Measurement 34 (2003) 39–48

www.elsevier.com / locate / measurement

Widely, strongly and weakly defined measurement

Ludwik Finkelstein*
Measurement and Instrumentation Centre, School of Engineering and Mathematical Sciences, City University, London EC1 V 0 HB, UK

Abstract

The paper discusses the concept of measurement. Measurement, in the wide sense, is defined as a process of empirical,
objective assignment of symbols to attributes of objects and events of the real world, in such a way as to describe them.
Strongly defined measurement is measurement that conforms to the paradigm of the physical sciences. Weakly defined
measurement is measurement in the wide sense, but which is not strongly defined. Strongly and weakly defined
measurements are analysed and compared. Other forms of symbolic representation are distinguished from measurement.
 2003 Elsevier Science Ltd. All rights reserved.

Keywords: Measurement, fundamental concepts; Measurement, widely defined; Measurement, strongly defined; Measurement, weakly
defined

1. Introduction symbolic information that is not numerical, has given

Measurement is an essential enabling tool of
modern thought. It is a means by which we describe
the world. Its application has been extended from the
physical sciences and technologies, in which it had
its origins, to other domains of knowledge, such as
the social, behavioural and decision sciences. In our
thinking we have adopted the programme of Galileo:
to count what is countable, measure what is measur-
able, and to make measurable what is not measur-
able.
The requirements of those wider domains of
knowledge and reasoning have led to the broadening
of the concepts of measurement, so as to encompass
other forms of objective symbolic description.
The development of computing, which can handle
a further impetus to the evolution of the theory of
symbolisation, and of the concepts of measurement
widely defined.
The broadening of the concept of measurement
has been disputed, and measurement in the physical
sciences remains the normative paradigm In any
case, it is generally accepted that some forms of
measurement, widely defined, are stronger than
others.
This paper deals with the aspects of more widely
defined measurement, and the distinction between
strongly and weakly defined measurement.
2. Outline of the historical development of
measurement theory

Since this paper is concerned with the evolution of

*Tel.: 144-020-7040-8139; fax: 144-020-7040-8568. the concepts of measurement it is useful to begin by
E-mail address: l.finkelstein@city.ac.uk (L. Finkelstein). a brief outline of their historical development.

0263-2241 / 03 / $ – see front matter  2003 Elsevier Science Ltd. All rights reserved.
doi:10.1016 / S0263-2241(03)00018-6
40 L. Finkelstein / Measurement 34 (2003) 39–48
It is not proposed to discuss in any detail the
historical development of the theory of measurement
proper. Guidance to more general topics may be
sought in relevant reference works [1–5]. The works
concerned give guidance to more detailed infor-
mation.
The practical pursuit of measurement in com-
merce, crafts, and surveying arose in the earliest
stages of the material culture of man. It grew in the
course of the Urban Revolution in Mesopotamia and
Egypt and in parallel developments in China. It was
based on an intuitive understanding of the process.
The ancient Greeks were the first to investigate the
philosophical foundations of measurement. The
school of Pythagoras was concerned with the
philosophy of the relation between numbers and the
real world and hoped to make arithmetic the fun-
damental study in physics. Plato’s Academy de-
veloped an extensive theory of magnitudes, though it
was concerned more with the nature of numbers than
with an analysis of the nature of measurement.
Aristotle studied the concepts of measurement in his
Metaphysics.
The Middle Ages saw much scholarly study of the
theory of measurement, though scholars were not
concerned with the application of measurement to
observation of the physical world.
With the rise of modern science, Galileo put
mathematics at the heart of physics and established a
programme of application of measurement to the
observation of the physical world. Newton in his
development of mechanics provided the first com-
prehensive mathematical theory of a domain of
physics. Measurement was increasingly applied to
scientific observation, culminating in full mathemati-
cal theories of domains of physical science such as
the celestial mechanics of Lagrange and the electro-
magnetic theory of Maxwell. These profound de-
velopments of the application of measurement were
not, however, accompanied by studies of the fun-
damental concepts of measurement.
