Brown Incom Reconsidered

7/23/2019 Brown Incom Reconsidered

1/21

Incommensurability reconsidered

Harold I. Brown

Department of Philosophy, Northern Illinois University, DeKalb, IL 60115, USA

Received 28 April 2004; received in revised form 19 July 2004

Abstract

In his later writings Kuhn reconsidered his earlier account of incommensurability, clarify-

ing some aspects, modifying others, and explicitly rejecting some of his earlier claims. In

Kuhns new account incommensurability does not pose a problem for the rational evaluation

of competing scientific theories, but does pose a problem for certain forms of realism. Kuhn

maintains that, because of incommensurability, the notion that science might seek to learn thenature of things as they are in themselves is incoherent. I develop Kuhn s new account of

incommensurability, respond to his anti-realist argument, and sketch a form of realism in

which the realist aim is a pursuable goal.

2005 Elsevier Ltd. All rights reserved.

Keywords: Incommensurability; Thomas Kuhn; Rationality; Skill; Theory choice; Translation.

1. Background

The incommensurability rubric was introduced by Kuhn and Feyerabend in 1962;

claims associated with this rubric have been subjected to many interpretations, ref-

utations, and defenses. (For recent discussions and an extensive bibliography see

Hoyningen-Huene & Sankey, 2001.) In the years that followed Kuhn returned to this

notion many times, seeking to clarifyand sometimes modifyhis account. In his

last papers, he introduced some new ideas that seem to involve significant changes

0039-3681/$ - see front matter 2005 Elsevier Ltd. All rights reserved.

doi:10.1016/j.shpsa.2004.12.008

E-mail address: [email protected](H.I. Brown).

Stud. Hist. Phil. Sci. 36 (2005) 149169

Studies in Historyand Philosophyof Science

www.elsevier.com/locate/shpsa
mailto:[email protected]:[email protected]


2/21

in the import of incommensurability. In this paper I will explore Kuhns modifica-

tions, defend some of them, and assess the status of incommensurability as a

meta-concept for understanding the development of science.1

Kuhn introduced his original account of incommensurability in a specific historicalcontext, and in opposition to the prevailing logical empiricist view of science. To

understand Kuhns idea, we must be clear on the aspects of logical empiricism he

was opposing. Consider, first, the theory of meaning that was central to logical empir-

icism. According to this theory, we have a basic vocabulary made up of terms that de-

rive their meaning directly from experience. This vocabulary, dubbed theobservation

language, could be shared by all human beings with normal sense organs (variations are

a matter of the particular experiences one has had), and the meanings of its terms are

established independently of any of our beliefsand a fortioriindependently of any

scientific theories. Different natural languages associate different phonemes and gra-

phemes with experienced items, but terms that are associated with qualitatively iden-tical bits of experience have the same meaning. Thus, the terms of the observation

language are precisely and mechanically translatable among all natural languages.

We also have a large body of auxiliary terms that are introduced for convenience,

and ultimately derive their meaning from the observation language. However, the ex-

act relation between the observation language and the auxiliary language was a matter

of some debate. Ideally, all auxiliary terms would be translatable into the observation

language, so that all meaningful discourse would be expressible in this language.

Unfortunately, certain key terms resisted translation. In philosophy of science theoret-

ical terms constituted the most important class of resistant terms; as logical empiricism

developed there were several attempts to assimilate theoretical terms to this frame-

work.2 By 1962 there was a widely accepted account of theoretical terms that had been

introduced in 1920 by Campbell (1957)and independently rediscovered by Carnap

(1956). This view held that a theory consists of a set of axioms containing the theoret-

ical terms plus a set of correspondence rules that relate the theoretical terms (more or

less directly) to the observation language. There were residual disputes about the exact

nature of these correspondence rules, and about whether the relations between theoret-

ical terms established in the axioms contribute to the meaning of these terms. But it was

generally agreed that the empirical import of a theory derives from its relation to the

observation language. Once this relation is established there is an upward seepageof meaning from the observational terms to the theoretical concepts (Feigl, 1970, p.7).

The observation language plays a second role for logical empiricists, in addition

to its role in semantics: it provides the basis for comparing and choosing between

competing theories. This follows directly from the thesis that all empirically signifi-

1 I will follow Kuhns practice of referring toThe structure of scientific revolutions (1962) as Structure;

Kuhns later papers are collected in The road since Structure,Kuhn(2000); I will cite specific papers from

this collection, and give the original publication date in square brackets.2 For reviews of these attempts seeBrown (1979); Feigl (1970); Hempel (1965, 1970); Scheffler (1963).

Quines (1953) attack on the analytic/synthetic distinction provided an additional challenge to this

programsince translations were supposed to be expressed by analytic propositionsand provided part

of the background of Kuhns challenge to logical empiricism (1962, p. viii).

150 H.I. Brown / Stud. Hist. Phil. Sci. 36 (2005) 149169


3/21

cant claims are expressible in the observation language. As a result, if two theories

are genuine competitors, their disagreements will amount to alternative claims about

what we will observe under specified conditions. These disagreements are completely

expressible in the observation language, and decidable (except for possible pragmaticdifficulties) by appropriate observations.

There is one more issue that was not actually discussed by logical empiricists be-

fore the challenges of 1962, but that is easily solved given their theory of meaning.

Note, first, that empirical generalizationsboth in science and everyday lifeare

typically expressed in the auxiliary language. At this level logical empiricism is com-

patible with the existence of competing generalizations that make use of radically dif-

ferent concepts. Indeed, from a logical empiricist perspective a major part of the

creative challenge of scientific research is finding auxiliary conceptsthat is, ways

of classifying observablesthat yield reliable predictions. Historically, the develop-

ment of science has often required replacement of those classifications that first leaptto inquiring minds. Familiar examples include the classification of all solids as in-

stances of earth, and all liquids as instances of water; treating hot and cold as onto-

logically equivalent contraries; and treating rest and uniform motion as different

dynamical states. I have already noted the logical empiricist account of how the con-

sequences of such competing generalizations can be compared. The additional issue I

want to introduce concerns the resources that allow someone who has learned to

think about some domain using a particular set of auxiliary concepts to make the

transition to thinking in terms of a very different set. The logical empiricist answer

is straightforward, at least in principle: empirically significant differences between

two generalizations can be expressed in the observation language, and one can move

from one set of auxiliary concepts to another by making the appropriate translations

into this language. A parallel account holds for the ability to learn new theoretical

terms.

Kuhn challenged this entire framework by challenging the existence of an obser-

vation language and the theory of meaning built on it. In effect, his alternative ac-

count of meaning begins from the picture of a theory as an axiom system

connected to observation by correspondence rules, but reverses the direction in

which meaning flows. Focusing on theoretical terms, Kuhn maintained that their

meaning is determined by their interrelations. Moreover, if observation is to playa role in evaluating a scientific theory, descriptions of what we observe must be ex-

pressed in the language of that theory. As a result, different theories that deploy dif-

ferent concepts will yield different descriptions of a single body of observations. In

this sense, at least, observation language is theory-laden. It is important to be clear

that the observation language, rather than perceptual events in someones psyche, is

the important factor in this context. This is because logical empiricists treated con-

firmation (and disconfirmation) as logical relations, and logical relations hold be-

tween statements (or propositions if you prefer, this is a metaphysical dispute that

we need not consider here).3 To my knowledge, Kuhn never challenged this aspect

3 The key role of observation statements, rather than observations, was especially emphasized by

Popper in his 1934 book Logic der Forschung(seePopper, 1992).

