8285 applications for,93
intentional,78 biomedical research and,9596
proximate vs ultimate,7779 genetic information and,7677
proximate,7985 intentionof,92
ultimate,8591 principle of, 78,85
biological teleology,8,87 recognizing variety of difference-
biology making causes,94
antireductionist theses about, sourceof,96
defense of,146 specificity and,8385
causal relationshipsin,82 causal explanation
causal selectionin,81 difference-making,80
creationist,5556 interventionist viewof,78
developmental,84 theoryof,80
evolution and,68 causal graph theory,80
interactionist consensusin,74 causal information,7778
philosophy of,68,74 casuality, assumptions about, 12, 234,
biomedical research,9395 238,247
black-boxing, 10, 16568,177 causal pie diagrams,235
efficiency of,171 causal relations, 80,24647
irrelevance and,17072 causal selection,81
Blair, Tony, 23031,242 causal specificity, 78, 7983,90
Blake, William,285 causal webs, 16163, 168,234
blame, 23132, 243,250 causation
Bombelli, Raphael, 20, 21, 28, 30,38,40 adaptive information and,89
Boorse, Christopher,5152 explanation and,123
Boyd, Richard, 112,98 interventionist criterion of,81,82
Boyd, Robert,9697 justification and,271
Boyle, Robert, 214,224 linear models of,245
Brake, Elizabeth,231 linearity of,247
Breaking the Spell (Dennett), 276,280 local principles of, 23849, 242,244
bridge principles, 36, 153,159n4 philosophyof,80
Britten, Benjamin,205 unification and,151
Broad Street cholera epidemic certainty, 234,246
(1854),16162 channel conditions,80
Brothers Karamazov, The chemicals, behavior of,144
(Dostoyevsky),285 child abuse, 22930, 24245, 250,251
Buber, Martin,285 cholera epidemic,16162
Buridan, John, 102, 103, 104, 116,118 citizen juries, 10,205
Bush, Vannevar,21415 citizen science,223
Clausius, Rudolf,107n5
Cantor, Georg,21 clergy, nonbelieving, 275, 285,287
capitalism,204 climate change, 194, 195, 197,
Cardano, Rafael,20 19899,2034
[302]Index
Clutton-Brock, T.H.,69 democracy,18889
cognitive labor, division of, 910, 166, degeneration of,2034
171, 190. See also epistemic labor, expertise and, 188,190
divisionof history of,213
cognitive relations, discoveries and,39 ideal conversations and, 195, 196,
collective good,221 198,20910
combinatorial claims,34 neoliberal framework for, 199,205
compartmentalization, 910,16671 progress and,22122
complex arithmetic, development project of,208
of,2125 promoting human
complex numbers flourishing,21821
detectionof,25 science and, 1011, 190, 19192,
discoveryof,24 19499, 201, 2023, 206, 20912,
conception, unity of,5253 218,22122
conceptual incommensurability,119 self-image of,222
conjunction, principle of,157 social pathology in,2035
Connelly, Peter, 22930, 241,250 as social technology, 10, 191,196
consistency, 34,11011 vulgar, 188, 195, 197,211
constructivism, 99, 100, 114,119 as work in progress,203
contextual completeness,16869 democratic socialism,200
contextual irrelevance, 16465, De Motu antiquiora (Galileo),103
167,16972 Dennett, Daniel,28387
contextual relevance,17273 dependency relationships,
continuum hypothesis,34 12526,14041
convergent realism, 102, 103,117 autonomy in,141
Cooper, Harris,235 independence in,141
cooperation, 256,25760 irrelevancies and,144
interference with, 268, 269,271 representations of, 134,13536
Copernicus,214 derivability, 34,159
correlational information, 8586,89 Descartes, Ren,214
correspondence truth, 42, 44,118 design
counterfactual theory,161 creationists conception of,5556
covering-law models, 148,15051 function and, 45,5556
Craver, Carl,13839 indirect connectionto,54
creationist biology,5556 intentional,50
Crick, Francis, 82,8384 selection and,4749
cultural selection,9697 desires, resolving conflicts
cultural transmission,9697 about,26465
Cummins, Robert, 46, 5758n16, De Sitter, Willem,113
6266,68,70 detection,25
cumulative selection,48 notionsof,39
cytology, 15455, 159,16364 rhetoric of,2021
development
