The edenic theory of reference

Abstract

I argue for a theory of the optimal function of the speech act of referring, called the edenic theory. First, the act of singular reference is defined directly in terms of Gricean communicative intentions. Second, I propose a doxastic constraint on the optimal performance of such acts, stating, roughly, that the speaker must not have any relevant false beliefs about the identity or distinctness of the intended object. In uttering a singular term on an occasion, on this theory, one represents oneself as not having any confused beliefs about the object to which one intends to refer. This paves the way for an intentionalist theory of reference that circumvents well-known problems, which have not been adequately addressed before in the literature.

Keywords:

• Reference
• confusion
• identity
• intentionalism
• pragmatics

Acknowledgements

This paper has benefited from discussion, suggestions, and encouragement from Brian Ball, Daniel Harris, Cressida Gaukroger, Dominic Alford-Duguid, Ben Phillips, François Recanati, Stephen Neale, Gary Ostertag, Michael Devitt, Deirdre Wilson, Eliot Michaelson, Matt Moss, Finnur Dellsén, Thomas Hodgson, Rachel McKinney, Daniel Deasy, Maria Baghramian, Hrafn Ásgeirsson, and Indrek Reiland. Thanks also to audiences at PLM 3 in Oslo, SPE 8 in Cambridge, The M&L Seminar at Institut Nicod, at the 2016 Pacific APA in San Francisco, and thanks to several anonymous referees. Finally, many thanks to Alexandru Radulescu, who provided me with very helpful and engaging comments at the Pacific APA.

Notes

No potential conflict of interest was reported by the author.

1 See, for example, Gauker (2008); King (2014); Lewis (2013); Perry (2012); Reimer (1992); Speaks (2016).

2 See, e.g. Bach and Harnish (1979); Carston (2002); Grice (1989); Neale (2005); Schiffer (2014); Wilson and Sperber (2012).

3 See, for example, Grice (1969); Harris (2016); Loar (1986); Neale (1992; 2005; 2016); Schiffer (1972,1982, 128–137, 2003,2006,2014). For a more skeptical view, see, e.g. Lance and O’Leary Hawthorne (1997, 287–297).

4 It should be noted, however, that in the Gricean project of explicating linguistically encoded meaning, the notion of non-linguistic acts of reference – in communication – must play a foundational role. And, further, non-communicative and audienceless uses of language must be given a separate treatment. These issues are not to be resolved in this paper.

5 For example, Strawson (1950, 17–19). See Bertolet (1987, 202–206); Schwarz (1976, 67–68); Stine (1978, 51–53). The Strawsonian tradition is recently enjoying resurgence, see, especially, Hanks (2011,2015); Soames (2010,2014, 105). Soames speaks of the act types of targeting and predicating, combinations of which give rise to full propositional act types.

6 I focus on beliefs only for simplification. Clearly, any other propositional attitude would do just as well. Note also that when there is no hearer, we stipulate that $S=H$ (Schiffer 1972, 76–80).

7 Following Sperber and Wilson (1995, ch. 1), among others, the definition of RW skips Grice’s third condition for speaker meaning, stating roughly that H’s recognition that S intends (1) should function, in part, as a reason for (1). The condition may well be on to something important about the nature of speaker meaning and, derivatively, speaker reference, but it involves particular complications that I wish to sidestep here. See in particular, García-Carpintero (2001, 102–103); Neale (1992, 547–549); Wharton (2009, ch. 2).

8 This proviso is added because, as is well known, Grice (1989, 106–112) and others gave counterexamples to (1) as it occurs in the explication of speaker meaning. In an oral history exam, the examiner asks the student when the Battle of Waterloo was fought. Student replies ‘1815’ saying thereby, it seems, that it was fought in 1815. But of course the student does not intend to produce this belief in the examiner, for he (the student) knows that she already knows that it was fought then. Thus, Grice distinguished between exhibitive and protreptic utterances. For the latter (1) is left unchanged, for the former, it becomes: to produce thereby in H, a belief that S believes that p. McDowell (1980, §5) and Neale (1992, 545–547) point out that this makes our (exhibitive) utterances about our own mental states rather than about the world itself, which appears to be implausible. Based on recent work in relevance theory on hearers’ epistemic vigilance – especially Sperber et al. (2010) – I incline to the view that this is entirely acceptable. All of our utterances could be exhibitive but we could still explain how communication provides speaker/hearers with knowledge of the world. I keep the protreptic form here only for simplification.

9 I thank an anonymous referee for this journal for providing this example and pointing out the issues it raises.

10 See, for example, Reimer (1992,2004); Perry (2012); Kripke (1977); Wettstein (1984); King (2014).

11 Åkerman (2010); Bach (1987, 183–186, 1992); Neale (2005, 184–185); Strawson (1974, 52); Unnsteinsson (2017).

12 The latter is a possible gerrymandered object in the actual world. Whether the former is a possible object at all is controversial but we need not make metaphysical assumptions of this sort here. For convenience I sometimes assume that the Superman who is not Clark Kent is a non-actual possible object, but it can also be assumed that this is a non-actual impossible object. Other models of the mental state of confused identity could have been used or appropriated here with similar results – e.g. those of Millikan (2000), Camp (2002), Lawlor (2007), or Recanati (2012). For a full defense and elaboration of the model itself, see Unnsteinsson (2016).