H.I. Brown / Stud. Hist. Phil. Sci. 36 (2005) 149169 151


4/21

of logical empiricism. As a result, even if perception were to yield experiences that

are independent of our beliefs, these experiences do not become relevant to theory

evaluation until they have been described in the language of that theory.4 This shift

undermines the logical empiricist account of theory comparison since it eliminatesany theory-neutral observational evidence. In effect, each theory has its own empir-

ical data. Note that holding observation to be theory-laden in this sense does not im-

ply that every observation will support a theory. Interpreting data in terms of a

particular theorys concepts is quite compatible with recognizing that the data con-

tradicts specific theory-generated expectations. Indeed, this is the source of the

empirical anomalies that, for Kuhn, provide one major locus of normal-scientific re-

search. But the lack of a common observation language would seem to block com-

parisons of theories that describe observations using different concepts. This yields

one of two incommensurability problems that Kuhn originally raised: the absence

of some common unit for comparing theories that deploy different concepts. Atthe time he wrote Structure Kuhn agreed with logical empiricists that full objective

comparisons of competing fundamental theories requires a common language in

which the consequences of the competitors can be expressed; they disagreed on the

availability of this language. We will see in the next section that while Kuhn contin-

ues to stress this form of incommensurability in his last papers, and continues to con-

sider it important, he reconsiders its significance for the problem of theory

comparison.

A second incommensurability problem derives from another Kuhnian thesis: that

observation and logic are not sufficient to resolve disputes between fundamental the-

ories (1962, p. 93;Kuhn2000c [1997], p. 204); additional methodological criteria are

required, and these additional criteria are internal to specific theories. Familiar

examples that Kuhn discusses include the requirement in ancient astronomy that

celestial motions be circular; Newtons gravitation law and the requirement that

acceleration be proportional to force in, say, the nineteenth century; and conserva-

tion of energy.5 The key point is that we do not have a theory-neutral set of meth-

odological rules that will allow us to compare competing theories. Presumably, this

problem would remain even if we had a relevant body of theory-neutral data. Kuhn

does not address this form of incommensurability in his late papers, but we will see

that his revised approach to incommensurability has consequences for this problem.In addition to the two problems just noted, incommensurability provides Kuhn

with the basis for an attack on scientific realism.6 The term scientific realismis used

to describe many claims; here I am interested only in the thesis that the history of

4 InBrown (1995)I distinguished six different senses in which it has been claimed that observation is

theory-laden, and explored the sustainability and significance of each. The deepest of these is the claim that

our theories infect what we perceive, and Kuhn adopts this view in several passages. I have challenged the

relevance of this claim for scientific research in Brown (2005). For present purposes only the version

introduced in the main text is relevant.5 This form of incommensurability was stressed inDoppelt (1978).Examples of recent discussions will

be found inBird (2002)and Brown (1996).6 This is not a point on which Kuhn differs from logical empiricists since they were generally anti-

realists as well.



5/21

science exhibits a progressively improving set of approximations to a correct descrip-

tion of reality, as it exists independently of our beliefs. The two forms of incommen-

surability considered above seem to undermine this thesis because they allow for the

possibility that successive theories in a field are just different, without any reason tothink that the later theory is more nearly accurate. This line of argument will be

sharpened by Kuhns reconsideration of the role of incommensurability in theory

choice.

2. Reconsideration

I begin my account of Kuhns reconsideration with his clear rejection of an impor-

tant feature of his earlier practice. For a long time Kuhn approached conceptual

incommensurability in terms of his initial encounter with the phenomenon: his first at-tempt to understand Aristotelian physics from the perspective of his own training in

Newtonian physics. In one report of this experience (Kuhn, 2000h[1987], pp. 15

20) he prefaces the account with this remark: The road I traveled backward with

the aid of written texts was, I shall simply assert, nearly enough the same one that ear-

lier scientists had traveled forward with no text but nature to guide them (p. 15). In

Structure Kuhn placed much emphasis on a comparison of the conceptual frameworks

of Aristotelian and Newtonian physics. But for the scientists involved these were never

genuine competitorsa point that Kuhn acknowledged soon after the remark just

quoted: In recent years I have increasingly recognized that my conception of the pro-

cess by which scientists move forward has been too closely modeled on my experience

with the process by which historians move into the past(Kuhn, 2000e[1989], p. 87).

The point and the example are sufficiently important to warrant some elaboration.

Westfall notes that when Newton came on the scene Aristotelian physics was no

longer a major player: As far as men active in the study of nature were concerned,

the word overthrown is not too strong. For them, Aristotelian philosophy was

dead beyond resurrection (Westfall, 1983, p. 14). Cartesian physics was at center

stage, and was the view Newton sought to replace in the Principia. Descartess ma-

ture physics is found in Parts II and III of his Principles of philosophy(1991). There is

no doubt that Newton was thoroughly versed in this material since he wrote a de-tailed critique of it in an unfinished manuscript generally referred to as De gravitati-

one(Newton, 1962). Although NewtonsPrincipiacontains few explicit references to

Descartes, Newton systematically argues that the vortex theory of planetary motion

is incompatible with each of Keplers laws and with the motion of comets. Much of

this argument is in Section 9 of Book II and its concluding Scholium. Newton opens

the General Scholium that appears at the end of the second and third editions of the

Principiawith a summary of the case against Descartess vortex hypothesis (Newton,

1999, pp. 939940); Cotes Preface to the second edition states the case at greater

length (ibid., pp. 393398).7 Newton returns to the critique of Cartesian physics in

7 References to Newtons Principiaare to the 1999 translation unless otherwise specified.



6/21

the Queries that he included in the final edition of his Opticks (Newton, 1952, pp.

362365, 368369, 397400).

Commentators note that Newtons full title,Philosophiae naturalis principia math-

ematica, is intended to refer to Descartess Principia philosophiae. On the title pagesof the first and second editions of Newtons Principia the words Philosophiae and

principiaare in larger type than the other two words of the title; in the third edition

these two words are also printed in red. In addition, the third edition begins with a

half-title page containing just the two words Principia philosophiae plus Newtons

name (Newton, 1972, pp. 17). At the beginning of Principia there is considerable

similarity between Newtons and Descartess language. (For discussion of this and

other connections between Newton and Descartes see Cohen, 1983, pp. 182193

and Brackenridge, 1995, pp. 1724.) For example, Newtons second definition

(D2) reads: Quantity of motion is a measure of motion that arises from the velocity

and the quantity of matter jointly. This is followed by the comment: The motionof a whole is the sum of the motions of the individual parts, and thus if a body is

twice as large as another and has equal velocity there is twice as much motion,

and if it has twice the velocity there is four times as much motion (p. 404). Verbally,

this passage could have been written by Descartes who also introduces a concept of

the quantity of motion and writes: when one part of matter moves twice as fast as

another twice as large, there is as much motion in the smaller as in the larger . . . (1991, p. 58). To be sure, Descartess concern is rather different from Newtons, since

Descartes is here discussing the total quantity of motion in the universe which, he

maintains, is constant. Still, like Newton, Descartes considers this total quantity

to be the sum of the quantities of the individual parts. Thus both hold that the quan-

tity of motion in a compound body is the sum of the quantities of the constituent

parts, and that this quantity is measured by the bodys speed multiplied by some

measure of the amount of matter in the body. However, Newton and Descartes have

different measures of this amount of matter. Although Descartes never explicitly

states it, his identification of matter with extension makes it is clear that his measure

of quantity of matter must be the bodys volume.8 Newtons first definition reads:

Quantity of matter is a measure of matter that arises from its density and volume

jointly (p. 403), which is a considerably different conception from Descartess. But

Newton eases us into his new framework by initially providing text that will be famil-iar to readers of Descartes. I want to consider some further near parallels, but must

first introduce a bit more of Descartess physics.