Darwin, Charles, 48, 70, 202, 210, alternative pathways for,89
213,279 dichotomous viewof,77
Daston, Lorraine, 1011, 22328,272 environmental role in,7576
Dawkins, Richard, 273, 27778, 279,287 gene-centric approaches to,9294
Deaths in Venice:The Cases of Gustav von genetic role in, 75,7778
Aschenbach (Kitcher),5 developmental biology,84
deliberation, threats from,220 developmental systems theory,85
Index [303]
Dewey, John, 11819, 149, 202, 228, ethical authority,202
24849,271 ethical beliefs, falsityof,12
difference-making ethical practice, 22526. See alsoethics
explanatory relations and,12728 collective nature of,267
factors in,162 function of, 25360,26667
facts about, systematicity in,169 performance of,26667
kairetic account of,16162 ethical progress, 25559, 262,
relationships for, 12829, 134, 268,26970
135,14041 Ethical Project, The (Kitcher), 5, 12, 226,
transitivity of,172 253, 260, 262, 267269,272
Dilthey, Wilhelm,225 ethics,5
directed graphs,13536 collective construction of,27071
directed selection,5758 conflicts about,26465
discovery,2223 conversation about,22627
accountsof,23 evolution of, 5, 96,97227
cognitive relations and,39 expertise in, 19091, 198, 201,211
mathematical,2224 function of, 25354,271
disease, origins of, 7576,94 practice of,22526
disunity, 150,176 pragmatic,251
divide et impera (divide-and-conquer) progress in, 216,228
strategy, 101, 102, 103,115 science and, 18788,19597
DN model,13135 social experimentation and,25052
Dobzhansky, Theodosius,68 as social technology, 10, 191,
doctrinal indefiniteness,277 19596,267
domino causation, 234,235 ethical truth, 25354, 256, 258, 260,262
Douglas, Heather, 249,250 behavior and,26667
drift,6061 Kantian accounts of, 12,26366
Duhem, Pierre,227 noncognitive accounts of, 12, 254,
Dupr, John, 92, 149, 150, 153, 176, 26163,270
200205 etiological analysis of functions, 7,
4546, 4953,6873
EBP. See evidence-basedpolicy etiology,
Edwards, Paul N.,207 ultra-strong notion of, 7, 5761,63
Einstein, Albert,113 middling-strong notion of, 7,
Eliot, T.S.,286 6167,7173
emotions, aptness of, 26263,270 EU1 (explanatory unification1), 9,
empathy,219 12526,14146
empiricism, 99,100 EU2 (explanatory unification2), 9,
Engel, Eduardo,143 12526, 141, 143,145
entailment relation,133 Euler, Leonhard, 21,30,40
environmental factors evaluation from within,117
developmental role of,7576 Evangelical movement,281
inherited information and,85 evidence-based policy,23334
epigenetic control systems,84 evo-devo,97
epistemic labor, division of, 209, 21012. evolution
See also cognitive labor, divisionof biology and,68
epistemic modesty, 277,278 contingencies and,213
epistemological empiricism, cultural selection and,9697
100101,108 cultural transmission and,9697
ethical approaches, Kantian,26366 function and,69
[304]Index
future trajectoryof,88 explanatory autonomy, 910, 159,164
gene-centric approaches to,9294 explanatory framework,172
morality and,96,97 explanatory irrealism,131
natural selection and,9697 explanatory models, overlapping,
religionand,5 at higher organizational
scientific evidenceand,2 levels,16768
teleosemantic viewof,78 explanatory realism,131
evolutionary biology,68,76 explanatory reductionism, 9, 10, 15354,
evolutionary debunking,9697 158, 160, 16364,174
evolutionary psychology,9596 explanatory relations,12628
evolutionary success condition,86 explanatory relevance, 15658,160
exaptations,47,57 causal models and,162
existence theorems, 15, 17,2425 systematicity in,169
expertise, democracy and, 188,190 transitivity of,159
explanation,4 explanatory unification, 3, 99, 12526,
anthropocentric considerations 141, 14243,14546
and,141 explication,8283
bringing together all objective externalism
difference-makers,168 anthropocentrism and,141
causation and, 123,148 explanation and, 13136,140
competition and,130
completeness of,16869 facts
contextual irrelevance in, anthropocentric,139
16566,167 discovery of,2021