The true foundations of the modern theory of
measurement were laid by Helmholtz [6] in a
thorough logical analysis of the epistemology of
counting and measuring published in 1887. The work
was part of the beginning of studies in the logical
foundations of mathematics with which the theory of
measurement is closely connected. An important
development of the work of Helmholtz was the
axiomatization of measurement of additive quantities
by Hoelder in 1901 [7].
The British physicist, N.R. Campbell provided in
1920 a lucid and thorough analysis of the fundamen-
tal basis of the measurement of physical quantities in
his Physics: the Elements [8]. The book can still be
read with pleasure and profit. Campbell’s theory was
based on the measurement of physical quantities for
which an empirical operation of addition could be
constructed, what is now known as extensive mea-
surement. The theory became generally accepted
under the influence of logical positivism and
operationalism [3].
The theory of measurement from Helmholtz to
Campbell, and those who developed their work, was
concerned with physical measurements, which are
based on additive quantities and quantities derived
from them. In the social and behavioural sciences,
however, there are many properties that cannot be
empirically added nor derived from additive quan-
tities. This created great philosophical difficulty. A
report of a committee of the British Association for
the Advancement of Science, published in 1940,
which considered quantitative methods, rejected
measurements not based on additivity, and hence the
possibility of psychological measurements. It even
questioned the status of the thermodynamic scale of
temperature [9].
These rigid positions of the classical theory of
measurement were broken down by work in the
social and behavioural sciences. The concern of the
social sciences with the concept of utility lead
through the work of the early writers, Bentham and
Pareto, to the classical work of von Neumann and
Morgenstern in 1944 [10]. Their axiomatic formula-
tion of the theory of utility has been the basis of
much work in the theory and practice of measure-
ment in the social sciences. In psychology S.S.
Stevens carried out much fundamental work on
developing an appropriate analysis of the nature of
measurement [11–13].
The proceedings of a conference in the United
States [14], published in 1959, presented a review of
the classical approaches to measurement as extended
to needs of the social and behavioural sciences.
 L. Finkelstein / Measurement 34 (2003) 39–48
Ellis has provided, in 1966, a most useful and
interesting philosophical analysis of measurement. It
is not formal in approach and is principally
concerned with physical measurement, though it
takes into account non-classical theory [15].
Modern formal measurement theory may be said
to be based the work Tarski on relational systems
and model theory. This contribution is well described
in Ref. [16]. The theory, which may be termed
representational theory of measurement, considers
measurement, loosely speaking as the establishment
of a correspondence between a set of manifestations
of a property and the relations between them, and a
set of numbers and the relations between them.
Suppes and Zinnes provided in 1963 an early clear
exposition of the theory, in the development of
which Suppes has been one of the key workers [16].
The representational theory of measurement was then
developed and well documented in a group of
important books published between 1971 and 1990.
Pfanzagl published in 1968 the first monographic
work devoted to the theory of measurement with a
representational approach [17]. Krantz et al. pub-
lished between 1971 and 1990 a very detailed and
thorough account of the representational foundations
of measurement, in three volumes [18]. Other
thorough accounts are those of Roberts [19] and of
Narens [20].
The representational theory has now been accepted
widely in the social and behavioural sciences. How-
ever, while it embraces physical measurement, the
theory has not found a recognised place in the theory
and practice physical sciences.
Modern measurement theory has been introduced
to the measurement and instrumentation technology
community by Finkelstein, who has regularly re-
viewed its development and progress [21–27]. The
topic has been taken up and debated in the communi-
ty, particularly by some of the authors represented in
this issue.
Representational measurement theory has not pro-
gressed significantly in recent years, as it seems to
have proved adequate for the purposes for which it
has been developed. However the rapid development
of symbolic computation and soft computing has
offered prospects for its further development and
extension [26,27].
41
3. Wide sense definition of measurement
It is proposed here to adopt a wide definition of
measurement, as a basis of discussion.
Measurement, in the wide sense, is thus defined as
a process of empirical, objective assignment of
symbols to attributes of objects and events of the real
world, in such a way as to represent them, or to
describe them [27]. Measurement theory provides the
logical basis for the analysis of the conditions of
representation.
The essential elements of this wide-sense defini-
tion will be analysed informally.