Unlike Newton, Descartes was seeking to replace Aristotelian physics. One of

Descartess major departures from the tradition he inherited is a change in the exten-

sion of the concept of a state. Descartes also endeavors to ease his readers into his

new account. Like his predecessors, Descartes views a state as a property (i.e., a

mode) of a body that does not change spontaneously. It has long been recognized,

Descartes notes, that objects do not change their shape or begin to move without

an external cause; shape and rest serve as paradigm examples of states. Descartes

8 This is especially clear in his account of condensation and rarefaction ( Descartes, 1991, p. 4142, 48).



7/21

now claims that uniform motionmotion in a straight line at constant speedis

also a state. The idea that a form of motion might continue spontaneously was

not completely new since in ancient and medieval astronomy the eternal circular mo-

tion postulated in the heavens exhibited this characteralthough in this case noforces exist that could change the motion. Descartes, however, explicitly denies that

circular motion is a state (1991, p. 60). In the terrestrial realm, the Aristotelian nat-

ural motionswould not count as Cartesian states since these motions spontaneously

cease when the moving object reaches its natural place. So Descartes is making sig-

nificant changes in what counts as a state, but working in terms of notions that

would have been familiar to Aristotelians. They might disagree, but there is no rea-

son to think they would not understand.

Newton built on this extended notion of a state. Verbally, Newton s third and

fourth definitions and first law of motion could have come from Descartes s pen.

D3. Inherent force of matter is the power of resisting by which every body, so far

as it is able, perseveres in its state of either of resting or of moving uniformly

straight forward. (p. 404)

D4. Impressed force is the action exerted on a body to change its state either of

resting or of moving uniformly straight forward. (p. 405)

L1. Every body perseveres in its state of being at rest or moving uniformly

straight forward, except in so far as it is compelled to change its state by forces

impressed. (p. 416)

L1 should also be compared with a comment Descartes makes on his own first law.

Descartess law reads: each thing, provided that it is simple and undivided, always

remains in the same state as far as is in its power, and never changes except by exter-

nal causes(1991, p. 59). Descartes then adds that if such a body is at rest, we do not

believe that it will ever begin to move unless driven to do so by some external cause.

Nor, if it is moving, is there any reason to think that it will ever cease to move of its

own accord and without some other thing which impedes it (ibid.).

However, in spite of this verbal similarity Newton makes major conceptual

departures from Descartes. For example, although Descartes considers both rest

and uniform motion to be states, he views them as fundamentally different states.This is clear, for example, in Descartess rules of impact (ibid., p. 6469) where

cases in which one body is at rest before impact are treated differently from cases

in which both bodies are initially in motion.9 Newton considered rest and uniform

motion to be the same state; rest is just motion with a speed of zero. Thus, for

Newton, when a moving object reverses direction it passes momentarily through

a state of rest, but this involves no difference of principle from any other speed that

the object passes through. Descartes denies that a state of rest occurs in such cases

(e.g. 2001, pp. 7576). Newton also holds that change of direction and change of

speed are the same phenomenon, while for Descartes these are fundamentally

9 Descartess accounts of motion and rest are conceptual quagmires that we need not discuss for present

purposes.



8/21

different. Descartess clearest argument for this claim occurs in his Optics where he

notes that in order to change the speed of a tennis ball I must change the force

with which it is hit, while to change its direction I need change only the angle

of the racket (ibid.). Descartess distinctions between rest and motion, and betweenchange of direction and change of speed are combined in his rules of impact which

include cases in which (as he describes them) there is change of direction with no

change in quantity of motion.10 The upshot, then, is that there are deep conceptual

differences between Newtonian and Cartesian physics, so that the import of the

verbally similar claims I have noted is quite different in the two systems. Still,

the language Newton uses at the beginning of his exposition provides a bridge be-

tween Descartess physics and his own.

It is worth noting that Ptolemaic astronomy was also dead by Newton s day. To

the extent that Copernicanism had a competitor, it was Tycho Brahes system in

which the traditional planets move around the sun while that entire system movesaround the earth. The status of Brahes account is reflected in an interesting way at

the beginning ofPrincipia Book III where Newton applies the mathematical results

of the two prior books to the planetary system. Early in Book III Newton intro-

duces six phenomenaempirical generalizations about major constituents of the

solar system that will provide the basis for his account. Newton is temporarily

agnostic about the motion of the earth. Phenomenon 4 reads: The periodic times

of the five primary planets and of either the sun about the earth or earth about

the sunthe fixed stars being at restare as the 3/2-powers of their mean distances

from the sun(p. 800, italics added).11 The phenomena that precede and follow this

statement make it clear that he has Brahe s alternative in mind. Phenomenon 3

states that the five primary planets encircle the sun, while Phenomenon 5 tells us

that radii from these planets to the sun pass through equal areas in equal times,

while this does not hold for radii from the planets to the earth. Newtons resolution

of the questiongiven in Book III, Proposition 12 and its corollaryis not quite

what any of his predecessors expected. Newton introduces the hypothesis (p. 816)

that the center of the world is at rest. But the common center of gravity of the

earth, sun, and planets is at rest (Book III, Proposition 11), and The sun is engaged

in continual motion but never recedes far from the common center of gravity of all the

planets (p. 816). This common center of gravity is to be considered the center of

the universe(p. 817). Since the sun, unlike the planets, is always close to this center,

treating the sun as stationary is often a good approximation. Thus, Newton rejects

a central assumption of all previous astronomy: that the center of rotation must be

occupied by a specific body. This is a revolutionary move, although we are brought

to it in a way that should not be at all difficult for astronomers of the time to

understand.

10 There are considerable further complications in Descartess account of motion that we need not

consider here. SeeGarber (1992),Chap. 68, for a detailed discussion.11 The five primary planets are Mercury, Venus, Mars, Jupiter, and Saturn. Newton sometimes call the

satellites secondary planets and sometimes uses planet to encompass both sets.



9/21

As these examples illustrate, the conceptual gaps that must be closed in typical

cases of scientific theory choice are often much smaller than the gaps that an his-

torian must cross. We might, then, have a series of theory replacements running

from T1 to Tn, where T1 and Tn have little in common although there is a greatdeal in common between any two adjacent members of the series. Kuhn recog-

nizes the point. Discussing an historical narrative describing a series of changes

in scientific belief, Kuhn concludes: By the end of the narrative those changes

may be considerable, but they have occurred in small increments, each stage his-

torically situated in a climate somewhat different from that of the one before(2000g [1991], p. 112). The difference between the small changes that occur in sci-

entific practice and the larger changes that we find when we look at a longer his-

torical development are central to Kuhns later views on the rationality of theory

choice and on scientific realism. For Kuhn these are distinct issuesa point

that he underlines by distinguishing between questions concerning the rationalityof belief and the rationality of change of belief. In effect, the former is the ques-

tion of realism, the latter that of theory change. I will focus first on theory

change.

