SORITES, ISSN 1135-1349
Issue #11. December 1999. Pp. 6-14.
Reference Change of Natural Kind Terms
Copyright © by SORITES and Luis Fernández Moreno
Reference Change of Natural Kind Terms
by Luis Fernández Moreno
1. Introduction
Kuhn's incommensurability thesis asserts that the languages in which successive or rival scientific theories are formulated are not mutually translatable. Nevertheless, Kuhn restricted the scope of this thesis in two senses. First, the translation failure between the languages of theories only concerns a sort of terms, namely, the kind terms; second, that failure of translation only affects some of the kind terms of those languages or rather a small subset of interrelated terms.Foot note 1_1 Thus Kuhn claims that incommensurability has only a local character, since most terms common to rival theories are mutually translatable.Foot note 1_2
Although the incommensurability thesis so conceived, to wit, as a semantic thesis, concerns the concept of translation and hence meaning, this thesis is often justified on the basis of the concept of reference. The thesis of referential incommensurability -- as it may be called -- would assert that kind terms from rival theories have a different reference. This thesis rests on the claim that a change of theory entails changes in the reference of kind terms common to those theories or, in short, that theory change involves reference change. This thesis could be called «thesis of reference change».
The acceptability of the thesis of reference change will depend on how the reference of kind terms and especially of natural kind terms is determined. Since according to the causal theory of reference the thesis of reference change should be rejected, it is understandable that Kuhn examined the causal theory of reference. Kuhn [1979] already contained some remarks on this theory, but Kuhn's last judgment on the causal theory is formulated in two later writings, Kuhn [1989] and [1990]. In these two writings Kuhn examined critically the causal theory of reference and especially Putnam's natural kind terms reference theory such as it is formulated in Putnam [1975a]. Though Kuhn asserts that he restricts himself primarily to Putnam's theory,Foot note 1_3 Kuhn claims that his objections against Putnam's theory would also apply to other versions of the causal theory. One of Kuhn's objections, and indeed his main objection, concerns the question of whether the reference of natural kind terms may change. On examining this objection I will pay attention to the factors which might involve changes of reference and to the doctrines which may lend support to the thesis of reference change. I shall argue that though the reference of natural kind terms is open to change, the proponents of the thesis of reference change have not conclusively established their thesis.
2. Kuhn's Objections to Putnam's Theory
It is advisable to begin with a brief characterization of Putnam's natural kind terms reference theory. According to this theory the reference of a natural kind term is determined by two factors, to wit, by some initial act of dubbing samples of the kind -- let us call the samples there involved «original samples» -- and by the relation sameness-of-kind; this relation is constituted by the so-called essential properties, i.e. by properties concerning the internal structure of samples of the kind, which are discoverable only by scientific research. Thus, in Putnam's words, there is a «contribution of the environment» to the determination of the reference of natural kind terms. Putnam mentions another sort of contribution too, namely, the «contribution of the society»; by this it is meant that there is a linguistic division of labor or, more precisely, there is a subset of members of the linguistic community, the set of «expert» speakers -- they will be in the main members of a scientific community -- who have a more reliable knowledge than laymen of the reference of natural kind terms; Putnam claims that non-expert speakers rely on the judgment of expert speakers about the reference of these terms.Foot note 1_4 Kuhn agrees with Putnam's assertions on the linguistic division of labor and therefore with Putnam's theory of reference transmission, but he does not accept some central claims of Putnam's theory of reference determination, which entail the rejection of the thesis of reference change. According to Putnam the original samples of the kind and the relation sameness-of-kind are independent of our theories; thus theory change does not involve reference change and therefore the thesis of reference change has to be rejected.