Firstly, measurement is defined as the assignment
of symbols to properties of objects and events, and is
thus the description of properties of objects or events
and not of the objects or events. Measurement is
based on a clear concept of a property, as an abstract
aspect of a whole class of objects of which in-
dividual instances, or manifestations, are the subject
of measurement.
The definition states that the assignment of sym-
bols in measurement is such that the symbols
describe the property of the object or event. The
meaning of this can be explained as follows. Consi-
der that a symbol, or measure, is assigned by
measurement to the property of an object, and other
symbols are assigned by the same process to other
manifestations of the property. Then the logical
relations between the symbols or measures, in the
symbolisation system adopted, imply and are implied
by empirical relations between the property mani-
festations.
The next aspect of the definition of measurement,
which requires discussion, is the fact that measure-
ment is an objective process. By this is meant that
the numbers assigned to a property by measurement
must, within the limits of error, be independent of
the observer.
The informal definition of measurement presented
above stresses the fact that measurement is an
empirical process. This means first that it must be the
result of observation and not, for example, of a
thought experiment. Further, the concept of the
property measured must be based on empirically
determinable relations and not, say, on convention.
It should be noted that all measurement involves a
42 L. Finkelstein / Measurement 34 (2003) 39–48
comparison of the measurand with a standard. either
material, or conceptual. Measurement theory pro-
vides the logical basis for that comparison.
4. Pragmatics of measurement
The adoption of a wide definition of measurement
necessitates a comparison between widely defined
measurement and other forms of symbolic descrip-
tion [27].
Firstly, measurement is free of the ambiguity and
vagueness of other forms of description, say those of
natural language.
Secondly, measurement is not mere naming, but
provides information about the relation of a par-
ticular manifestation of a quality and other mani-
festations of the same quality.
Thirdly, and most significantly, the objectivity of
measurement makes measures invariant and indisput-
able in logical discourse.
Finally, measurement is based on empirical ob-
servations and empirical relations. It thus provides
strong knowledge.
These pragmatic strengths of measurement apply
to widely defined measurement, as much as to
measurement more strongly defined.
5. Strongly and weakly defined measurement
The wide definition of measurement is often
disputed by those who consider the paradigm of
measurement in the physical sciences as normative,
or, at least, require measurement to be a numerical
representation in which reflects an order.
For this reason it is convenient to distinguish
between strongly and weakly defined measurement.
5.1. Strongly defined measurement
Strongly defined measurement is defined as a class
of widely defined measurement, which follows the
paradigm of the physical sciences.
Measurement in the physical sciences is based on:
(i) precisely defined empirical operations,
(ii) mapping on the real number line on which an
operation of addition is defined,
(iii) well-formed theories for broad domains of
knowledge.
5.2. Weakly defined measurement
Measurement that constitutes representation by
symbols of properties of entities of the real world,
based on an objective empirical process, but lacks
some, or all, of the above distinctive characteristics
of strong measurement, may be termed weakly
defined.
6. Elements of the formal theory of
measurement
6.1. Representation by numbers
For the sake of completeness the concept of
measurement, informally presented above will now
be presented formally using the representational or
model theory approach. The presentation follows that
given in [27] and the literature quoted therein.
A representational theory of measurement has four
parts:
(i) an empirical relational system corresponding to
a quality,
(ii) a number relational system,
(iii) a representation condition,
(iv) a uniqueness condition.
These will now be considered.
( i) Quality as an empirical relational system
A quality is a property of an object.
Consider some quality and let qi represent an
individual manifestation of the quality Q, so that we
can define a set of all possible manifestations as
Q 5 hq1 . . . j (1)
Let there be on Q a family R of empirical relations
Ri
R 5 hR 1, . . . R n j (2)
 L. Finkelstein / Measurement 34 (2003) 39–48
Then the quality is represented by an empirical
relational system
Q 5 kQ, Rl,
( ii) Numerical relational system
Let N represent a class of numbers
N 5 hn 1 . . . j
Let there be on N a family P of relations
P 5 hP1 , . . . ,Pn j.
Then
N 5 kN, Pl
represents a numerical relational system.
Commonly N is just the real number line Re, with
numerical relations defined on it.