In his later work, Kuhn clearly holds that in science change of belief is rational

even though incommensurability occurs. For example, after reviewing the tradi-

tional demand for observations that are neutral with respect to all beliefs, Kuhn

writes:

From the historical perspective, however, where change of belief is what s at

issue, the rationality of the conclusions requires only that the observationsinvoked be neutral for, or shared by, the members of the group making the

decision, and for them only at the time the decision is being made. By the same

token, the observations involved need no longer be independent of all prior

beliefs, but only of those that would be modified as a result of the change.

The very large body of beliefs unaffected by the change provides a basis on

which discussion of the desirability of change can rest. It is simply irrelevant

that some or all of those beliefs may be set aside at some future time. To

provide a basis for rational discussion they, like the observations the discus-

sion invokes, need only be shared by the discussants. (Kuhn, 2000g [1991],

p. 113)

Kuhn still maintains that there is incommensurability between the competing views

involved in such local changes, but this is because he holds that incommensurability

is just untranslatability, where translation is to be understood as a quasi-mechanical

activity governed in full by a manual which specifies, as a function of context, which

string in one language may, salva veritate, be substituted for a string in another lan-

guage(Kuhn, 2000e[1989], p. 60). As a result, Incommensurability thus becomes a

sort of untranslatability, localized to one area or another in which two lexical taxo-

nomies differ (Kuhn, 2000f[1991], p. 93). But, Kuhn now insists, translation is not

required either for communication or for rational theory choice. A different cogni-tive processwhich he describes as interpretation and as language learning is

required.



10/21

Translation is, of course, only the first resort of those who seek comprehension.

Communication can be established in its absence. But where translation is not

feasible, the very different processes of interpretation and language acquisition

are required. These processes are not arcane. Historians, anthropologists, andperhaps small children engage in them every day. (Kuhn, 2000b[1983], p. 53;

cf. 2000a [1993], p. 238)

A few years later, discussing historians, with their frequent need to cross larger con-

ceptual gaps than that of scientists involved in actual theory choice, Kuhn writes:

Faced with untranslatable statements, the historian becomes bilingual, first

learning the lexicon required to frame the problematic statements and then,

if it seems relevant, comparing the whole older system (a lexicon plus the sci-

ence developed with it) to the system in current use. Most of the terms used in

either system will be shared by both, and most of these shared terms occupy

the same positions in both lexicons. Comparisons made using those terms

alone ordinarily provide a sufficient basis for judgment. (Kuhn, 2000e[1989],

p. 77)12

Kuhn also tells us that anything which can be said in one language can, with

imagination and effort, be understoodby a speaker of another. What is prerequisite

to such understanding, however, is not translation but language learning (ibid., p.

61). And, with sufficient patience and effort, [one can] discover the categories of

another culture or of an earlier stage of ones own (Kuhn, 2000d [1991], p.

220). At this point it looks as ifincommensurability is irrelevant for questions of the-ory choice.

Kuhn also backs off from his metaphor of a scientific revolution as a gestalt shift

(although this may still be an appropriate analogy for particular historical studies).

To speak, as I repeatedly have, of a communitys undergoing a gestalt shift is to

compress an extended process into an instant, leaving no room for the micropro-

cesses by which the change is achieved (Kuhn, 2000e [1989], p. 88). Indeed, Kuhn

even clams that the possibility of significant comparisons of competing modes of sci-

entific practice was never for me in question (Kuhn, 2000b[1983], p. 55). Leaving

aside considerations of whether Kuhn is correctly reporting his earlier views, once

we acknowledge the possibility of mutual understanding, there is no residual prob-lem of the rationality of theory comparison. Moreover, this applies not only to the

use of different concepts, but also to differences in evaluation standards and in con-

ceptualization of the data. In all these cases, there may be genuine disagreements

disagreements that are more severe than those acknowledged by logical empiricists

but there is no reason why failures of rational discussion need occur. To be sure, such

12 At his death Kuhn was working on a new systematic account of scientific change in which the notion

of a lexicon appears to be the centerpiece (see, e.g., 2000a [1993], p. 239). Perhaps the manuscript will be

edited and published at some point. This concept is often mentioned in Kuhn s last papers, but there is no

published systematic account. While the term appears in some passages that I quote, nothing in the present

discussion turns on its exact characterization.



11/21

failures may occur among those who do not approach the problem with sufficient

effort, patience, and imagination, but there is no problem of theory comparison that

transcends rational mediation. Such mediation is possible because scientists are not

limited to thinking in the language of a single theory. Given the ability to understandincommensurable theories, and the occurrence of these theories in a wider shared lin-

guistic framework, they can derive results that are expressible either in this common

language or in the language of each theory. Let us consider some examples of such

mediation.

Innovators and early adopters of a new framework are often masters of the pre-

vious view and thus able to find means of catching the attention and interest of those

they would convert. Galileos use of the rock dropped onto a moving ship provides

an example. Once the test is done, there is no particular difficulty in seeing where the

rock lands. And while die-hard Aristotelians may attempt to explain away the result,

Galileos correct prediction could also provide an opportunity for an opponent towonder how Galileo had arrived at his result, and undertake to learn his approach.

Note also that one of Galileos major projects in his Dialogue is to show that Aris-

totelian arguments against the motion of the earth commit specific logical fallacies

that would be familiar to Aristotelians. (Finocchiaro, 1980,provides detailed analy-

ses of many of these arguments). Another example is provided by the shape of the

planetary orbits for Descartes and Newton. Since they both required that their the-

ories explain these shapes, Newton could argue that his theory gets them right while

Descartess theory cannot do this. In a similar way, in presenting special relativity

Einstein starts off from two well known problems: a problem in the interpretation

of Maxwellian electrodynamics, and a problem of consistency between two postu-

lates that others had already found attractive. He then resolves these problems while

working within the established mode of mathematical physics, and in a way that pre-

serves Maxwells equations and explains why Newtonian physicswhich is super-

sededworks as well as it does. From this perspective, Kitchers (1978) account

ofreference potential, which allows for the flexible identification of some items coun-

tenanced by a later theory with items invoked by a predecessor, is one technique that

can be used by both historians and innovators to build bridges between competing

theories.

An important feature of Kuhns later approach to theory comparison is its invo-

cation of human cognitive abilitieswhich introduces scientists (in addition to ab-

stractly formulable linguistic structures) into his account of theory evaluation.

This introduction of scientists into philosophy of science was a central theme of

Structure, although it dropped into the background in much of Kuhns later work.

I want to review the role this theme played in Structure. Again, the discussion is best

set in the context of the philosophical situation in 1962.

