Inferentialism and semantic externalism: a neglected debate between Sellars and Putnam

ABSTRACT

In his 1975 paper “The Meaning of ‘Meaning’”, Hilary Putnam famously argued for semantic externalism. Little attention has been paid, however, to the fact that already in 1973, Putnam had presented the idea of the linguistic division of labour and the Twin Earth thought experiment in his comment on Wilfrid Sellars’ “Meaning as Functional Classification” at a conference, and Sellars had replied to Putnam from a broadly inferentialist perspective. The first half of this paper aims to trace the development of Putnam’s semantic externalism, situate his debate with Sellars in it, and reconstruct the two arguments he presented against Sellars. The second half of this paper aims to reconstruct how Sellars replied to Putnam. I argue that Sellars not only accepts the social character of language but also suggests how inferentialists can accommodate the contribution of the world. Sellars’ key idea is that substance terms have a ‘promissory note aspect’ which is to be cashed out in a successor conceptual framework. I reconstruct Sellars’ position as ideal successor externalism, and compare it with temporal externalism.

KEYWORDS:

• Inferentialism
• semantic externalism
• history of analytic philosophy
• Wilfrid Sellars
• Hilary Putnam

Acknowledgements

This paper was submitted when I was a visiting fellow at the Tilburg Center for Logic, Ethics, and Philosophy of Science (TiLPS). Thanks to the TiLPS for providing a visiting fellowship and a wonderful research environment. I would also like to thank Fuse Academic Foundation for Fellowship for Early Career Researchers. I am also grateful to Bill deVries, Jeremy Koons, Jun Kuzuya, Shigaku Nogami, and two anonymous referees for helpful comments on earlier drafts. Versions of this paper were presented at the University of Tokyo and at the Japan Forum for Young Philosophers 2020 in Tokyo. I would like to thank the audience on these occasions.

Notes

1 Although Putnam’s 1973 paper, “Meaning and Reference”, also contains the idea of the linguistic division of labor and the Twin Earth thought experiment, as we will see in section 3, his comment on Sellars predates it.

The other speakers at this conference are Donald Davidson, Michael Dummett, and Saul Kripke, and the other commentators are Daniel Dennett, Gilbert Harman, David Kaplan, David Lewis, Charles Parsons, Barbara Partee, and Willard van Orman Quine. The proceedings of this conference, published in Synthese in 1974, includes all the papers, comments, and general discussions except for Kripke’s paper “Vacuous Names and Fictional Entities” and the comments on it. According to the editors, Kripke’s paper was not included in the proceedings “[f]or reasons beyond the editor’s control” (Troyer and Wheeler, “Preface”, 307). It is now collected in Kripke, Philosophical Troubles.

2 One exception is Scharp, “Sellars’ Anti-Descriptivism”. But he does not discuss what I take to be of the most importance in Sellars’ reply, that is, the idea that substance terms have a ‘promissory note aspect’. To elaborate this idea is the aim of the second half of this paper.

3 Externalism about mental contents was proposed and developed after the “Meaning of ‘Meaning’” by McGinn’s “Charity, Interpretation, and Belief” and Burge’s “Individualism and the Mental”. In chapter 1 of Reason, Truth, and History, Putnam also embraces externalism about mental contents, and explores its implications for scepticism about the external world. For a further discussion on externalism about mental contents, see Lau and Deutsch, “Externalism about Mental Content”.

4 But semantics and metasemantics may not be sharply separated from each other. See Burgess and Sherman, “Introduction”.

5 As Murzi and Steinberger point out (“Inferentialism”, 218n2), Brandom’s distinction between formal and philosophical semantics seems to correspond to the distinction between semantics and metasemantics. See Brandom, Making It Explicit, 143–5, “Reply to Danielle Macbeth”, 340, and “Reply to Michael Dummett”, 342.

6 See the title of Part I of the Twin-Earth Chronicles.

7 It may also be worth noting here that in his Meaning and the Moral Sciences, Putnam appropriates Sellars’ terminology – ‘language entry rules’ and ‘language exit rules’ – without crediting it to Sellars (see Putnam, Meaning and the Moral Sciences, 110–1).

8 Putnam goes on to say “I frequently heard him [Sellars] speak in group discussions, however, and I soon realized that there were serious disagreements between my position and that of both Feigl and Sellars”. The disagreements Putnam has in mind here are concerned with the “analytic/synthetic dichotomy” and “Adolf Grünbaum’s views on physical geometry” (Putnam, “Intellectual Autobiography”, 43). It is beyond the scope of this paper to discuss these issues.

9 For a recent review of Putnam’s arguments, see Haukioja “Internalism and Externalism”.

10 But according to his autobiography, Putnam first explained and defended semantic externalism at his lectures in the 1968 Summer Institute in Philosophy of Language in Seattle (Putnam, “Intellectual Autobiography”, 105n97). Burge counts “Is Semantics Possible?” as one of “the great trio of articles” in which the core of Putnam’s contributions to semantics emerged (“Putnam’s Contributions”, 235). The others are “Explanation and Reference” and “The Meaning of ‘Meaning’”.

11 In another retrospective essay, Putnam says that “I wrote at the end of 1972 [a paper] entitled ‘The Meaning of ‘Meaning’’” (“A Half Century”, 196).

12 In “Explanation and Meaning” (1973), Putnam uses the ‘elm’ and ‘beech’ example but does not use the term ‘the linguistic division of labor'.