( iii) Representation condition
The representation condition requires that mea-
surement be the establishment of a correspondence
between quality manifestations and numbers in such
a way that the relations between the referent property
manifestations imply and are implied by the relations
between their images in the number set. Formally,
measurement is defined as an objective empirical
operation M
M: Q → N
so that
n n 5 M(qn ).
such that Q 5 kQ, Rl is mapped homomorphically
into (onto) N 5 kN, Pl.
The above homomorphism is the representation
condition.
Firstly it implies that if qn is related to qm by an
empirical relation R kk , that is R kk (qn , qm ), Pk is the
numerical relation corresponding to R k,, , n n 5 M(qn )
is the image of qn in N under M then R kk (qn , qm )
implies and is implied by Pkk (n n , n m ).
Measurement is a homomorphism, rather than an
isomorphism, because M is not one-to-one, it maps
separate but indistinguishable property mani-
festations to the same number.
S 5 kQ, N, Ml
constitutes a scale of measurement for Q.
43
n n 5 M(qn ), the image of qn in N under M is
called the measure of q j on scale S
(3)
( iv) Uniqueness condition
The requirement that the fundamental measure-
ment procedure of a scale should map the empirical
relational system Q homomorphically into the nu-
(4) merical relational system. N does not determine the
mapping uniquely.
There is an element of arbitrary choice in the
(5) setting up of scales of measurement. For example in
the case of scales based on additive combination, for
instance, the choice of the unit standard is arbitrary.
(6) The requirement of homomorphism thus defines a
class of scales that may be called equivalent. The
class of transformations, which transform one
member of a class of equivalent scales into another,
is called the class of admissible transformations. The
conditions which admissible transformations must
satisfy, are known as the uniqueness conditions.
6.2. Representation by general symbols
Measurement as defined above can be seen as a
special case of general representation of entities by
symbols.
An object termed the symbol may be said to
represent, or to carry information about another
(7)
object or event termed the referent, by bearing a
known relation to it.
We can now briefly present the formal theory of
(8)
symbolic representation, which essentially follows
the discussion of measurement above.
Let qn be a referent entity. Consider further that qn
is a member of a family or set of similar entities Q,
Q 5 hq11 . . . j, (10)
Q is termed the referent set. Let there be on Q a
family R of relations
R 5 hR 1, . . . ,R n j. (11)
We may term Q 5 kQ, Rl, the referent relational
system.
Let now z n be a symbol entity. Consider further
that z n is a member of a family or set of similar
(9) entities Z,
Z 5 hz 1 , . . . j, (12)
44 L. Finkelstein / Measurement 34 (2003) 39–48
Z is termed the symbol set. Let there be on Z a
family P of relations
P 5 hP1, . . . ,Pn j.
We may term Z 5 kZ, Pl the symbol relational
system.
Let there be a mapping
M:Q → Z
so that z i 5M(q i ). Further let there be a mapping
F:Q → Z
so that Pn 5 F(R n ).
We may define
C 5 kQ, Z, M, Fl
where C is the representation code, and its inverse,
the interpretation code. z n is termed symbol of or for
qn .
6.3. Language
For the purposes of this paper it must suffice to
state briefly that a language is a collection of
symbols, together with the rules of their combina-
tion, or syntax. in other words, a general symbol
representation system.
Thus using the kind of notation employed above a
language L for the description of objects of the real
world and their relations, may defined as
L 5 kZ, P, Gl
where Z is a set of individual symbols, P is a set of
relation symbols and G is a set of function symbols.
A typical sentence in such a language is Pi (z 1 , z n ).
A code, such as C above, describing the corre-
spondence of the linguistic symbols to the real
world, or more generally to extra-linguistic entities
and their relations, constitutes the semantics of the
language.
6.4. Theory
A theory T in a language L is a set of sentences in
L.
A theory may be a representation, or model, of a
domain in the extra-linguistic world under a set
appropriate codes C1 , C2 , . . . ,Cn .
7. Comparison of strongly and weakly defined
measurement
(13)
Having defined strongly and weakly defined mea-
surement it is now proposed to compare their
characteristics.
Strongly defined measurement is based on the
mapping of an empirical relational system into a well
(14) defined theory in a mathematical language. The
terms language and theory are used in the senses
explained above.