Recall that logical empiricists drew a central distinction between context of discov-

eryand context of justification. The latter deals with logical relations between obser-

vation statements, on one hand, and those generalizations and theoretical claims that

go beyond observation statements, on the other. Logical empiricists held that thephilosophical analysis of sciencewhich includes the analysis of its epistemic sta-

tusis concerned only with logical relations. Any considerations of the psychology



12/21

of the actors in the development and acceptance of scientific claims was considered

irrelevant to epistemic evaluation and was relegated to the context of discovery. We

should be clear that the context of discovery was not rejected as unimportant, but

only as irrelevant to the task ofphilosophicalanalysis. The psychology of discovery,for example, is a legitimate field of scientific researchbut the evaluation of its re-

sults depends (it was held) on their meeting the appropriate criteria for the evalua-

tion of scientific theoriescriteria which must be established independently of any

particular scientific results.

In Structures introductory chapter Kuhn indicated that he would be challenging

the distinction between the two contexts (1962, pp. 89). This challenge is captured

especially in Kuhns thesis that observation and logic are not sufficient to account for

revolutionary theory change, and his attempt to close the gap by taking into account

aspects of the psychology of scientists and the social interactions in scientific commu-

nities.13 While many rejected this as an inappropriate intrusion of psychology andsociology into epistemology,14 Kuhns move is more accurately interpreted as a pro-

posal to rethink the epistemological relevance of psychological and social factors.15

In other words, Kuhns claim is that in order to understand and evaluate scientific

theory choice we must attend to the scientific processas well as the scientificproduct.

The relevant aspects of the scientific process are not passing quirks of the scientists

involved, but rather the skills that scientists develop through their training and

continuing scientific work. Skills are lodged in individual scientists, but are no more

subjectivein a pejorative sense than is the ability to drive a car. These skills include

the cognitive resources that Kuhn is invoking when he writes, in his later papers, of

the need forand availability ofpatience and imagination in understanding a

competing theory. The point, then, is that human theory evaluation is dependent

on human psychology and we cannot give an adequate account of this process with-

out taking human psychology into account. Note especially that this dependence on

our psychology is not just a limiting constraint on the prospects of human knowl-

edgeit is also a feature that enables the development of knowledge. Our ability

to respond with intelligence and sensitivity overcomes the gaps left by failures of

translation which are inevitable given that early conceptualizations are often quite

inadequate, and that scientific progress requires both the introduction of new con-

cepts and elimination of older concepts which no longer have a role to play in theresearchers repertoire.

We are, however, not completely finished with incommensurability. In Kuhns

late works the impact of incommensurability appears in the evaluation of beliefs

that is, the question of realism. If, as Kuhn maintains, the development of science

13 Kuhn discusses two situations in which logic and observation are not sufficient for theory choice. One

occurs in normal science, where shared principles close the gap. The other more severe case occurs in

revolutionary situations where these principles are among the items being challenged.14 For example: On my first reading of Thomas S. KuhnsThe Structure of Scientific Revolutions(1962)

I was so deeply shocked at his repudiation of the distinction between the context of discovery and the

context of justification that I put the book down without finishing it (Salmon, 1991, p. 325).15 In this regard, Structure is continuous with the naturalistic approach to epistemology that was

emerging at that time.



13/21

requires the introduction of new concepts that are not translatable into existing con-

cepts, then it seems impossible to assess whether successive frameworks are moving

closer to a correct description of items in their domain. Commenting on the question

of sciences zeroing in on, getting closer and closer to, the truth, Kuhn contends thatsuch claims are meaningless, and that this is a consequence of incommensurability(2000a [1993], pp. 243244). Elsewhere Kuhn writes: I am not suggesting, let me

emphasize, that there is a reality which science fails to get at. My point is rather that

no sense can be made of the notion of reality as it has ordinarily functioned in phi-

losophy of science(2000g [1991], p. 115). In what I take to be his clearest statement

of the basis for this position, Kuhn begins with the untranslatability of a lexicon into

its successor. As a result, the earlier statements are immune to an evaluation con-

ducted with [the later] conceptual categories. But, he adds, The immunity of such

statements is, of course, only to being judged one at a time, labeled individually with

truth values or some other index of epistemic status. Another sort of judgment ispossible, and in scientific development something very like it repeatedly occurs(Kuhn, 2000e [1989], p. 76). The passage quoted earlier on becoming bilingual fol-

lows, and Kuhn continues:

But what is then being judged is the relative success of two whole systems in

pursuing an almost stable set of scientific goals, a very different matter from

the evaluation of individual statements within a given system.

Evaluation of a statements truth values [sic] is, in short, an activity that can be

conducted only with a lexicon already in place, and its outcome depends upon

that lexicon. If, as standard forms of realism suppose, a statement s being trueor false depends simply on whether or not it corresponds to the real world

independent of time, language, and culturethen the world itself must be

somehow lexicon dependent. Whatever form that dependence takes, it poses

problems for a realist perspective, problems that I take to be both genuine

and urgent. (Ibid., p. 77)

Although Kuhns argument turns on the supposed impossibility of evaluating indi-

vidual statements in a theory, he does not invoke the DuhemQuine thesis here.

Rather, his argument depends on his view that the meaning of scientific concepts

is determined by their relations to other concepts. Kuhn treats the view that meaningis dependent on relations among scientific terms as itself incompatible with realism.

I suggest that, with the help of some ambiguity in the notion of evaluation being

dependenton the lexicon, Kuhn has confused three different issues. First, there is the

claim that theconceptual contentof the claims of a theory is dependent on the system

of concepts in which it occurs. Second, the DuhemQuine thesis supports the view

that only entire theoretical systems are subject to epistemicin particular, empiri-

calevaluation. But neither of these speak to the thirdissue: what itmeans to attri-

bute truth values to individual statements in a system. Suppose we have a scientific

theory that embodies a set of interrelated concepts, and that the available evidence

supports this theory. It makes good sense not only to accept the entire theory, butalso to hold that each of the sentences constituting the theory is truewhere this

means that each sentence correctly describes the items it speaks about. This is just



14/21

a different question from that of how the meaning of these sentences is determined.

What we must avoid is the all-too-common confusion between the meaning of a

claim, the evidence for (or against) it, and what it means to say that a claim is true.

Although the evidence may support only the theory as a whole, this does not blockattribution of truth to the individual sentences of that theory. Moreover, if a theory

is replaced, a bilingual historian or scientist aware of the new evidence, canthink-

ing in terms of the older theoryconclude that certain claims in the theory are false,

and explore which claims carry over to the new theory (recall Kuhns remark, cited

above, that most of these claims will be retained), or have close successors in that

new theory.

Two further points are worth making. First, there is no reason why realism

understood as a quest for the correct account of things-in-themselvesneed be tied

to the view that scientific knowledge is apportioned to individual sentences. There is

no bar to a version of realism which holds that a theoretical system is the minimumunit of correspondence. Thus even if one rejects attribution of truth-values to indi-

vidual sentences, a robust form of scientific realism remains a possibility. Second,

Kuhns argument from incommensurability does provide grounds for rejecting the

view that science pursues such accounts in a linearly progressive fashion. But this

is not the same as rejecting the claim that science pursues correct accounts, nor does

it eliminate all grounds for thinking that, as science develops, our ability to pursue

this goal improves. I want to consider such an alternative approach.