deductive-nomological modelof,3 ultimate,122
dimensions of,13641 faith,73
epistemic approach to, 131,139 faith fibbing, 275,283
epistemic dimensions of,13641 feedback loops, 238,239
externalism and, 121, 13136, feminism, 258,271
140,141 Field, Hartry,26
fragmentation of,168 Final Overview Report of Serious
general account of,147 Case Review re Daniel Pelka
internal considerations of,13641 (Coventry Children Safeguarding
internalism and, 121, 13136, 140, Board),24245
141,15051 Finnegans Wake (Joyce),285
irrelevancies and,14145 Fisher, R.A., 15556,179
kairetic account of,16064 Fishkin, James,201
noncausal, 128,151 fitness, environment and,88
ontic approach to, 131,13738 fitness advantage,71
origin and development pattern fluids, universal behavior of,14142
of,12223 focal traits,72
pragmatic theory of,147 Fodor, Jerry, 153, 158,175
relevance relations and,14748 folk biology,77
scientific, styles of,14748 force, Newtonian,16061
success of, 13637, 147,148 formalism,2930
unification and, 3, 8, 99, 121, 12326, criticism of,37,38
127, 131, 14143, 14546,157 formsof,31
explanation-seeking questions, 147, resistance to,3436
177,180 frameworking,17274
explanatory asymmetries, 134n4,135 freedom, degrees of, variables and,136
Index [305]
free fall, theories of, 1027,11518 generalization
free postulation,19 independence of,145
Frege, Gottlob, 18, 21, 36, 38,227 inductive,4041
Fresnel, Augustin-Jean, 101, 1023, problems with,65
112,115 genes
Friedman, Michael,131 alternative splicingof,83
Friedmann, Alexander,114 assortment of, 15455, 159,175
function,3 causation and,77
accidental contribution of,5455 environment and,80
ahistorical evolutionary view expressionof,83
of,8789 narrow focuson,94
analysis of, contrasted with personalized medicine and,9495
etiological analyses,46 phenotypes and,91
artifact,52 reshuffling of,7273
causal role accountof,7 sequenceof,83
changes in, over time,5759 as signal source,80
design and, 45,5556 specificityof,83
etiological accounts of, 7, 46, 47, traits and,97
5661,64,67 transcription factors for,14445
evolution and,69 transmission of, 14849,150
explanatory roles for,6364 genetically modified foods,19294
malfunction,51 genetic determinism, 7, 8, 75, 93,95,96
natural selection and,53 genetic information, outcomes
natureof,53 and,7677
ontological groundingof,53 genetics,3,126
pathological processesas,63 autonomy of,148
pluralist theoryof,67 classical, 3, 12324, 126, 148,
roles of,45,63 150,15354
selection and, 4546, 53,7071 molecular, 68, 15354,155
teleological/teleonomic notionof,50 molecular biology and,150
tenselessness and,5051 Gibbard, Allan, 262,270
valueof,49 Giere, Ron,99100
functional analysis,62 given that/because
functional compartmentalization, distinction,17374
910,16667 GM foods,19294
functional conflict,26768 Gdel, Kurt, 19n6,21
functional difference-making,10 Gdel sentences, 3233,43
functional stratification, 910,16667 Godfrey-Smith, Peter,58
function ascriptions,49 Goldbachs Conjecture,33
function-dysfunction distinction, 63, Goodman, Nelson,180
6465,67 Gore, Al,280
function-nonfunction distinction,63 Gottlieb, Gilbert,76
function pluralism, 4546,49 Gould, Stephen Jay, 68, 6970,
7273,97
Galileo Galilei, 100, 1027, 116, 118, government, and protection of child
214,224 abuse,23031
game formalism, 30,4243 gravity, 102, 103, 125,163
gas, behavior of, 129, 137,13940 Gray, Russell, 75, 7980,92,97
gay gene,77 Griffiths, Paul E., 75, 7980, 92, 93,
gene-environment interaction, 8,7475 94,95,97
[306]Index
Hamilton, William Rowan,40 inductive generalization,4041
Hardie, Jeremy, 245,250 infant mortality,15556
Hare, Robert, 231,232 information
Harris, Sam, 273,27778 adaptive,89,91
Harvey, William, 4546, 53,268 biological, types of,7779
Hecht, Gabrielle,207 causal notionof,90
Helmholtz, Hermann von, 106,214 correlational,86
Hempel, Carl G., 131, 13234, 137, 148, dissemination of, 191,19394
15051,157n3 distribution of,28081