(15) In strongly measurement:
(i) the definition of the quantities is based on the
theory,
(16)
(ii) the empirical relational system is also defined
by the theory,
(iii) the symbolic relational system is rich,
(iv) the theory for the domain is complete.
Weakly defined measurement, has one or more of
the following features:
(i) it is based on an ill-defined concept of the
quality,
(ii) there is significant uncertainty in the empirical
relational system that it represents,
(iii) the symbolic relational system has limited rela-
tions defined on it,
(iv) there is no adequate theory relating the measure-
ment to other measurements in the same do-
(17)
main.
It is to be noted that representation by symbols
that is not objective, empirical, or does not represent
an empirical relational system on the quality, does
not constitute measurement.
The various aspects of strongly and weakly de-
fined measurement will now be considered in more
detail.
8. Concept definition
The establishment of a scale of measurement of a
quality must be based on a satisfactory concept of
the measurand.
The formation of a quality concept is normally a
 L. Finkelstein / Measurement 34 (2003) 39–48
recursive process. The establishment of a basic
concept leads to a scale of measurement defined on
it. Observations using the scale of measurement lead
to a refinement of the concept and hence to a
development of the scale. Measurement with a
refined scale may ultimately lead to a rich theory of
the domain, which establishes a definition of the
concept and strongly defined measurement scales.
An example of such recursive concept formation
in the physical sciences is the development of the
concept of temperature. Starting with a vague con-
cept of ‘the degree of hotness of a body’, scales of
temperature based on practical instruments were
developed. Such scales led to the refinement of the
concept of temperature and improved thermometers.
Observations with improved thermometers led to
theories of heat and ultimately to an integrated
theory of thermodynamics. It is on such an integrated
theory of thermodynamics that the temperature scale
is currently defined in physical science [28,29]
Problems of concept formation arise commonly in
social science and psychology.
Consider for example the concept of poverty,
which is much discussed in political discourse in
natural language. It is extensively discussed in
sociological literature, and is the subject of much
sociological measurement. The development of the
concept is clearly considered in [3] and [4], which
cite more detailed literature. Sociologists distinguish
between absolute and relative poverty. Absolute
poverty occurs when people fail to receive sufficient
resources to maintain health and efficiency, often
expressed in terms of calories, or other nutritional
measures. Relative poverty is defined by the general
standard of living in a society, and by what is
culturally deemed as poor. The relative poverty of a
household may be described by reference to some
percentile of income distribution, which is deemed to
represent the poverty line. Relative poverty thus
varies between societies and for the same society
over time. It is not possible here to consider the
measurement of poverty in any detail, but it can be
seen from the above that the definitions of poverty
are theory-laden and hence measurement of poverty
is weak in the sense defined above.
Another example of the problem of concept
formation is the notion of intelligence [30]. The term
intelligence is much used in natural language dis-
45
course, but its definition is vague. Ref. [31] provides
an excellent introduction to the problem of the
formation of the concept of intelligence and its
assessment, and cites more detailed literature. Many
tests are available for measuring intelligence, but
there is no certainty as to what intelligence is, or
what intelligence tests are measuring. It may be that
they are merely measuring the ability to score on the
same form of intelligence test [32]. The measure-
ment of intelligence by intelligence tests, is an
objective and empirical process, but is, in the sense
defined above, a weakly defined measurement.
9. Uncertainty
All measurement is accompanied by uncertainty.
However, there are differences in kind between
strongly and weakly defined measurement with
respecy to the nature of the uncertainty involved.
Referring to the notation above we may distin-
guish between uncertainty in the mapping M and
uncertainty in the empirical relational system R.
In strongly defined measurement R is defined by a
well-founded and well-formulated theory. M is again
theoretically founded. Uncertainties are considered to
be imperfections of the measurement process, and
are estimated by corresponding analysis.
In weakly defined measurements, both R and M
are based on empirical observations and empirical
laws, in the determination of which there is uncer-
tainty,
Good examples of uncertainty in weakly defined
measurement are provided by psychophysical mea-
surements of perceptions of sensory stimuli, such as
brightness and loudness. The problem is reviewed in
Ref. [32], which gives references too more detailed
literature. It makes clear that human judgements of
stimuli are poor, and that substantial uncertainty in
both the concept and scale formation arises.
