Expressivism and the Value of Truth

Expressivism in its most theoretically virtuous forms aspires to be an account of all evaluative claims. In a recent paper, Lynch (2009) has argued that expressivism cannot accommodate claims about the value of truth, since an expressivist account of any normative claim requires a ‘normatively disengaged standpoint’ which is unavailable in the case of truth (one cannot cease to value truth while still being an inquirer). In this paper I argue that Lynch’s objection to expressivism rests on an ambiguity. The expressivist can distinguish between a standpoint that is committed to certain evaluations and a standpoint that employs those evaluations in its explanations.

So why think that expressivism, so understood, cannot provide an account of (TG)? Lynch's argument can be summarised as follows.
(1) In order for an expressivist account of any normative claim to make sense it must be possible to a reach a standpoint disengaged from the making of that normative claim (85).
Hence, more particularly (2) In order for an expressivist account of the normative claim (TG) to make sense, it must be possible to make sense of a standpoint disengaged from the making of (TG), that is, an 'epistemically disengaged standpoint' (86).

But
(3) It is not possible to make sense of an epistemically disengaged standpoint ('the epistemically disengaged standpoint is an illusion': 86).

Hence
(4) An expressivist account of the normative claim (TG) doesn't make sense.
One worry with this argument is that the conclusion is understated. It is not just expressivism that requires a standpoint disengaged from a particular normative commitment in order to be understood. Any theory that seeks to stand outside a particular normative commitment and give an account of what is involved in holding that commitment, without asserting the commitment itself, seems to require such a standpoint. Indeed, this 'standing outside' of normative commitments and 'placing' them in a wider understanding of the beings who hold them is one of the most commonly accepted desiderata of meta-normative theories. 2 In so far as Lynch's argument suggests that, for the case of (TG), this desideratum cannot be satisfied, it counts against all metanormative theories of (TG) that accept the desideratum of placement. This would include not merely expressive views, but realist views (according to which normative commitments are representations of a genuine normative reality) and error-theoretical views (which agree with realists about the nature of the commitments, but disagree about the genuineness of normative reality). In other words, Lynch's argument, if successful, counts against all 'external' or 'Archimedean' views of normative commitments, that is, views that provide 'commentary on normative thinking, concepts, and their truth-makers that [isn't] part of normative thinking itself' and therefore that 'purport to stand outside a whole body of belief, and to judge it as a whole from premises or attitudes that owe nothing to it'. 3 Archimedeans, including expressivists, need not worry unduly, however, since Lynch's argument is not successful. It fails because premise (3) is false.
To motivate this attack, first consider the intuitive implausibility of conclusion (4). This says that expressivism cannot give an account of the commitment expressed by a sincere utterance of (TG). But expressivism, as Lynch admits, is just the view that such utterances express attitudes, so the expressive account of (TG) is just that sincere utterance of (TG) expresses a pro-attitude to the state of affairs of believing (within a suitably defined range) all and only propositions that are true. This account may be implausible, but it is not unintelligible, as Lynch's conclusion claims. Something must have gone wrong. I suggest the problematic premise is (3).
What is the 'disengaged' standpoint that expressivism (or any Archimedean view) requires? Lynch does not give a general account, but does give two examples: of a morally disengaged standpoint and of an epistemically disengaged standpoint. First the morally disengaged standpoint is the stance from which the expressivist wishes to 'give a story about how ethical thought functions' (85).
Further, such a standpoint is distinct from the stance we take when we 'employ evaluative language' (85). As an example, a naturalistic view of human beings, which sees them as 'frail complexes of perishable tissue' (Blackburn 1998: 48) is commonly regarded as a disengaged standpoint from which we may hope to understand our moral commitments. Insofar as a naturalistic view of human beings does not seem to involve making any moral judgements about them, it is a morally disengaged standpoint.
Since morality is just one example of a normative domain (epistemology another) we can generalise this account as follows. A standpoint is normatively disengaged insofar as (i) it attempts to explain what it is hold particular (type of) normative commitment and (ii) that explanation doesn't itself make any explicitly normative claims (of that type).
Given this general account of a normatively disengaged standpoint, what of the possibility of an epistemically disengaged standpoint, that is, a 'standpoint where we can explain our epistemic evaluations without engaging in them' (86)? Here the relevant epistemic evaluation is (TG). Lynch argues that an epistemically disengaged standpoint is not possible. But in fact there is a buried ambiguity in the notion of an epistemically disengaged standpoint that undermines Lynch's argument. We can distinguish: (A1) The epistemically disengaged standpoint is a standpoint from which we can explain our commitment to epistemic evaluations (e.g. our commitment to (TG)) without thereby being committed to any such evaluations ourselves.