13 On this basis, Camp (2002, 33) and Millikan (1994,1997) both insist that combinatory confusion cannot be described in terms of belief in a proposition, because the relevant singular representations – of Bill as separate from Biff, and vice versa, in this case – are unavailable to the speaker.

14 Note that I assume, like the majority of philosophers working on these issues, a distinction between relational or object-dependent and satisfactional or descriptive contents; confused identity only arises for properly relational attitudes, i.e. attitudes whose content is a proposition with an object dependent truth condition (see, e.g. Bach 1987; Recanati 2012).

15 For discussion, see, for example, Camp (2002, ch. 2); Evans (1982, ch. 11); Field (1973).

16 A full statement of the edenic theory also involves a related explication of speech acts of singular coreference but, for reasons of space, that’s a story for another occasion.

17 Some may think that the edenic theory is more or less similar to Jerry Fodor’s view in The Elm and the Expert (1994) and, so, is subject to all the objections already mounted against the arguments in Chapter 2 of that book. See, especially, Arjo (1996); Aydede and Robbins (2001); Segal (1997); Wakefield (2002). Schneider (2005) defends the view.

Fodor argues, unsuccessfully by most accounts, that Frege cases are ceteris paribus exceptions to psychological laws; pathological cases which can be idealized away. Richard Heck (2012, 136n6) reports that Fodor abandoned the view. The edenic theory is not in the business of arguing that identity confusion gives rise to exceptions to intentional generalizations and, further, it does not claim that Frege cases are rare; in fact, it is assumed here that they are ubiquitous. The theory is also exclusively concerned with a certain specific communicative speech act, defined partly in terms of speakers’ mental states. Certainly, however, the big picture motivation is very similar to Fodor’s: developing a strongly externalist or information-theoretic account of reference while conceding as little as possible to Frege-type puzzles.

18 So, the edenic theory is not about reference ‘in thought’, unless, perhaps, if it assumed that we think in a language and that this must be some natural, public language. This is not assumed here.

19 A group of theorists unsympathetic to the Gricean program label it the ‘Lockean’ or ‘communicative’ conception of language, arguing that it is naïvely mentalistic and in need of significant revision or replacement. See, in particular, Bar-On (2013); Carruthers (1996); Gauker (2003); Green (2007). These authors are, at least in those publications, united in the belief that utterances express mental states directly, rather than being intended to produce some cognitive effect in an audience.

20 See also Elgin (2007); Mizrahi (2012); Dellsén (2016).

21 Stephen Neale (2004, 71–72, 2016) argues that this ought to be considered the ‘master question’ – i.e. the question in terms of which theoretical terms are to be justified – in philosophy of language: ‘How rich of an explanation can we provide of our capacity to express and sharpen our thoughts, and to communicate information about the world and about our beliefs, desires, plans, commitments, hopes, fears, and feelings so efficiently – so quickly, systematically and consistently – using various noises, marks, and gestures?’ My aim is to answer this question for singular reference in particular.

22 As Chomsky (1980, 225) puts it, ‘pragmatic competence places language in the institutional setting of its use, relating intentions and purposes to the linguistic means at hand’. See also Carston (1998, 5–7, 2002, 10–11); Neale (2005, 188–189).

23 This should remind one of Timothy Williamson’s (2000, 243) so-called knowledge norm of assertion. I do not, however, endorse his view that the knowledge norm is constitutive of assertion – but it is part of speakers’ semantic or pragmatic competence. Similarly, if the reference rule is constitutive of anything it is merely constitutive of the theoretical notion of RW, which, on my view, also captures an important aspect of pragmatic competence with singular terms.

24 Cf. Evans (1982, 310); Williamson (2000, 252n6).

25 The sense in which confusion is corruptive is explained and argued for in more detail in Unnsteinsson (2016, 219–224).

26 For a healthy dose of skepticism about the notion of deferred reference, see Greenberg (2014).

27 Compare what Millikan says about the imperative mood: ‘If no token of the imperative mood ever effected more than an abortive attempt or intention to comply with it, it is clear that speakers would soon cease to use the imperative forms at all or to use them as they now do’ (1984, 56). Note, however, that I am only endorsing this as a legitimate form of explanation and I don’t necessarily accept Millikan’s particular explanation of the imperative mood. Much care is needed in formulating normal conditions for utterance types. See Origgi and Sperber (2000) for some guidance on to how to do this.

28 But apparently not to others, see, e.g. Bach (1987, 157–159).

29 Michael Devitt (1974, 201, 1981) makes a very similar point.

30 So, even if e-reference is partly defined in terms of a prior notion of ‘referring with’, the claim is that e-reference is the notion of reference that is explanatorily fundamental in theorizing about communication. To be blunt, the hypothesis is that a notion of reference that abstracts away from Frege-type puzzles will play the reference-role in our final theory.

31 Similar arguments would apply to Evans’ (1973, 11) ‘Madagascar’ example.

32 Michaelson explores some similar examples in his (2013) and ‘Speaker’s Reference, Semantic Reference, Sneaky Reference’ (n.d.).

33 Talking in terms of two distinct names is only a simplification here. It would perhaps be more accurate to talk of one name with two bearers. The speaker would still be introducing a new convention, namely the convention that Jones’ name can now also be used to refer to Smith.

Additional information

Funding

This work was supported by the Icelandic Research Fund [grant number 163132-051]; Irish Research Council [grant number GOIPD/2016/186].