Kuhn's main objection to Putnam's theory is that this theory rules out and therefore cannot explain changes of reference. More precisely, Kuhn questions that the samples determining the reference of natural kind terms and the relation sameness-of-kind have remained stable through theory change. Kuhn formulates this objection on analysing Putnam's Twin-Earth thought experiment concerning the term «water». It is noteworthy, however, that Kuhn's remarks will not concern the term «water» as employed in everyday life or by laypeople, but as used «within the community of scientists and philosophers to which Putnam's argument needs to be applied.»Foot note 1_5
Kuhn attends specially to the second part of Putnam's thought experiment and to the referential history of the term «water» as used on Earth, more precisely, to the following passage from [1975a]:
[L]et us roll the time back to about 1750. At that time [...] [t]he typical Earthian speaker of English did not know water consisted of hydrogen and oxigen [...] Yet the extension of the term `water' was just as much H2O on Earth in 1750 as in 1950 [...]Foot note 1_6
Kuhn questions that the term «H20» has the same extension as the term `water' such as this term was used in 1750. The extension of the term «water» from the perspective of present chemistry or of 1950's chemistry is the set of samples of H20, regardless of whether these samples are in solid, liquid, or gaseous state. But Kuhn claims that this was not so from the perspective of 1750's chemistry. According to Kuhn, at that time, namely, before the so-called «Chemical Revolution», which took place in the 1780's, it was considered that different chemical substances corresponded to the three states of aggregation, to wit, to the solid, liquid, and gaseous states, as it was regarded that a chemical substance could only exist in one of these three states, and water was conceived in 1750's chemistry as having the essential property of being a liquid.Foot note 1_7 Hence the term «water», such as this term was used in 1750's chemistry, would not be co-referential with the term «H20», but with the term «liquid H2O» (or «close-packed H2O particles in rapid relative motion»). From here it follows that the term «water», such as it was used in 1750's chemistry, is not co-referential or coextensive with the term «water», such as it is used in present chemistry; from the perspective of 1750's chemistry a piece of ice would not belong to the extension of the term «water», while it does belong to the extension of the term «water» from the perspective of present chemistry or of 1950's chemistry. Thus in this case the relation sameness-of-kind has not remained stable through theory change, at least through changes in our scientific or rather meta-scientific theories about the notion of sameness-of-kind concerning chemical substances.
This is Kuhn's main objection against Putnam's theory and in general against the causal theory of reference, an objection which concerns the thesis of reference change, but Kuhn formulates in [1989] and [1990] two further objections against the causal theory, which bear on the so-called essential properties. On the one side, Kuhn attributes to the causal theory the thesis that only a single essential property determines the reference of each natural kind term, and then he objects that more than one essential property is required to determine the reference of the term «water» such as it was used in 1750, namely, the properties of being H2O and of liquidity.Foot note 1_8 However, the causal theory does not seem to be committed to the thesis that Kuhn attributes to the proponents of this theory; there is no reason why Putnam or other causal theorists could not accept that the reference of a term is determined not by a single property but by a conjunction of properties. On the other side, Kuhn questions the frequent assimilation by Putnam and in general by causal theorists of theoretical properties to necessary or essential properties and of so-called superficial properties or macroscopical properties to contingent ones. Kuhn claims that the so-called superficial properties are as necessary as the theoretical ones, since if a theory that posits the relevant theoretical properties could not predict the superficial properties or at least some of them, it would not be taken seriously. Thus Kuhn claims that, as theoretical properties have been posited to explain and predict superficial ones, the latter will be as necessary as the former.Foot note 1_9 Kuhn's remark is motivated by a certain underestimate of the so-called superficial properties by Putnam and in general by causal theorists; the point of the causal theory is just that internal structural properties are more determinant of a natural kind term's reference than macroscopical properties, since samples which share the former but not the latter ones should be still considered as samples of the kind, while samples which share the latter but not the former ones should not be regarded as samples of the kind. This contrast between internal structural properties and macroscopical properties could be formulated in the following way: every member of the kind has to have the internal structural properties that determine the kind, but it is not necessary that every member of the kind have all the macroscopical properties usually associated with the kind. Nevertheless, this contrast between internal properties and macroscopical ones is compatible with Kuhn's point that macroscopical properties play an important role in the determination of the reference of natural kind terms; they contribute to specify the relevant relation sameness-of-kind, since the internal properties which constitute this relation will be those that are responsible for such macroscopical properties.
2. Causal Theory and Reference Change
After having taken into account these two further objections let us return to Kuhn's main objection against Putnam's theory and in general against the causal theory of reference. Kuhn argues that the extension of the term «water» has changed between 1750 and 1950, since the relation sameness-of-kind has not remained stable through theory change. Now, if the causal theory of reference does not allow changes of reference, then the causal theory should be rejected.
The claim that theory change may involve changes of reference is very plausible and can be justified on the basis of the thesis that the relation sameness-of-kind does not only depend on the world, but also on our theories or conceptions about the notion of sameness-of-kind; thus changes in these conceptions may entail changes in the relation sameness-of-kind and so may result in changes in the reference of natural kind terms. This conclusion is illustrated and supported by Kuhn's historical example. Concerning this historical example I find advisable to make the following three remarks. First, I do not commit myself to the veracity of Kuhn's historical example; nevertheless, if someone would question the historical accuracy of Kuhn's example, I would ask him to regard it as a thought experiment. Kuhn's historical example, regarded as a thought experiment, illustrates and supports Kuhn's conclusion that the relation sameness-of-kind does not need to remain stable through theory change. Second, the scope of Kuhn's conclusion is, however, not very far-reaching since changes in our scientific theories do not usually entail changes in our conceptions about the notion of sameness-of-kind and so in the relation sameness-of-kind. Third, Kuhn's historical example may also be interpreted as supporting the plausibility of the claim that there have been modifications in the sorts of samples involved in the determination of the reference of natural kind terms.