Herrnstein, Richard,186 genetic, amplificationof,84
H-H (Hodgkin-Huxley) model, 12930, inherited, 79, 8586,89,90
138,13940 as metaphor,94
hierarchy, 216, 218,259 mutual,8182
high-level theories, lower-level proximate, 8, 78,85,90
understanding of,155 removal of,162
Hilbert, David,43 specificity and,82
Hinde, Robert,76 teleological notionof,8
Hipparchus, 102, 104, 116,118 teleosemantics and, 79,80,86
history ultimate, 8, 78,8591
haphazard,213 information theory,78,80
learning from,22526 infotel theory,85
reflection on,227 Ingham, Elaine,193
of science, 206,2078 instruments, new, introductionof,41
speculative, 11, 21617,22425 intelligence, measurement of, 186,187
Hitchens, Christopher, 273,27778 intelligent design,5,213
Hodgkin-Huxley model,12930 intentional information,7778
Hooke, Robert,214 intentional selection,46,47
Hopf, Eberhard,143 intentionality,78
Hubble, Edwin,113 interactionism, 74,75,77
human error,24142 internalism, 13136, 140,151
Human Fertilisation and Embryology Internet
Authority (HFEA; UK),198 citizen science and,223
human flourishing, 21821, 222,22728 scientific discussions and,
humanist organizations,287 191,19495
human sociobiology,96 intertheoretical reduction,153
Humboldt, Wilhelm von,228 intervention, 80, 82, 128,13335
Hume, David, 148, 151,152 invariance, 80,128
Huneman, Philippe,64 irrelevance, 158, 173,174
black-boxing and,17072
ideal conversations, 195, 196, 198, contextual, 10, 16367,16972
20910, 21921, 222,228 unification and, 12526, 127,15051
idle wheels, 101, 102,115
ill-ordered science, 189,198 jackrabbits, ears of,5556
imaginary numbers,1920 James, William, 11819, 149, 271,
impetus theory, 102, 103,104 284,285
incompleteness theorems,4344 Jasanoff, Sheila,207
incredulity, antireductionist argument Joule, James Prescott,107
from,17475 Joyce, James,5
independence, 12526,14546 Joyces Kaleidoscope:An Invitation to
induction,247 Finnegans Wake (Kitcher),5
Index [307]
justice, global, 219,221 mad cow disease,19293
justification Maddy, Penelope,27
causal explanation and, 261,262 maintenance selection,5758
causation and,271 Making Things Happen (Woodward),147
male/female relations, ethics
kairetic theory,16064 and,25862
Kant, Immanuel, 43, 99, 100, 118, 120, Mann, Thomas,5
190, 228, 253, 26364,271 Mantzavinos, Chrysostomos, 262,270
Kepler, Johannes, 162,214 Marcellesi, Alexandre,24852
Kim, Jaegwon, 126, 13132,149 Mars, orbit of,162
King Lear (Shakespeare),285 Massimi, Michela,11419
Klebsiella planticola,19394 materials, universal behavior of,14142
knowledge mathematicalclaims
accessto,27 acceptabilityof,28
causal constrainton,39 aim of,29,31
causal theoryof,26 internal justification for,2627
detection and,27 truthof,28
explanatory constrainton,26 mathematical discovery,2224
Kohn, Linda,24142 mathematical objects, existenceof,14
Kronecker, Leopole,28n8 mathematical pragmatism,42
Kuhn, Thomas, 3, 11720, 148,150 mathematical procedures,
soundnessof,28
labeling theory,23031 mathematical truth, theoriesof,26
labor mathematics
cognitive, division of,190 anti-platonism and,67
epistemic, division of, 164,216 applicationof,44
scientific, division of, 910,16669 bridge principlesin,36
Lancet, The,194 diversity within,151
Lange, Marc,151 as domain of inquiry,3435
language games,42 formalist view of, 17,3031
LaScola, Linda, 276, 285,287 as games, collection of, 18,3032
Lash, Nicholas,285 history of, 2, 19,41,42
Laudan, Larry, 8, 98, 102, 115,11618 language of, 15n3, 1819,
lawlikeness,157n3 3536,42,44
Lehrman, Daniel,76 original contentof,42
Leibniz, Gottfried Wilhelm,214 philosophy of,2,35
Lematre, Georges,114 platonists approachto,19
Le Verrier, Urbain,23,41 subject matterof,16
Lewis, David,157 measurement, historical approachto,44
Lewontin, Richard, 68, 6970, 7273, mechanistic explanation,46,62
75, 92,93,97 medical research,150
Lindbeck, George,285 meiosis, 15455,164
linearity, 23437,241 meliorism, 203,250
Living with Darwin:Evolution, Design, and Mendel, Gregor, 20, 21, 38,120