10. Strength of measurement scale
10.1. In strong measurement
In the classical approach to strictly, or strongly,
defined measurement such measurement is confined
46 L. Finkelstein / Measurement 34 (2003) 39–48
to fundamental, or extensive measurement [24]. In
such measurement, the empirical relational system in
the quality Q, is
R 5 ( ¯ , p, o),
consisting of an empirical partial order, and an
empirical concatenation operation. The measurement
operation M maps this Q into the real number line
Re, with 5, 1 defined on it.
For qualities for which this is not possible, we
may define conjoint measurement structures, on the
basis of empirical relational systems established for
qualities conjointly [18].
Alternatively, it is possible to treat qualities that
are not fundamentally measurable as derived vari-
ables [24]. The scales of measurement are then
derived from an empirically established relation
between two fundamentally measurable quantities,
which maps manifestations of the variable into
numbers, which have the same order as the mani-
festations have in the concept of the quality.
The approach to measurement adopted in the
paradigm of the physical sciences is to establish a
theory for a domain. The theory is mathematical. The
concept of a variable is then defined by the theory.
Measurement is an objective empirical operation,
which maps manifestations of a variable into sym-
bols of the theory.
10.2. In weak measurement
In weakly defined measurement, scales of mea-
surement are classified in the literature according to
the uniqueness conditions and their invariance under
a class of transformations [17–20,24]. Usually scales
of measurement are classified as nominal, ordinal,
interval and ratio. The wider the class of admissible
transformations of a scale, the narrower is the class
of meaningful statements in terms of the symbols of
the scale.
For example, in nominal measurement the quality
Q has only an empirical indifference relation ¯ on
it. The symbol system Z is one in which the symbols
Z may be very general and the symbolic relational
system P consists only of the identity relation 5.
The scale is invariant under any one-to-one trans-
formation. Statements implying order, or distance,
are thus not meaningful in the scale.
An example of a nominal measurement is the
classification of members of a population according
to their first language, if it is objective and based on
empirical observation. Such nominal measures may
be used for empirical observations, relating first
language to, say, income, religion and educational
level. The class of meaningful statements in terms of
nominal measures is limited. In statistics it is mean-
ingful to state a mode of a distribution, but not a
median, or a mean.
It should be noted that wide measurement admits a
very wide class of empirical relations R, and corre-
spondingly a variety of symbolic systems Z. There
are, as yet, no measures developed for the strength of
diverse wide symbolisation systems.
Staying with linguistic examples, it is possible to
classify languages by empirical and objective criteria
[33]. in such a way that it is possible to draw a
graphical tree describing a genetic classification of
languages to determine say, a distance of one
language from another. For example Polish is close
to Slovak, more distant from Russian, yet more
distant from German, and even more distant from
Arabic.
In the case of biological classification it is possible
by empirical observation a living organism as Homo
sapiens, or Rana rana, and to determine the distance
between them, by graph theoretical measures [34].
11. Theory for strongly and weakly defined
measurement
In strongly defined measurement the measure of a
quantity is related to other measures of other quan-
tities on the same, or other, objects by a well
formulated and extensive theory. For example mea-
sures of force on an object are related to measures of
mass and acceleration.
In weakly defined measurement such a theory is
commonly lacking. For example, intelligence test
scores may not be related with confidence with other
measurements. To give an example from physical
measurements, measures of hardness on the Mohs’
 L. Finkelstein / Measurement 34 (2003) 39–48
scale [24] are not related by an adequate theory to
other measures of properties of the object.
12. Symbolisations which are not measurement
It is useful for the sake of completeness to
consider here assignments of numbers, or other
symbols, to properties in such a way as to describe
them, but which are not measurements.
One example is classification of industries by a
standard industrial classification. Assignments, on
such a classification, of numbers to enterprises, are
objective and based on an empirical observation.
They represent the relation of the enterprise to others
symbolised using the same classification. However,
they are not measurement, to the extent that the
system of classification is not based on an objective
empirical system of relations of industries [35].