Anyway, in regard to Kuhn's main objection against Putnam's theory and in general against causal theories of reference it must be noted that, though classical versions of the causal theory seem to make reference change impossible, not every version of the causal theory has to be committed to the immutability of reference. In fact, it may -- and should -- be allowed that the reference of a term is not only determined by the use of the term in the supposed initial dubbing but also by subsequent uses of the term.Foot note 1_10 Modifications in our uses of terms -- due to mistakes or to deliberate choice -- may be accompanied by changes of reference, especially if those modifications involve changes in the samples determining the reference of the term or in the notion of sameness-of-kind.
In regard to the first sort of change it is advisable to mention a possible interpretation of Putnam's reference theory according to which the samples determining the reference of a natural kind term are not the original ones, but the so-called «paradigmatic samples», which do not need to be the same as the former ones and which will be determined by experts. In a seldom quoted paper, Putnam [1975b] -- written in 1974 -, he makes some claims which could be interpreted in that way; for instance, after asserting that «we may no longer care about the original use of [a] term»,Foot note 1_11 Putnam regards as dubbers and as initiators of a chain of transmissions of a term «the original dubber, or the relevant expert».Foot note 1_12 Thus the reference of natural kind terms would be determined by paradigmatic samples and by the relation sameness-of-kind. Now, if the choice of paradigmatic samples is taken by experts, it is plausible to assume that their choice would depend partly on their theories so that theory change might involve changes in the paradigmatic samples and therefore may result in changes of reference.
Nevertheless, one can accept that there have been changes of reference, while rejecting the thesis of reference change; this thesis is stronger since it amounts to the claim that a change of theory always entails some changes of reference.
3. The Justification of the Thesis of Reference Change
Kuhn's justification for the thesis of reference change rests on two sorts of doctrines, one semantic and the other ontological. The ontological doctrine is a sort of antirealism which has been called by some authors, such as R. Nola, «relativist idealism».Foot note 1_13 According to this doctrine a change of theory involves a change of world, or rather of phenomenal world, though not of the world-in-itself. Now, since the concept of reference expresses a relation between language and world, and since the world-in-itself is assumedly inaccessible to us, the reference of terms must be given in the phenomenal world corresponding to a theory. And as a change of theory entails a change of phenomenal world, a change of theory would also involve changes of reference. Relativist idealism makes the thesis of reference change very plausible, but if the former were the only justification for the latter, the thesis of reference change would have no interest by itself, since the acceptability of this thesis would depend on the previous adoption of a certain ontological position.
However there is another way to vindicate the thesis of reference change, namely, to assume a sort of description theory of reference. According to the description theory, in its modern or cluster form, the reference of a natural kind term is determined by a cluster of descriptions that speakers associate with the term, where it is allowed that some descriptions are more central than others for the determination of the reference; to the extension of a term will belong the entities which satisfy, in John Searle's words, a «sufficient but [...] unspecified number» of such descriptions.Foot note 1_14 This last condition can be modified in order to obtain stronger and weaker versions of the description theory; a strong version would demand the satisfaction of all or, at least, of most the descriptions associated with the term. The description theory may incorporate the distinction between experts and non-experts mentioned by Putnam and claim that the relevant cluster of descriptions consists of descriptions that experts associate with the term or, at least, that they are the most central ones. But since it has to be expected that experts who support rival theories will associate different and even incompatible descriptions with a term, the cluster description theory may make plausible the thesis of reference change. Although Kuhn did not specify precisely how he thinks that the reference of natural kind terms is determined, he made some proposals about how natural kind terms are learned which may also be interpreted as proposals about how the reference of natural kind terms is determined, and which agree with the approach of the description theory. According to these proposals kind terms are learned -- and their reference is determined -- by recourse to exemplary members of their extensions -- which will be ostended to or described -- and to symbolic generalizations which contain these terms. Kuhn characterizes symbolic generalizationsFoot note 1_15 or, for short, generalizations as sentences of a theory which have the form of universal sentences or which can be easily put in that form; to the generalizations of a theory belong specially the laws of the theory. Kuhn's proposals about how the reference of natural kind terms is determined seem to constitute, at least partly, a sort of description theory, where the respective descriptions will be extracted from generalizations containing the terms. Now, insofar as theory change involves a substantial alteration of generalizations of the theory, it is likely to be accompanied by change of reference; thus the thesis of reference change becomes plausible, and especially if a strong version of the description theory is endorsed which requires the satisfaction of most the descriptions obtained from generalizations.