We must not forget the central role that empirical evidence plays in driving

research (cf.Brown, 1990, 1995, 2001). This occurs in two respects. First, much re-

search is directly elicited by experience. In recent decades many have argued for a

central role of theory in driving research. While I think this is basically correct,

the point was often overstated because it emerged as part of a critique of logical

empiricism which focused mainly on the empirical side of science, and gave theory

only a secondary role. By now we can see that experience and theory are more nearly

equal partners in generating scientific problems. Consider a well worked example: at

the beginning of planetary astronomy the wandering motions of the planets need not

have seemed problematic; they could have just been listed among the observed facts.

It required the hypothesis that all true planetary motions are circular to generate the-

oretical research. But let us not forget the other side. The belief in circular motionwould not have generated a research problem without the observation of celestial

items that appeared to violate this hypothesis. The point is especially dramatic when

nature impinges on researchers in unexpected wayssuch as in the initial observa-

tions of sperm, X rays, and radioactivity.16 To be sure, none of these phenomena

16 The case of sperm is less well known than the other two. Sperm were discovered in seminal fluid in the

1670s by the early microscopists van Leeuwenhoek and Hartsoeker. Initially it was unclear what role, if

any, they played in reproduction (Farley, 1981; Gasking, 1967, p. 54). For a substantial period after their

discovery many naturalists believed that sperm were parasites of the testes that had no reproductive

functiona view that survived into the nineteenth century (Farley, 1982, pp. 4347). The exact role that

sperm played remained a subject of dispute for some 200 years (see Farley, 1982 for an extended

discussion).



15/21

would have seemed surprising without some theoretical background that indicated

what to expect, but these new lines of research were initiated by the new observa-

tions. In more or less dramatic ways, this same interplay holds throughout the his-

tory of scientific innovation. One side or the other may dominate in a particular case,but both are required to generate new research.

But while the theoretical and experiential sides may be roughly equal in generat-

ing new research, the decision to accept a theory depends ultimately on its ability to

handle the results of our interactions with nature. This is thesecondrole of evidence

noted above: it is the final arbiter of scientific acceptability. Sometimes, in a well

developed science, a piece of research may be primarily driven by theoretical con-

siderations. Diracs determination to construct a relativistically correct quantum

theory that uses only first derivatives is as clear an example of successful theory-dri-

ven science as we are likely to find. The quest carried him through a significant

mathematical innovationintroduction of matrices where previously standardpractice would require numbers or vectorsand the introduction of a new con-

ceptantimatter. But the work also had empirical consequencessome already

known, some new. It is only because of its empirical successes that the theory pre-

vailed. In the face of empirical failures, any theoretical principlecircular celestial

motions, conservation of energy, direct proportionality of force and acceleration,

stability of species, the total separation between space and timecan be reconsid-

ered and replaced no matter how well founded it may once have seemed in experi-

ence and reason.

Now, a proper understanding of empirical evidence in science requires another

break with the classical empiricist traditionone that, I think, Kuhn never made.

The epistemic significance of empirical evidence doesnot derive from its dependence

on our senses. Rather, we pursue evidence pertaining to presumed items in the world

by attempting to interact with those items.17 We evaluate claims about items in a do-

main by attempting to probe them in various ways, and the greater the variety of

probes at our disposal, the richer the body of evidence we have for these claims.

The development of instrumentationbeginning with Galileos use of the telescope

and exploding in the twentieth centuryhas greatly increased the variety of ways in

which we probe nature; it has also increased the precision of the results of these

probes.18

On this view our senses play a pragmaticnot a foundationalrole ingathering empirical evidence: our senses are the means by which information about

items in the world enters into our cognitive systems. But the information does not

reduce to the sensationstheory-laden or notthat provide this access.

17 Kuhn approaches this point when he writes that a science students world is determined jointly by the

environment and the particular normal-scientific tradition that the student has been trained to pursue

(1962, p. 112), and by his frequent invocation of the role of nature in producing anomalies. But the

emphasis in these passages is usually on the role of the tradition, and Kuhn does not systematically pursue

the role of items that are independent of our beliefs in scientific research.18 As one indicator of this explosion, consider that we no longer just have telescopes that gather light. In

addition to optical telescopes, we have radio, infra-red, ultraviolet, X-ray, and neutrino telescopes. For

extended discussions of this view of evidence see Brown (1987, 1995, 2001, 2005).



16/21

From this perspective, the process of conceptual change in science is, in part, a

process by which we seek concepts that describe items that exist independently of

our theories. These items form the subject matter of our theorizing; they are the

items we seek to learn about when we construct and test theories. Sometimeswe suc-ceed in picking out items (such as Mars) or kinds of items (such as gold) that provide

a stable focus for research as our ideas about them change. Even if the concepts we

use to think about these items undergo radical change, it is clear that we are discuss-

ing the same things throughout, and no fundamental problem arises about compar-

ing successive theories because they are all clearly accounts of the same items. But

this type of research is not the onlyor even the dominantmode of scientific re-

search. It is not the kind of research that led to quarks, gluons, and weak-interaction

bosons; nor is it the kind of research that drives the search for the Higgs boson, or

Newtons identification of change of speed and change of direction as instances of

the same phenomenon, or the (limited) unification of space and time in special rela-tivity. Nor does it apply to cases such as the unexpected darkening of Becquerls pho-

tographic plate that led to the discovery of radioactivity. The last example suggests a

familiar variation on the notion that scientists identify and study stable referents:

that one introduces a new item as whatever caused an observed phenomenon, and

that research proceeds to seek out that cause. This is another kind of research that

does occur, but there is a great deal of research that does not fit this pattern either.

For example, the approach gives no insight into the development from the Aristote-

lian search for the cause of the continued motion of projectiles to the (rather differ-

ent) Cartesian and Newtonian conclusions that there is no cause. Nor does it help us

understand the introduction of isospin and its role in modern accounts of the stabil-

ity of atomic nuclei, or into the role of the weak interaction in understanding why

some nuclei undergo radioactive decay. Nor do either of these approaches give

any insight into the myriad cases in which classifications of items as the same kindare changed. Often cited examples include the changing classifications of the earth

and planets, and of the sun and stars, as well as the fate of the Aristotelian ele-

mentsnone of which are to be found on modern lists of the chemical elements.

Nor are any of the modern chemical elements on this ancient listseven though

some of these elements were familiar to the ancients.19 Another, more subtle type

of fundamental reclassification is exemplified by the shift from viewing rest and uni-form motion as different kinds to considering them as instances of the same kind,

and the similar shift with regard to change of speed and change of direction. Mean-

while, research into the structure of atoms has continued to multiply fundamental

kinds, moving from a compact account in terms of just electrons, protons, and neu-

trons in the early 1930s to the standard model which includes six leptons, six

quarks, 12 field bosons, the (as yet undetected) graviton and Higgs boson, plus an

anti-particle for each. That is 50 kinds of entities47 if neutrinos are their own

anti-particles. Theories that go beyond the standard model add additional kinds

of entities. Most of theseas well as some of the key properties by which they are

19 An ancient Chinese classification of elements into air, water, earth, metal, and wood suffered the same

fate (Leicester, 1971, pp. 5355;Brock, 1993, p. 6).