the Future of Faith (Kitcher),5 Mendelian effect,154
Locke, John,219 Mendels laws, 130, 14546,154
logical empiricism, 99, 148, 157,176 meta-ethics,96
Lotka-Volterra model,167 metamathematics,44
lower-level theories,123 microvariables, aggregation of,136
Lysenkoism, 211, 215,224 Mill, John Stuart, 2, 43, 44, 18889,210
[308]Index
Millum, Joseph,231 nominalists, on mathematics,17
mirroring, 20910,21921 nomological machines, 149,24849
MMR vaccine, 184, 18990,194 noncognitivism, 254, 26163,270
modal claims,34 nonface value semantics,16
moderate platonism,67 normative guidance, 255, 258,260
molecular biology,3 numbers, as concrete entities,1516
molecular details, relevance of,15455 number words, use for,38
Monsanto,193 Nurse-Family Partnership,23537
moral evaluation,5
morality, evolution and,96,97 objective relevance, 169, 172,173
moral progress,216 objectivity,234
MorganKeenan system,124 objects
Morrison, Margaret,9 abstract,14,22
Munro, Eileen, 24144,250 cognitive relationsto,39
Munro Review of Child Protection detectionof,20
(UK),23738 discoveryof,21
Murray, Charles,186 role of, in mathematical practice,19
mutations,7172 Olby, Robert,8384
mutual engagement, 21516,219 ontic facts, epistemic considerations
mutual information,8182 and,140
mythically self-consciousness, 277, operational explanation,46,62
283,284 Oppenheim, Paul, 157n3,159
order, 149,248
Nagel, Ernest,153 Oresme, Nicole, 102, 103,104
Nagel, Thomas,8 organisms, designof,63
Nanney, DavidL.,84 orientation model, for religion, 275,
naturalness,157 27678,28384
natural selection, 3, 46, 4748,62 origin and development (OD)
adaptation and,87 pattern,12223
adaptiveness and,87 others, responsiveness to,26768
creationists viewof,56 oversimplification,176
evolution and,9697 Owen, Wilfred,205
history and,70 Oyama, Susan, 75, 77, 76, 92,93,97
intentionality and,78
lacking foresight, 4849,50 pain, relieving of, 25658,262
moral knowledge and,97 Paley, William,47,48
nature, pure understanding of,228 Parker, Andrew,28081
Nature Biotechnology (Fletcher),193 Parker, Geoffrey,68
Nature of Mathematical Knowledge, The particle physics,144
(Kitcher),2,16 peacemaking,25455
Neander, Karen, 68,7072 Peirce, Charles Sanders, 42, 118, 228,
neopragmatist perspectivism,120 269,271
Neptune, discovery of, 23,25,41 Pelka, Daniel,24245
Nettle, Daniel,95,96 Penrose, Roger,21
Neurath, Otto, 229,24849 performance adequacy, standards
neurons, behavior of, 12930,13840 of,10413
New Atheists, 12, 273, 275, 27778, personalized medicine,9495
283,285 perspectival coherentism,105n2
Newton, Isaac, 44, 103, 106, 125, perspectivalism, 104106
160,214 perspectival justification,105n2
Index [309]
perspectival realism, 99100, 1034,108 profiling, 249,252
perspectivism, 11617,118 progress. See also ethical progress
phenotypes direction of,2023
alleles and,7980 moral,216
dependent on genes and truth and,269
environment,74 proof, as source of knowledge,24
developmentof,78 Psillos, Stathis,98
genes and,80,91 psychopathy, recidivism and,23132
as signal receiver,80 Putnam, Hilary, 98,159
physical connections, discovery of,125
physical detail, irrelevance of, 163,164 quango (quasi-autonomous
physicalism, 15859,160 nongovernmental
physics organization),19798
causal influence in,16061 Quine, W.V.,27
explanatory unification in,125 Quintavalle, Josephine,198
fundamental, explanatory relevance
of,15860 racial science, 18587,189
modern,160 randomized controlled trials, 1112, 233,
relevanceof,9 234, 240, 245, 246,250
platonism,6 rationality,4
case for, 14,16,17 Rawls, John, 202,252
criticism of, similar to claims against RCTs. See randomized controlledtrials
formalism,37 realism,99
hard-core,1921 anthropocentrism and,141
Kitchers opposition to, 1519,2125 arguments against,1012
moderate, 1921, 26, 29n,3942 explanatory,131
Moorean responseof,26 real realism, 8, 99100, 104,
varieties of,1920 11415,11820