Although at present pure description theories do not have many followers, there is a certain agreement that some sorts of descriptions have to play a role in the determination of the reference of natural kind terms; for this reason most of present versions of the causal theory are descriptive-causal ones. Among those sorts of descriptions are the following.Foot note 1_16 First, the determination of the reference of a natural kind term requires the association with the term of a sortal or categorial term which contributes to eliminate the indeterminacy of ostension, for instance, the sortal term «metal» in the case of the term «gold». Second, descriptions of certain observable properties of the samples and descriptions which ascribe to the samples certain causal powers are needed to specify the relevant relation sameness-of-kind, since the internal properties which constitute this relation will be those that are responsible for such observable propertiesFoot note 1_17 and causal powers. Third, in the case of reference to natural kinds which are unobservable, it is necessary to use a description of the causal mechanism through which it is assumed that unobservable entities produce certain observable phenomena.
But if it is conceded that such sorts of descriptions have to play a role in the determination of the reference of natural kind terms, it must be admitted that variations in those descriptions may result in changes of reference. In this regard it may be relevant to take into consideration a plausible claim concerning the reference of names made by Putnam in [1973]:
[U]nless one has some beliefs about the bearer of the name which are true or approximately true, then it is at best idle to consider that the name refers to that bearer in one's idiolect.Foot note 1_18
Now, this remark concerning the reference of names may be applied to the reference of the rest of the terms, including natural kind terms; thus the referent of a natural kind term must be so as to make true or approximately true at least some of our beliefs or theories about it or rather some of the beliefs or theories of experts about natural kinds. For this reason the question of what is the referent of a natural kind term, such as it is used by experts, cannot be completely independent of what their beliefs or theories are. Thus it may not be excluded that drastic changes in experts' beliefs or theories be accompanied by changes of reference.
The proponents of the thesis of reference change may regard this conclusion as rather disappointing, since their thesis is stronger, namely, a change of theory always entails changes of reference. But in order to justify this thesis, and according to the aforementioned remarks, they would have to argue for an unplausible antirealist position, such as relativist idealism, or for a strong version of the description theory of reference. It is their turn to put forward good arguments for either of these claims.
Devitt, M. [1984]: Realism and Truth. Oxford: Blackwell; 2nd. ed., 1991.
Devitt, M. y K. Sterelny [1987]: Language and Reality. Oxford: Basil Blackwell.
Kuhn, Th.S. [1962]: The Structure of Scientific Revolutions. Chicago: University of Chicago Press; 2d. ed., enlarged, 1970.
[1970]: «Postscript-1969». In Kuhn [1962] 2d. ed., pp. 174-210.
[1979]: «Metaphor in Science». In A. Ortony (ed.), Metaphor and Thought, Cambridge, Cambridge U.P., pp. 409-419.
[1983]: «Commensurability, Comparability, Communicability». In P.D. Asquith and T. Nickles (eds.), PSA 1982, East Lansing, Philosophy of Science Association, vol. 2, pp. 669-688.
[1989]: «Possible Worlds in History of Science». In S. Allén (ed.), Possible Worlds in Humanities, Arts, and Sciences, Berlin, de Gruyter, pp. 9-32.
[1990]: «Dubbing and Redubbing: The Vulnerability of Rigid Designation». In C.W. Savage (ed.), Scientific Theories (Minnesota Studies in Philosophy of Science, vol. 14), Minneapolis, University of Minnesota Press, pp. 298-318.
[1991]: «The Road Since Structure». In A. Fine, M. Forbes, and L. Wessels (eds.), PSA 1990, East Lansing, Philosophy of Science Association, vol. 2, pp. 3-13.
[1993]: «Afterwords». In P. Horwich (ed.), World Changes. Thomas Kuhn and the Nature of Science, Cambridge, MIT Press, pp. 311-341.
Nola, R. [1980]: «`Paradigms Lost, or the World Regained' -- An Excursion into Realism and Idealism in Science». Synthese, 45, pp. 317-350.
Putnam, H. [1973]: «Explanation and Reference», en G. Pearce y P. Maynard (eds.): Conceptual Change, Dordrecht, Reidel. Reimp. en Putnam [1975], pp. 196-214.
[1975]: Mind, Language and Reality. Philosophical Papers, vol. 2, Cambridge, Cambridge University Press.
[1975a]: «The Meaning of `meaning'», en K. Gunderson (ed.): Language, Mind and Knowledge (Minnesota Studies in the Philosophy of Science, vol. 7), Minneapolis, University of Minnesota Press. Reimp. en Putnam 1975, pp. 215-271.
[1975b]: «Language and Reality», en Putnam [1975], pp. 272-290.
Sankey, H. [1997]: «Incommensurability: The Current State of Play». Theoria, 12, pp. 425-445.
Searle, J. [1969]: Speech Acts. An Essay in the Philosophy of Language. Cambridge: Cambridge U.P.
Sterelny, K. [1983]: «Natural Kind Terms». Pacific Philosophical Quarterly, 64, pp. 110-125.
University Complutense of Madrid. Spain