An important class of the descriptive assignment
of numbers, the measurement status of which is
problematic, arises in educational testing [36]. Marks
in examinations may be objective, and are based on
an empirical process, but it is problematic what they
measure, other than the performance in a particular
test. It is doubtful whether, when marks are treated
as measures on a ratio scale, they are not in fact
measures on an ordinal scale. This affects the
meaningfulness of statistics on marks, such as t
calculations of averages and the like. The conflation
of marks, such as the calculation of weighted sums
of marks, contains an element of subjectivity in the
conflation scheme, which probably disqualifies such
conflated marks from being considered measure-
ments.
As a final example of descriptive numerical as-
signments that have doubtful measurement status, it
is important to consider utility evaluation in formal
decision making. Problems of decision sciences were
an important stimulus to the development of wider
concepts of measurement [19]. Formal concepts of
utility, with considerable power, have been de-
veloped. However, it appears, that utility, however
strongly defined, is only a description of the subjec-
tive judgement of the decision-maker. It is not a
measure of any objectively defined characteristic of
the object evaluated.
47
13. Measurement and natural language
There is a strong relation between description by
measurement, in the weakly defined sense, and
description by natural language, which is in some of
its functions a general form of symbolic representa-
tion.
It is not possible within the scope of this paper to
consider natural language, even in outline. The barest
skeleton of the comparison between measurement
and description by natural language must suffice.
In measurement the meaning of a symbol is its
reference. In natural language there are other views
of meaning. Meaning may be related to an idea in
the mind of the originator or receiver of an utterance,
or it may be considered as determined by conven-
tional use [37].
The function of measurement is informational.
Natural language has other functions such as aes-
thetic or phatic [37].
A linguistic symbol, even in its informational
function may, in addition to its denotation, convey
other meanings, such as emotional colour.
As discussed above the essence of measurement is
that it is an objective, empirical process. Description
by natural language may be derived from empirical
observation, though it is not necessarily so. It may be
subjective, though it may have a high level of
objectivity.
Description by natural language has often a degree
of ambiguity, and vagueness. Description by mea-
surement has generally a high degree of precision.
Finally description by measurement is generally
more concise than description by natural language.
14. Conclusions
The paper has proposed the concepts of widely,
strongly and weakly defined measurement. It has
analysed their respective characteristics.
Acknowledgements
The author wishes to thank the Directors of the
Measurement and Instrumentation Centre of City
University for their encouragement in his work.
48 L. Finkelstein / Measurement 34 (2003) 39–48
References
[1] P.H. Sydenham, Measurement and instrumentation: history,
in: L. Finkelstein, K.T.V. Grattan (Eds.), Concise Encyclo-
pedia of Measurement and Instrumentation, Pergamon, Ox-
ford, 1994, pp. 1997–2001.
[2] T. Honderich (Ed.), The Oxford Companion to Philosophy,
Oxford University Press, Oxford, 1995.
[3] N. Abercrombie, S. Hill, B. Turner, in: The Penguin Diction-
ary of Sociology, 4th Edition, Penguin Books, London, 2000.
[4] G. Marshall (Ed.), Oxford Dictionary of Sociology, Oxford
University Press, Oxford, New York, 1998.
[5] D.W. Pearce (Ed.), Macmillan Dictionary of Modern Econ-
omics, 4th Edition, MacMillan, Basingstoke, London, 1992.
[6] H. v. Helmholtz, Zaehlen und Messen Erkentiss—theoretisch
betrachet, Philosophische Aufsaetze Eduard Zeller gewidmet,
Fuess, Leipzig, 1887.
[7] O. Hoelder, Die axiome der quantitaet und die lehre vom
mass, Ver. Koen Sachs. Gessel. Wiss. Math. Phys. Klasse 53
(1901) 1–64.
[8] N.R. Campbell, in: Physics, The Elements, Cambridge
University Press, Cambridge, 1920.
[9] British Association for the Advancement of Science, British
Association for the Advancement of Science. Final Report,
Adv. Sci. 2 (1940) 331–349.
[10] J. von Neumann, O. Morgenstern, in: Theory of Games and
Economic Behaviour, Princeton University Press, Princeton,
NJ, 1944.
[11] S.S. Stevens, On the theory of the scales of measurement,
Science 103 (1946) 677–680.