17/21

characterized and differentiatedwere not conceived of in, say, 1900. In some cases

the question whether two items are instances of the same kind is downright mislead-

ing since what counts as the same kindvaries with the research context. Protons and

neutrons are the same with respect to the strong interaction, but not with respect tothe electromagnetic interaction. Isotopes of an element share some properties, but

not others; thus most chemical elements are clusters of a number of different kinds.

Thorium, for example, has more than 25 known radiactive isotopes with half-lives

running from microseconds to billions of years; H2O in which the hydrogen is

deuterium is toxic. The list expands when we consider isomers of a compound and

ionization states of an atom.

We must recognize that there are multiple forms of scientific research and theory

change, with different features providing continuity through different changes. This

is quite in accord with Kuhns late view of the microstructure of scientific change.

Thus we return to the question whether we have any good reasons for believing thatwe are making progress towards the correct description of items in the world as they

are apart from our theorizing. Given the scope of the conceptual changes that occur

as science develops, an argument on behalf of such progress must allow for a highly

non-linear approach in which we may be on the wrong track for substantial periods

of timeperhaps for most of the history of subjectand for cases in which a new

theoretical development actually moves us further away from the description we

seek. An account of progress that is compatible with this kind of development can

be built on the above remarks about the development of instrumentation yielding

a wider variety of means of interaction with nature, along with results of much high-

er precision, than in the past. All of these interactions provide constraints on our the-

orizingconstraints that come from nature. A theory that meets contemporary

constraints has thus passed tougher tests than were available in the past, while the

range and precision of such tests continues to grow. As a result, we have a strong

reason for believing that contemporary theories provide a better account of nature

than their predecessors, even though we cannot measure how close we are. More-

over, as the process of theory testing continues, the constraints on successful theories

continue to grow.20 Note especially that the theory-dependence of observationin

the sense that observational results must be interpreted in terms of the concepts of

the theory being evaluatedsupports this point. It is the pursuit of such interpreta-tion that allows us to recognize cases in which empirical results are incompatible

with a particular theory, and to consider other theories with which they are compat-

ible. Note also how incommensurabilityunderstood as the inability to translate

newly introduced concepts into a previously available frameworkhas dropped

out of the discussion. Empirical evaluation of a theory can take place within the

framework of that theory. Failures of the theory can be recognized, and attempts

to construct or learn an alternative can begin.

20 There is a tension between this account of improving grounds for accepting theories and the

pessimistic induction from the failures of previously well supported theories. The response is to note that

the inductive evidence for many scientific theories is stronger than the inductive evidence for the

pessimistic induction. For details seeLevin (1979); Brown (1990).



18/21

We can now also get beyond two further points that have generated some confu-

sion in the literature. First, given that we need a theoretical framework to carry out

coherent research, Kuhn (1962), Lakatos (1970), and others have maintained that

theory evaluation is not just between nature and a theory, but always involvestwo competing theories. However, this thesis runs together two quite different points.

We can agree that scientists do not reject an established theoryleaving themselves

with no basis for organized researchunless they have an alternative to adopt. In

this sense, theory evaluation is comparative. But scientists do not need an alternative

theory in order to recognizethat the prevailing theory is empirically or conceptually

defective, and thus seek an alternative. The empirical failings and internal inconsis-

tency of Bohrs theory of the atom were well known, but it took some time until Hei-

senberg and Schrodinger provided an appropriate successor.

Second, the kind of incommensurability that remains at this point in our discussion

does not involve even a hint of relativism. It does involve a large dose of fallibilism:recognition that science proceeds by means of theories that are subject to reconsider-

ation and replacement by radically different theories. But, we have seen, the replace-

ment process is based on specific comparisons between theories. This does not mean

that evaluations will be simple, straightforward, or algorithmiconly that there will

be sufficient grounds for coherent debate, which may include specification of further

tests that could lead to a decision. Most importantly, as long as we are doing science,

we must accommodate theories to the results of empirical probes, and it is not the case

that any theory can be defended come what may. Note especially that those who in-

voke the DuhemQuine thesis to argue that we can protect any given thesis acknowl-

edge that once we have an incorrect prediction from a theoretical complex, something

is wrong. Thus the attempt to protect a favored thesis comes only at the cost of mak-

ing other changes in this complex. It is then an open question whether specific changes

made to protect a favored hypothesis generate further empirical anomalies (see

Greenwood, 1990andBrown, 2001for further discussion).

A key question now emerges: Do the empirical constraints on our theories ever be-

come sufficiently powerful to require acceptance of a single theory? There is no simple

answer to this question, and it may be different in different domainssuch as the

cause of polio and the fundamental constituents of the material world. Still, we must

not forget that elimination of specific theories, or classes of theories, from serious con-sideration, is an important form of progress in our knowledge of the world.

There is one more form of incommensurability that remains to be considered: the

psychological problem that arises for many people in adapting to new concepts and

evaluation criteria. Three points are worth making in this regard. First, human cog-

nitive history shows thatas a specieswe are capable of carrying out this task. To

be sure, some people are more able to make such adjustments than others, and the

number of people who introduce new ideas is considerably smaller than the number

of those who can learn them. No doubt some are left behind in the process. But

unanimous acceptance is not a requirement for genuine progress. Second, as Kuhn

has recognized, the gaps that must be crossed to introduce and learn new frame-works are considerably smaller than have sometimes been supposed. Even the tran-

sition to a strikingly new framework can result from relatively small systematic



19/21

changes in an available framework. As a result, there are conceptual bridges that can

help those who put in the required work to make the transition. Third, it is worth

repeating that innovators and early adopters of a new framework are often masters

of the previous view and thus able to find means of generating interest in the mem-bers of an existing community and constructing the required bridges.

Beyond these three observations, there remain such problems as a detailed under-

standing of how people adapt to new concepts, and why some adapt more easily than

others. More generally, there is a problem that Kuhn maintains was always his cen-

tral concern: What was and is at issue is not significant comparability but rather the

shaping of cognition by language, a point by no means epistemologically innocuous(2000b [1983], p. 55). But these are empirical questions to be pursued by the appro-

priate sciences.

3. Conclusion

Kuhns later remarks on incommensurability change the account of its signifi-

cance that is commonly attributed to him. These remarks also provide an opportu-

nity for reconsidering the significance of incommensurability for philosophy of

science. One important feature of these remarks is Kuhns return to a theme that

has pervaded his writings, but that he never fully exploited: debates between advo-

cates of competing views can draw on a wide range of human cognitive resources;

we are not limited to matching words in a translation manual. An anecdote fromoutside of science may help underline the point. When the first joint performance

of the Spanish soprano Montserrat Caballeand the American tenor Richard Tucker

was being planned, the organizers had some concerns about their ability to work to-

gether because she did not speak English and he did not speak Spanish. But she

spoke German and he spoke Yiddish, and they had no trouble communicating. Sci-

entists who share a wider cultureboth scientific and extra-scientificand who are

genuinely trying to understand each other can do as well. As a result, the introduc-

tion of new ideas into science does not undermine the process of rational assessment.

This is fortunate because once we recognize that humanity did not begin its intellec-

tual journey already possessing all the concepts and methodological tools that wouldever be required, incommensurability becomes a requirement for progress.