plenitude, principle of, 2829n8 Real Realism (Kitcher), 9899,115
pluralism, 183,210 reduction,125
plutocracy,204 reductionism, 9, 10, 125, 155. See also
Poincar, Henri,143 explanatory reductionism
Poisson, Simon Denis,101 autonomy and, 15354, 160,168
policy fall of,153
cause-effect relationship and,235 regularities, explanations of,162
ethically compromised,249 relationships
evaluation of, 234, 24041,24344 causal vs. noncausal,80
evidence-based,11 discovery of,125
experimentation with,251 explanation and, 127,151
implementation of,23334 preexisting, discovery of,140
morals and methods mixed for,233 relevance,12526
outcomes of,23337 contextual,9
overconfidence in,233 objective,16566
science and,11 types of,172
political philosophy, for science,18384 religion,5
population, evolutionary trajectoryof,88 approaches to,27576
population ecology,167 atheist critics of,273
pragmatic naturalist approach,25354 attendance trends in,28182
pragmatism, 118, 147, 149, 2023,251 authority of,282
predation, rate of,167 belief model of, 1213, 275,
principle, strategy and,27475 27677,282
[310]Index
creeds and, 275,284 democracy and, 1011, 190, 19192,
defenders of,278 19499, 201, 2023, 206, 20912,
demise of,280 218,22122
dogmatic,21011 developmentof,4
evolution of,279 disunity of,124
freedom of,273 division of labor in, 16669,
language of,284 17172,177
loss of,28687 ethics and, 5,19597
orientation model of, 1213, 275, funding for, 214,215
27678,28384 future-oriented,222
as policy issue,274 goals of,188
predominance of,279 government interference with,224
refined,28485 history of, 100101, 114, 206,
value of, 27980, 283,28687 2078,21317
as way station toward secular human flourishing and, 207,208
humanism,285 interacting with other
religious diversity,273 institutions,248
renormalization techniques, 126, layer cake model of, 167,183
142,143 maladaptation in,213
representation,137 as means of accumulating truth,182
reproduction,69,87 methods of,229
research as nomological machines,24849
attraction to,223 philosophers of, 146, 149, 157,182
banning of,189 philosophy of, 121, 148, 172, 176,
democratic decisions about,18889 206,207
ethics and,18788 pluralism in,183
funding of, 185, 189,199 political philosophy of, 18384,
progress in,2023 190,199
subjects for, 183,18485 as political problem,183
theories and,247 popularization of,223
truth and,186 progress and,22122
values and,249 project of,208
Resnik, Michael,30 promoting human flourishing,
Richerson, Peter,9697 21821,222
Rilke, Maria Rainer,285 and the public good,207
Rimbaud, Arthur,285 public perception of, 19294,
Rosen, Gideon,284 2089,24647
Royal Society,214 public policy and, 10, 11, 19697,214
Russell, Bertrand,247 pure,214
realismin,98
Salmon, Wesley, 131, 137, 138,147 relation of, to society,4
Santa Barbara paradigm,95,96 resisters and,195
Sarkar, Sahotra,82 role of, in thought and action,24849
science self-image of,222
applied, 18485,21415 as social technology, 10, 19192,196
authority of,208 as source of knowledge,182
autonomy of, 11,21115 as system of public knowledge,248
consensus and, 194,195 unity of, 124, 149, 159, 172,183
contexts for,2078 values and, 188, 195, 206, 210,229
conversation about,19499 Science, Truth, and Democracy (Kitcher),
decisions about,190 4, 149, 18586, 199200,286
Index [311]
Science in a Democratic Society (Kitcher), Smith, Michael, 73, 26772,284
45, 185, 190, 194, 199200,206 social engineering,250
ScienceThe Endless Frontier social experimentation,25052
(Bush),21415 social friction,268
science studies,206 social policy, 24546. See alsopolicy
scientificclaims social problems, systemic nature of,242
assessments of,105 social progress,203
success criteria for,10613 social technologies, 19697,267
scientific communities, success of,107 societies, developing ethics,5
scientific creationism,2 special sciences
Scientific Explanation and the Causal antireductionist theses about,
Structure of the World defense of,146