13 After the conference, Michael Dummett wrote a paper on Putnam’s comment on Sellars, which was first published as “Postscript” in the proceedings and reprinted as “The Social Character of Meaning” in his Truth and Other Enigma (1978).

14 This paper was originally presented at a Boston Colloquium in the Philosophy of Science in December 1963, where Sellars also presented a paper entitled “Scientific Realism or Irenic Instrumentalism”.

15 Here, psychological states are to be understood as ‘narrow’ psychological states (see Putnam, “Meaning of ‘Meaning’”, 137; Mind, Language, Reality, 221). McDowell rejects this view of the mind, and argues that not only the meaning but also the mind should be understood as world-involving (see McDowell, “Putnam on Mind and Meaning”).

16 I should note, however, that I omitted the word ‘knows’ in the above quotation. In his comment on Sellars, Putnam includes so-called wide mental states such as ‘knowing’ in the first assumption.

17 It should be noted that Crane’s interest is not in linguistic meaning but only in mental contents (Crane, “All the Difference”, 9, 11, 15). But the common concept strategy itself could also be applied to linguistic meaning.

18 Strictly speaking, this statement is made about ‘gold’ rather than about ‘water'. But as Sellars goes on to discuss ‘water’ on the next page, he seems to think that the same applies to ‘water'.

19 In “A Semantical Solution of the Mind-Body Problem”, Sellars suggests that he has borrowed the expression ‘promissory note’ from Feigl (SSMB, 196). Although Sellars does not refer to any particular paper of Feigl’s, see, for instance, “Operationism and Scientific Method”.

20 John Locke, who introduced the term ‘sortal’, uses this term in a general way which does not restrict it to countable nouns (see Essay, III.iii.15), and some analytic philosophers use ‘sortal’ as including mass nouns (see Lowe, More Kinds of Being, 14). Sellars seems to be following Locke in calling ‘gold’ and ‘water’ sortals, as he refers to Locke in his “Reply” (462).

21 Thing-kind expressions may include terms that are not substance terms. But I shall set aside this possibility and focus on substance terms.

22 Sellars himself seems to say that the referent of a substance term is the successor term (Sellars, “Reply”, 462). But I take it that while the view that the referent is determined by the successor term is a consequence of his idea of the promissory note aspect, the view that the referent is the successor term is derived from the conjunction of the idea of the promissory note aspect with his nominalism, which claims that “qualities, relations, classes, propositions and the like […] are linguistic entities. They are linguistic expressions” (AE, 627). Indeed, in his reply to Putnam, Sellars goes on to compare our temptation to reify what the ideal substance term would stand for with our temptation to construe what ‘triangular’ stands for as an abstract extra-linguistic object (“Reply”, 462). In this paper, I will focus on the question of how the referent of a substance term is determined, and set aside his nominalism.

23 This is not to say that Sellars is not a global temporal externalist. I leave the question open whether he is or not. My point is only that the above discussion is specific to substance terms.

24 Sellars discusses correspondence rules in a series of papers (LT, TE, SRII, SRT). See also O’Shea, Wilfrid Sellars, 32–47, 158–63, and Christias, “The Manifest-Scientific Image Distinction”.

25 In his “Abstract Entities”, Sellars distinguishes two sorts of criteria: immediate and ultimate (AE 645–6). Immediate criteria are those we are actually using and ultimate criteria are those we are to adopt in the end.

26 Although Brandom’s inferentialism can be seen as developing some of Sellars’ master-ideas, Brandom diverges from Sellars in refusing to appeal to such a “privileged perspective” (Making It Explicit, 599) and stressing the importance of the I-thou relation in determining the meaning and reference. In his recent book on Sellars, Brandom objects to “Sellars’ commitments to a Peircean end-of-inquiry science” by insisting that “it is very difficult to make sense of this notion” (From Empiricism to Expressivism, 79n43). Further, in his recent Book on Hegel, instead of appealing to the ideal perspective, Brandom provides a Hegelian historical account of the content of a concept in terms of a “process of determining” it (Brandom, A Spirit of Trust, 6). According to this account, in rationally reconstructing the prior applications of a term and thereby making them a progressive history, the present applications of that term both acknowledge the authority of that history and are authorized by that history, and similarly, the present applications of that term both petition its future users to acknowledge the authority of its present applications, and authorize its future applications. Brandom often elaborates this account by using an analogy with the evolution of legal terms in the common law tradition (see Brandom, Reason in Philosophy, chap. 3, and A Spirit of Trust, chaps. 12, 16, and Conclusion). In the next section, I will suggest how the ideal framework might work as the regulative ideal in our inquiry.

27 The matter will be more complicated if we consider the existence of isotopes. But the issues it raises are not specific to Sellars and hence set aside here.

28 This explanation can be seen as applying to the case of substance terms the basic strategy of Brandom’s inferentialist account of what we are doing when we use singular terms, demonstratives, or semantic terms (see Part Two of his Making It Explicit).

29 Dionysis Christias calls such a process the process of “preservation through transformation” (“The Manifest-Scientific Image Distinction”, 1312n15).

30 Of course, Sellars also famously claims that “the objects of the observational framework do not really exist – there are really no such things. They [correspondence rules] envisage the abandonment of a sense and its denotation [of observational terms]” (LT, §52, 126; emphasis added). But I take it that what he is claiming not to really exist is the objects of the observational framework as they had been originally conceived, before that framework was corrected and refined, and, in the same way, what he is claiming to be abandoned is the sense and denotation of observational terms as they had been originally conceived, before these terms were corrected and refined.