Incommensurability does raise a problem about certain forms of realism. I have

suggested a way of thinking about realism that builds on the empiricist tradi-

tionalthough with some important modifications. This approach provides a reason

for thinking that the realist aim is pursuable, and that it may even be achievable in,

at least, some domains.

References

Bird, A. (2002). Kuhns wrong turning. Studies in History and Philosophy of Science, 33A , 443463.

Brackenridge, J. (1995). The key to Newtons dynamics. Berkeley: University of California Press.



20/21

Brock, W. (1993). The Norton history of chemistry. New York: W.W. Norton & Co.

Brown, H. (1979).Perception, theory and commitment: The new philosophy of science. Chicago: University

of Chicago Press.

Brown, H. (1987). Observation and objectivity. New York: Oxford University Press.

Brown, H. (1990). Prospective realism. Studies in History and Philosophy of Science, 21, 211242.

Brown, H. (1995). Empirical testing. Inquiry, 38, 353399.

Brown, H. (1996). The methodological roles of theory in science. In B. Rhoads, & C. Thorn (Eds.), The

scientific basis of geomorphology (pp. 320). Sussex: Wiley.

Brown, H. (2001). Incommensurability and reality. In P. Hoyningen-Huene, & H. Sankey (Eds.),

Incommensurability and related matters (pp. 123142). Dordrecht: Kluwer Academic.

Brown, H. (2005). On the epistemology of theory-dependent evidence. In A. Raftopoulos (Ed.), Cognitive

penetrability of perception (pp. 7395). Hanppange, NY: Nova Science Publishers.

Campbell, N. (1957).Foundations of science. New York: Dover. (First published asPhysics:The elements.

Cambridge: Cambridge University Press, 1920)

Carnap, R. (1956). The methodological character of theoretical terms. In H. Feigl, & M. Scriven (Eds.),

Minnesota studies in the philosophy of science(Vol. I, pp. 3876). Minneapolis: University of MinnesotaPress.

Cohen, I. (1983). The Newtonian revolution. Cambridge: Cambridge University Press.

Descartes, R. (1991). Principles of philosophy (V. Miller and R. Miller, Trans.). Dordrecht: Kluwer

Academic. (First published 1644)

Descartes, R. (2001). Discourse on method, optics, geometry, and meteorology (revised ed.) (P. Olscamp,

Trans.). Indianapolis: Hackett. (First published 1637)

Doppelt, G. (1978). Kuhns epistemological relativism: An interpretation and defense. Inquiry, 27, 3386.

Farley, J. (1981). Sperm. In W. Bynum, E. Browne, & R. Porter (Eds.), Dictionary of the history of science

(pp. 397398). Princeton: Princeton University Press.

Farley, J. (1982). Gametes and spores: Ideas about sexual reproduction, 17501914. Baltimore: Johns

Hopkins University Press.

Feigl, H. (1970). The orthodoxview of theories. In M. Radner, & S. Winokur (Eds.),Minnesota studies inthe philosophy of science (Vol. IV, pp. 316). Minneapolis: University of Minnesota Press.

Finocchiaro, M. (1980). Galileo and the art of reasoning. Dordrecht: D. Reidel.

Garber, D. (1992). Descartess metaphysical physics. Chicago: University of Chicago Press.

Gasking, E. (1967). Investigations into generation 16511828. Baltimore: Johns Hopkins Press.

Greenwood, J. (1990). Two dogmas of neo-empiricism: The theory-informity of observation and the

QuineDuhem thesis.Philosophy of Science, 57, 553574.

Hempel, C. (1965). Empiricist criteria of cognitive significance: problems and changes. In idem, Aspects of

scientific explanation (pp. 101122). New York: Free Press.

Hempel, C. (1970). On the standard conception of scientific theories. In M. Radner, & S. Winokur (Eds.),

Minnesota studies in the philosophy of science (Vol. IV, pp. 142163). Minneapolis: University of

Minnesota Press.

Hoyningen-Huene, P., & Sankey, H. (2001). Incommensurability and related matters. Dordrecht/London:Kluwer Academic.

Kitcher, P. (1978). Theories, theorists, and theoretical change. Philosophical Review, 87, 519547.

Kuhn, T. (1962). The structure of scientific revolutions. Chicago: University of Chicago Press.

Kuhn, T. (2000a). Afterwords. In idem, The road since Structure (pp. 224252). Chicago: Chicago

University Press. (First published 1993)

Kuhn, T. (2000b). Commensurability, comparability, communicability. In idem, The road since Structure

(pp. 3357). Chicago: Chicago University Press. (First published 1983)

Kuhn, T. (2000c). Metaphor in science. In idem, The road since Structure(pp. 196207). Chicago: Chicago

University Press. (First published 1997)

Kuhn, T. (2000d). The natural and the human sciences. In idem, The road since Structure (pp. 216223).

Chicago: Chicago University Press. (First published 1991)

Kuhn, T. (2000e). Possible worlds in history of science. In idem, The road since Structure (pp. 5889).

Chicago: Chicago University Press. (First published 1989)



21/21

Kuhn, T. (2000f). The road since structure. In idem, The road since Structure (pp. 90104). Chicago:

Chicago University Press. (First published 1991)

Kuhn, T. (2000g). The trouble with the historical philosophy of science. In idem, The road since Structure

(pp. 105120). Chicago: Chicago University Press. (First published 1991)

Kuhn, T. (2000h). What are scientific revolutions? In idem,The road since Structure(pp. 1432). Chicago:

Chicago University Press (First published 1987)

Lakatos, I. (1970). Falsification and the methodology of scientific research programmes. In I. Lakatos, &

A. Musgrave (Eds.), Criticism and the growth of knowledge (pp. 91195). Cambridge: Cambridge

University Press.

Leicester, H. (1971). The historical background of chemistry. New York: Dover.

Levin, M. (1979). On theory-change and meaning-change. Philosophy of Science, 46, 407424.

Newton, I. (1952). Opticks. New York: Dover. (First published 1704)

Newton, I. (1962). On the gravity and equilibrium of fluids. In A. Hall, & M. Hall (Eds. & Trans.),

Unpublished scientific papers of Isaac Newton (pp. 121156). Cambridge: Cambridge University Press.

Newton, I. (1972).Philosophiae naturalis principia mathematica(3rd ed.) (A. Koyre, & I. B. Cohen, Eds.).

Cambridge, MA. Harvard University Press. (First published 1726)Newton, I. (1999).The principia: Mathematical principles of natural philosophy (I. Cohen & A. Whitman,

Trans.). Berkeley: University of California Press. (First published 1687)

Popper, K. (1992). The logic of scientific discovery. New York: Routledge. (First published 1959)

Quine, W. (1953). Two dogmas of empiricism. In idem, From a logical point of view: 9 Logico-philosophical

essays (pp. 2046). New York: Harper Torchbooks.

Salmon, W. (1991). The appraisal of theories: Kuhn meets Bayes. In A. Fine, M. Forbes, & L. Wessels

(Eds.),PSA 1990 (Vol. 2, pp. 325332). East Lansing, MI: The Philosophy of Science Association.

Scheffler, I. (1963). The anatomy of inquiry. Indianapolis: BobbsMerrill.

Westfall, R. (1983).Never at rest: A biography of Isaac Newton. Cambridge: Cambridge University Press.


Documents

Brown Incom Reconsidered