(Salmon),137 autonomy of, 9, 121, 123,154
scientific explanation,23 difference-making relationships in,128
scientific knowledge, 34, 99, 182,248 reductionist philosophy of
scientific models,183 explanation for,154
scientific perspective,106 unification and,121
scientific realism, 103,108 specificity
scientific research. See research causal,8485
scientists causal democracy and,8385
generalization problemsof,65 information and,82
public perceptions of,19294 measure of,8182
secondary causes,56 sourcesof,84
secular humanism, development of,287 speculative history,11
selection Sprat, Thomas,224
cumulative,48 Srinivasan, Amia, 258, 260, 266, 271,272
design and,4749 standard environment, defined,7980
directed,5758 statistical relevance theory,161
function and,7071 Stegmann, Ulrich,82
history of,6970 Sterelny, Kim, 78, 7980, 91, 94,260
maintenance,5758 Stotz, Karola,92
power of,7071 stratification, 910, 167. See also
presence and,60 functional stratification
types of,46,47 Street, Sharon,96
sequence hypothesis,82,83 Strevens, Michael,17681
sex ratios, 15556, 159,179 structural realists,1012
Shadwell, Thomas, 214,224 structure/substance dichotomy,1012
Shannon, ClaudeE.,8 subsumption relation, 125, 132,133
Shannon entropy,81 success, typesof,8
Shea, Nicholas, 8, 79, 8586, 8990,93,95 success from nowhere,113
Shiffrin, Seana,231 success from within, 8, 1028,11114
Shoesmith, Sharon, 230, 241,250 survival, reproduction and,87
signaling networks,90 survival value,87
significance,286 sympathy, 216,21921
significant truth,18283 systematicity,16972
simplicity,110 systems, sequence of,3233
Simplicius,102
slavery, abolition of,216 technology, emergence of,191
Slipher, V.M.,11213 teleosemantic information, 78, 79,85,91
Smith, Adam,209 teleosemantics
[312]Index
ahistorical,79 difference-making and,149
successof,93 explanation and, 12225, 127,131
theories irrelevance and,150
falsity of, 100103 Kitchers version of, 12224,136
semantic conception of,148 local,150
theory of the game,18 as regulative ideal,150
thermodynamics, 107n5,145 in science,14950
Thomae,J.,42 UnitedStates
Thomson, William (Lord Kelvin),107 Constitution of, interpretation
Tillich, Paul,285 of,222
Tinbergen, Nikolaas,69,87 freedom of religion in,273
To Err Is Human:Building a Safer Health sciences public role in,2089
System (Kohn etal.),24142 upper-level relationships, autonomy
traits in,12930
adaptation and,87,88 upper-level theories,123
ahistorical function and,8889
bound together,72 vaccination, 184, 18990,194
fashioningof,48 values
functionof,87 research and,249
inability to perform a function,51 science and, 188, 195, 206,210
performanceof,67 Van Fraassen, Bas, 147, 177,179
reproductive success and,69 vestiges, 5052,57
survival valueof,87 Von Liebig, Justus,214
typingof,67
vestigial,5052 Wakefield, Andrew,194
transcription,14445 Wallace, R.Jay,231
transcriptome,83 War Requiem (Owen),205
transgenerational environmental Waters, Kenneth,94
influences,76 wave theory of light,101
transitivity, 15657, 159, 172,176 welfare, policy outcomes for,23740
translation, 117,14445 well-ordered science, 4, 10, 11, 184, 189,
translational research,9495 193, 194, 201,209
truth,18283 climate change and,203
assessmentof,43 conditions for,211
progress and,269 desirability of,211
selective broadcasting of,275 ethics and,2045
significant,286 harmonizing science and
universal,240 democracy,201
value of,286 human flourishing and,219
truth-telling,280 ideal of, 21516,227
Two New Sciences (Galileo), 104,107 outcome of,196
process of,222
Ultra-Strong Selection utopianism of,212
Requirement,6061 Williams, Garrath,231
unification, 23, 121,148 Wittgenstein, Ludwig,30,42
approaches to,14849 women, rights for, 25860, 26162,271
aspects of,14145 Woodward, James (Jim), 81, 82, 84, 94,
comparative,135 14652,176
competition and,130 working posits, 101, 1023,115
constraints related to,122 Wright, Larry, 46, 4953, 64, 66
Index [313]