(A2) The epistemically disengaged standpoint is a standpoint from which we can explain our commitment to epistemic evaluations (e.g. our commitment to (TG)) without thereby invoking any such evaluations in that explanation.
(A1) is a stronger condition than (A2) and arguably cannot be met. The simple reason is given by Lynch. To attempt to explain our commitment to epistemic evaluations is a form of inquiry; to inquire is to figure out what to believe (89); and to believe is to accept the norm of truth for belief, namely that a belief is correct iff. it is true (79). Hence to inquire is to accept the norm of truth for belief. From here it is but a 'quick step' (83) to commitment to (TG). In short, to seek an account of the nature of our commitment to (TG) is to seek a good account, and that in turn is to accept that a good account is one that is true-which is to accept the claim (TG). Hence, if an epistemically disengaged standpoint is one that seeks to explain the nature of our commitment to (TG) and yet does not involve making that commitment then Lynch is right that such a standpoint is impossible.
It is important to note that the proceeding argument only mitigates against the possibility of an epistemically disengaged standpoint in sense (A1). If, instead of being interested in an account of (TG), we were interested purely in account of moral evaluations, the corresponding condition could be met. That is, compare (A1) to (A1M): (A1M) The morally disengaged standpoint is a standpoint from which we can explain our commitment to moral evaluations without thereby being committed to any such evaluations ourselves.
The naturalistic standpoint, previously mentioned, seems to meet this condition (as well as meeting the relevant counterpart of (A2)). The naturalistic standpoint is one from which some (expressivists) seek to explain the making of moral claims without thereby committing themselves to any such claims (or employing those claims in their explanations). So Lynch has correctly identified a disanalogy between an expressivism focused on moral claims and an expressivism that seeks to explain epistemic evaluations, such as (TG) (the latter category would include the theoretically virtuous 'global' versions of expressivism that seek to explain all normative claims). In the former case, an (A1)-style disengagement from the relevant normative standpoint is possible, in the latter case it is not.
But although it is true that, for an expressivist who seeks to give an account of epistemic evaluations, a disengaged standpoint in sense (A1) is not possible, this is unproblematic because another understanding of that standpoint is possible, namely that given by (A2). This point by itself is sufficient to undermine Lynch's argument, which relies on the claim that such a standpoint is an 'illusion'. However, it is also possible to argue independently for the conclusion that it is only in sense (A2) that expressivism requires an 'epistemically disengaged standpoint'. More particularly, in the case of providing an account of the claim (TG), it can be argued that all expressivism requires is that it is possible to give an explanation of our commitment to (TG) that does not itself invoke (TG) in that explanation.
The first argument for this last claim has already been noted above. Given that (A1) cannot be met by any Archimedean theory of epistemic evaluations, let alone expressivism, it is unreasonable to demand it. (To repeat, this is not to deny that the corresponding condition for a local moral expressivism-(A1M)-can be met. The point is rather that insisting on (A1) rules out any meta-normative account of epistemic evaluations, expressivist or otherwise.) The second argument can be extracted from Lynch's own characterisation, quoted above, of the 'morally disengaged standpoint'. Expressivism seeks to 'give a story' about how normative thought functions, that is, an explanatory story. There is of course a distinction between the content of an explanatory story and the commitments one acquires in the process of questioning, telling or accepting it (roughly, in the process of inquiring). The commitment to (TG) is (I am assuming) a commitment one acquires in the process of questioning, accepting or telling such a story. But it does not follow that (TG) is part of the story itself. Consider an analogy. In putting forward a defence of a scientific theory a scientist will thereby be committed to certain norms of truth and evidence, such as that a belief is correct iff. it is true and that one ought to believe something to the extent that the evidence supports it. But these normative commitments are not themselves part of the scientific explanation. It is not part of the explanation of why precipitation occurs when clouds are pushed over high ground, for instance, that one ought only to hold true beliefs. To hold otherwise would lead to the uncomfortable conclusion that all scientific theories are partly normative in content. Analogously, although anyone putting forward expressivism as a viable theory of our normative commitments will be interested in whether or not it is true-hence committed to the norm of truth-this norm is not itself part of the expressivist explanation. To be complete that explanation needs only explain, without employing normative terms, what sort of thing the normative commitments are, and how the practice of forming them might have arisen and perpetuated itself (both, of course, in the context of whatever worldview one is prepared to defend). It is not required that this explanation be such that one can accept it without having any normative commitments. More abstractly the point can be put by saying that there is a difference between the commitments to normative claims one acquires in the act of considering whether to accept an explanatory theory (even a theory of normativity) and the normative claims of the theory itself. Commitment to (TG) may be of the former type, but (TG) itself is not of the latter type, and that is enough to make conceptual space for 'disengaged' accounts of our commitment to it. Hence all that is required for expressivism to be a 'disengaged' view of normativity is that it explain our commitment to normative claims without employing normative claims in that explanation (clause (ii) above). More particularly all that is required for expressivism to be a disengaged view of normative epistemic claims is that it explain our commitment to those claims without employing any such claims in that explanation-which is what (A2) says.