[12] S.S. Stevens (Ed.), Handbook of Experimental Psychology,
Wiley, London, 1951.
[13] S.S. Stevens, Measurement, psychophysics and utility, in:
C.W. Churchman, P. Ratoosh (Eds.), Basic Concepts of
Measurement, Cambridge University Press, Cambridge,
1959, pp. 1–49.
[14] C.W. Churchman, P. Ratoosh (Eds.), Basic Concepts of
Measurement, Cambridge University Press, Cambridge,
1959, pp. 1–49.
[15] B. Ellis, in: Basic Concepts of Measurement, Cambridge
University Press, Cambridge, 1966.
[16] P. Suppes, J.L. Zinnes, Basic measurement theory, in: R.D.
Luce, R. Bush, E. Galanter (Eds.), Handbook of Mathemati-
cal Psychology, Wiley, New York, 1963, pp. 1–76.
[17] J. Pfanzagl, in: Theory of Measurement, Physica Verlag,
Wuerzburg, Vienna, 1968.
[18] D.H. Krantz, R.D. Luce, P. Suppes, A. Tversky, in: Founda-
tions of Measurement, Vols. 1–3, Academic Press, New
York, 1971–1990.
[19] F.S. Roberts, in: Measurement Theory with Applications to
Decisionmaking, Utility and the Social Sciences, Addison
Wesley, Reading, MA, 1979.
[20] L. Narens, in: Abstract Measurement Theory, MIT Press,
Cambridge MA, and London, 1985.
[21] L. Finkelstein, Principles of measurement, Trans. S.I.T. 15
(1963) 181–189.
[22] L. Finkelstein, Fundamental concepts of measurement, in:
Measurement and Instrumentation, Acta IMEKO VI, Vol. 1,
North Holland, Amsterdam, 1974, pp. 11–18.
[23] L. Finkelstein, Representation of symbol systems as an
extension of the concept of measurement, Kybernetes 4
(1975) 215–223.
[24] L. Finkelstein, Theory and philosophy of measurement, in:
P.H. Sydenham (Ed.), Handbook of Measurement Science,
Vol. 1, Wiley, Chichester, 1982, pp. 1–30.
[25] L. Finkelstein, in: Leaning Towards an Applied Theory of
Measurement, Acta IMEKO X, IMEKO, Budapest, 1985, pp.
287–296.
[26] L. Finkelstein, Measurement, information, knowledge: fun-
damental concepts, philosophical implications, applications,
in: From Measurement to Innovation, Acta IMEKO XIII,
Vol. 1, IMEKO, Budapest, 1994, pp. 11–18.
[27] L. Finkelstein, Foundational problems of measurement, in:
K. Kariya, L. Finkelstein (Eds.), Measurement Science: a
Discussion, Ohmsha Press, Amsterdam, 2000.
[28] T. Preston, in: The Theory of Heat, Macmillan, London,
1894.
[29] L. Michalski, K. Eckersdorf, J. McGhee, in: Temperature
Measurement, Wiley, Chichester, 1991.
[30] R.J. Sternberg, Intelligence, in: R.I. Gregory (Ed.), The
Oxford Companion to the Mind, OUP, Oxford, 1987, pp.
375–379.
[31] A.W. Heim, Intelligence: its assessment, in: R.I. Gregory
(Ed.), The Oxford Companion to the Mind, OUP, Oxford,
1987, pp. 379–381.
[32] D.T.J. Laming, Psychophysics, in: R.I. Gregory (Ed.), The
Oxford Companion to the Mind, OUP, Oxford, 1987, pp.
655–657.
[33] D. Crystal, in: The Cambridge Encyclopedia of Language,
Cambridge University Press, Cambridge, 1987.
[34] R.R. Sokal, P.H.A. Sneath, in: Principles of Numerical
Taxonomy, W.H. Freeman, San Francisco, CA, and London,
1963.
[35] UK Office for National Statistics, UK Standard Industrial
Classification of Economic Activities, HMSO, London,
1992.
[36] C.I. Chase, in: Measurement for Educational Evaluation,
Addison Wesley, Reading, MA, 1974.
[37] G. Leech, in: Semantics, Penguin Books, Harmondsworth,
1981.