Relativism and the expressivist bifurcation

Abstract

Traditional expressivists want to preserve a contrast between the representational use of declarative sentences in descriptive domains and the non-representational use of declarative sentences in other areas of discourse (in particular, normative speech). However, expressivists have good reasons to endorse minimalism about representational notions, and minimalism seems to threaten the existence of such a bifurcation. Thus, there are pressures for expressivists to become global anti-representationalists. In this paper I discuss how to reconstruct in non-representationalist terms the sort of bifurcation traditional expressivists were after. My proposal is that the relevant bifurcation can be articulated by appeal to the contrast between relativistic and non-relativistic assertoric practices. I argue that this contrast, which can be specified without appeal to representational notions, captures the core intuitions behind the expressivist bifurcation (in particular, it captures the anti-realist intuitions motivating many expressivist proposals).

Keywords:

• Expressivism
• anti-representationalism
• pragmatism
• relativism
• bifurcation thesis
• minimalism

Acknowledgements

Thanks to an anonymous referee for this journal and audiences at the Granada Workshop on Epistemic Expressivism and the 2017 EPISOC summer school for their feedback.

Notes

1. Compare Charlow (2015, 34): ‘to know a sentence’s semantic value is to be in a position to know which state of mind is constitutively involved in the acceptance of that sentence.’

2. The distinction between semantic value and informational upshot is needed because informational upshots (e.g. propositions constructed as sets of worlds) do not always compose, whereas semantic values are standardly expected to respect compositionality (see Lewis 1980; Rabern 2012).

3. Price (2011, 2015) proposes reconstructing the relevant bifurcation in terms of the notion of ‘e-representation’, which would be a non-representational relation of tracking or covariance with environmental features (e-representation would be distinctive of the types of speech that the expressivists wants to classify as descriptive). My suspicion is that this notion will either turn out to be actually representational or it will not suffice to characterize descriptive speech. Anyway, I will not pursue these worries here, but rather explore my own proposal.

4. Although Brandom tends to talk of the scoreboard of individual speakers, we can consider a common conversational scoreboard reflecting the commitments and entitlements shared by the speakers, as participants of the conversation.

5. In particular, Brandom (1994, ch. 3) argues that the audience acquires an entitlement by deferral to the proposition asserted. This means that the audience’s responsibility to vindicate their entitlement to the proposition (if suitably challenged) can be delegated to the speaker who made the assertion.

6. From this point of view, worlds would be characterized by maximally specific propositions, that is propositions that answer all possible questions (or at least, all possible questions relevant for the purposes of the conversation). In this way, a proposition characterizing a world will be incompatible with either p or −p for any relevant proposition p.

7. Brandom’s broader story includes connections between commitments to accepting propositions and practical commitments to acting in certain ways, and also connections between occupying certain perceptual positions and being entitled to accept some proposition (see Brandom 1994, ch. 4).

8. This is compatible with the existence of a technical notion of truth at a context and index playing a role in semantic theories to which the pragmatist may resort (see MacFarlane 2014; Yalcin 2011). The pragmatist is only committed to eschewing appeals to a substantive notion of truth in her ultimate account of why expressions mean what they do, not in technical explanations at the semantic level.

9. Or a set of such classifications, if we want to make room for normative uncertainty. I will leave these complications aside.

10. MacFarlane (2014) claims that retracting an assertion is a speech act made by saying things like ‘I take that back’. Likewise, rejections or challenges could be seen as explicitly expressed by saying something like ‘Take that back, you are wrong!’.

11. Another influential presentation of relativism is provided by Kölbel (2008, 2015).

12. This does not mean that MacFarlane is committed to an anti-representationalist version of pragmatism. It may be that the relevant practice is to be characterized by appeal to representationalist notions (see MacFarlane 2010 for discussion of pragmatist metasemantics).

13. Following MacFarlane (2014, ch. 4), I allow for the possibility that the truth value of a proposition depends both on features of the context of assessment and the context of utterance. For my purposes here, it is enough to focus on the dependence on the context of assessment. Therefore, in what follows I drop the mention to the context of utterance.

14. Note that such generalized norms can be used to give a unified account of relativistic and non-relativistic assertion. In non-relativistic domains, such norms will reduce to the non-relativistic ones, since the truth value of propositions will be insensitive to contexts of assessment.

15. Relativistic assertion so understood must be distinguished from contextualist or indexicalist speech, in which the same sentence can be used to assert different propositions in different contexts of utterance. What Harman (1975) calls ‘moral relativism’ is actually a version of indexicalism.

16. It is important to note, however, that asserting something like ‘Stealing is wrong‘ does not amount to reporting one’s endorsement of the relevant values or norms. Rather, one would be making an assertion that is true only in relation to contexts of assessment in which certain values are endorsed.

17. It may be hypothesized that in those areas of discourse where intersubjective coordination is more valuable, we will be more inclined to make use of the normative friction introduced by assertion, and we will be more reluctant to retreat to agent-relative claims merely reporting one’s attitudes.

18. Of course, we need to characterize the notion of relativistic assertion, and this is likely to involve some modifications in our semantic theory. But these modifications will be reasonably conservative, and the resulting picture will be continuous with standard theories of meaning (see MacFarlane 2014, ch. 5). Note, in particular, that although relativism introduces a technical notion of truth at a context of assessment, the monadic truth predicate used in ordinary speech will still behave in a standard way ( i.e. one will take assertions of ‘p is true’ to be permissible whenever assertions of ‘p’ are).

19. For discussion of the relations between relativism and expressivism, see MacFarlane (2014), Field (2009) and Stalnaker (2014).