Finally, it is worth disarming a potentially misleading train of thought that might suggest interpreting the notion of an 'epistemically disengaged standpoint' in a sense that includes (A1) rather than just (A2). Consider again a local expressivism that seeks to explain only moral evaluations. In this case, the naturalistic standpoint (for instance) is normatively disengaged in both the (A1) and (A2) senses: by taking it one is not committing oneself to any moral evaluations, nor do such evaluations play a role in any explanatory story to be told from such a standpoint. This might lead one to think that in all cases, what is required is a standpoint that is disengaged in both the (A1) and (A2) senses. But this generalisation is unwarranted. Even in the moral case, what is important is the possibility of an explanation of our commitment to moral claims that doesn't itself invoke any moral claims. It is a happy coincidence that in this case one can also embrace such an explanation without being committed to any moral claims. In the case of epistemic evaluations, though we cannot embrace any explanation of them without being committed to some epistemic evaluation (that is, (A1)-style disengagement is impossible) what matters is that the content of the explanation is evaluationfree. Excessive focus on the moral case, where (A1) and (A2)-style disengagement are generally co-extensive, can lead to insensitivity to what really matters (namely, (A2)-style disengagement).
Thus although there is a sense in which (3) is true-sense (A1)-it is not the sense that Lynch's argument requires. In the sense of 'epistemically disengaged standpoint' that expressivism requires-sense (A2)-premise (3) is false.
These reflections also allow us to say where Lynch's argument for (3) goes wrong. The argument is as follows: (3a) 'if we are to make sense of an epistemically disengaged standpoint, we need to consider the possibility…that someone could engage in inquiry without being committed to (TG)' (86). (3b) It is not possible that someone could engage in inquiry without being committed to (TG) (86).
This argument fails because (3a) is false. To make sense of an epistemically disengaged standpoint we need only make sense of an explanatory theory of our commitment to (TG) that doesn't involve the claim (TG) itself. This doesn't require imagining an agent or inquirer not committed to (TG), only imagining an explanation of that commitment that doesn't feature (TG) in the explanans. Thus Lynch's argument for premise (3b)-the example of King George (86-90)-is beside the point.
What can we learn from this? First, the Archimedean project of giving an external grounding for normativity is consistent with the claim that there are some normative commitments that all inquirers have. Should it turn out, for example, that there are some normative commitments that are constitutive of all inquirers (or even all agents) that would not determine how to understand those commitments. Generally, that a normative commitment is constitutive of inquiry, agency, humanity or whatever doesn't determine how to understand that commitment.
Second, Lynch's argument is not completely undermined. In his conclusion Lynch claims: Our reflections suggest that unlike our other values we cannot sufficiently abstract from the value of truth in order to be skeptical about it … What [this] tells us about value is that we cannot…take a skeptical attitude towards all of our values (95).
Nothing I have said here undermines this conclusion. That is, for all I have said it may be that it is impossible for us to cease to value truth and still count as inquirers. That is, it may be that as inquirers we cannot cease to have the normative commitment expressed by (TG). Lynch goes on: In particular, we cannot take an expressivist attitude towards the value of truth (95).
This does not follow from the previous conclusion and for the above reasons can be rejected. From the fact that a certain normative commitment is constitutive of inquirers nothing at all follows about how to understand that commitment. Further, expressivism is not-as this passage suggests-a skeptical attitude towards the value of truth. It is an account of what it is to value something such as truth (or loyalty or friendship or counting blades of grass). To say that we cannot, while remaining inquirers, cease to value truth does not settle the question of what it is to value truth. It is the latter question to which expressivism is addressed and, for all Lynch argues, remains a coherent account.