What is Normativity?
March 21, 2011
Another very quick and far too telegraphic post: the topic is – what do we mean when we talk about ‘normativity’? I take it that questions of normativity relate to what ought to be. Normative judgements are judgements of right and wrong. If we take there to be a right action and a wrong action, a right state of affairs and a wrong state of affairs, we are dealing with norms.
Normative judgements are classically contrasted with positive or descriptive judgements – where positive judgements pertain to matters of objective fact, as opposed to normative matters (matters of values). One of the things Brandom believes he has achieved in Making It Explicit (and I agree with him) is the demonstration that positive judgements are derivative of normative judgements – in explanatory terms, normative judgements (and normative practices) are more basic. This ought, I think, to be relatively intuitive and commonsensical: a statement of fact only makes sense as a statement of fact if we can be right or wrong about it – that is, if there is a good and a bad way of making a claim of this kind. Truth is also a norm. Nevertheless, Brandom’s order of explanation – and his collapsing of positive judgements into normative judgements (as part of his collapsing of representationalism into inferentialism) is controversial.
The main point I want to make in this short post, however, is that ‘normativity’ is a very broad category. Normative practices and normative judgements include any practice or judgement that can be right or wrong, good or bad. The senses of ‘right or wrong’ and ‘good or bad’ that are in operation here are not limited to the ethical. Thus there are a set of right ways and wrong ways to tie one’s shoelaces, or to swallow food. (That is to say: there are right ways according to the demands of such-and-such a group of socially(-biologically)-enacted norms – as we will see, normativity is always, for Brandom, to be understood in social-perspectival terms, and there can always be both contestation and contextual shifts around any given normative requirement). Similarly, there are a set of right ways and wrong ways to handle a weapon if you want to torture somebody. When we are talking about norms, we are not talking about intrinsically admirable things, or intrinsically important things – we are talking about a basic resource of human judgement and practice, regardless of whether those practices are good or evil, ethically freighted or banal: to participate in a normative practice simply involves having the ability to regard one way of doing things as right, and another way of doing things as wrong, irrespective of the substantive content of ‘right’ or ‘wrong’ (and of whether we ourselves agree that the ‘right’ way is in fact the right way, in any of the relevant senses of that term.)
It’s perhaps also worth saying that there is no fundamental distinction to be found in Brandom’s system between ethical and instrumental reason: at the meta-theoretical level at which Brandom’s work is pitched, there is no classificatory principle that would enable us to distinguish between intrinsic means and ends (I’ll try to expand on this in future posts). Valuing liberty, and selecting a five dollar bill to pay for your cola, are both equally normative in the relevant sense (and I think Brandom is right about this, too).
This last claim can sometimes worry people: it is sometimes suggested that if the ethical distinctions we value most highly cannot be located at the very well-spring of normativity, our values have been robbed of all legitimacy, and ethics (or reason) must be given up as lost. This concern is, in my judgement, an unwarranted one. As I (again) plan to discuss further in future posts, the Brandomian metatheoretical flattening of the (at times) desired hierarchy of norms does not result in our inability to assert those norms (or indeed assert the hierarchy) – it simply locates such an assertion within the same normative field we are analysing, and at the same metatheoretical plane as those normative judgements we wish to contest. In this way we lose the aura of metaphysical privilege we may have coveted – but we do not lose the norms.
Apologies for the scrappy and underdeveloped nature of this post – I’ll try to address all of these themes in more detail when I have more time.
duncanlaw.wordpress.com 
March 21, 2011 at 8:38 am
Hi Duncan,
I guess one could object to Brandom that truth is a semantic and not a normative notion. There’s a difference arguably between saying that something is ‘right’ – as in, ‘This is the right way to proceed’ – an saying that a statement is right (i.e. ‘true’). The first implies a normative claim directly – one ought to proceed so. The second statement is metalinguistic not normative. It assigns a semantic property to the statement (that one can analyze in various ways, obviously). But it implies nothing about how one ought to act on its own.
One could object that saying ‘s is True’ (where s is some sentence) implies that the normative claim ‘One ought to believe that s’. But this entails that for any s ‘s is True’ implies one ought to believe s, which doesn’t seem to be an observed norm. Clearly, to accept ‘s is true’ implies that one believes that s, but that is just a conceptual consequence of our notion of belief, arguably.
So I’m not convinced that there’s any prescriptive content to the notion of truth. Clearly, we can have norms involving truth – e.g. one ought not to believe a contradiction. But the fact that a certain property features in normative claims does not make it intrinsically normative.
Regards,
David
March 21, 2011 at 11:55 am
David,
I’m a little out of my league here, but it seems your objection here fails to take stock of Brandom’s project as a whole, according to which the sharp distinction between semantics and normativity is untenable. That is because Brandom attempts to show that semantic content is to be understood on the basis of normative statuses, themselves understood in terms of the practical expression of normative attitudes. In other words, the semantic content of a proposition just is the set of rules governing the appropriate circumstances under which it used, and the appropriate consequences that should follow from its use. The determination of this propriety is itself understood in terms of the social practices of taking or treating usage as proper.
In that context, your claim here doesn’t make sense: “It assigns a semantic property to the statement (that one can analyze in various ways, obviously). But it implies nothing about how one ought to act on its own.” Semantic properties, for Brandom, are nothing but the implicit and explicit normative attitudes of a linguistic community about how one ought to act, or more specifically, how one ought to use certain propositions.
If you have problems with Brandom’s argument for a normative-pragmatic account of semantic content, fair enough, but you can’t take an opposing perspective for granted without showing why it is superior to his. You certainly must at least address why you find this approach objectionable, considering that Duncan is writing in the context of interpreting Brandom. Otherwise, your comment misses the point.
A final note: how do you make sense of the concept of truth without construing it in normative terms? Truth is just a form of correctness, the ‘ought’ corresponding to the ‘ought not’ of falsity. If truth is not the exemplar of normative notions then what is it?
March 21, 2011 at 1:56 pm
Hi reidkane
I’m no expert on Brandom’s work so any critique I can offer here is programmatic at best.
First, I guess we need to distinguish different kinds of semantic property here. A set of rules governing correctness seems like a fair gloss on ‘semantic content’. I’ve no deep objection to a use theory of meaning – though here I might cite Davidsonian objections to the idea of language as a set of rules, at all. Maybe a theory of understanding is more basic than a theory of use. But lets allow, in some sense, that usage fixes what we want to call semantic content.
However, if by truth we mean the semantic property of sentences fixed under an interpretation function which fixes a model for the language then it seems pretty obvious that inferential role cannot determine truth. This obviously apparent in areas like set theory where – as the Lowenheim-Skolem theorem shows – the meaning of concepts like ‘countable’ is not invariant between models of the same axiomatic system (basically, LS shows that a theory with an uncountable model also has countable ones – so what does ‘uncountable’ mean in a theory interpreted onto a countable model? – Putnam famously developed this argument in ‘Models and Reality’).
Similarly, it seems perfectly reasonable to suppose that the inferential roles of terms in a language can’t fix the intended model of the theory. If that’s right, then inferential role cannot suffice to determine truth conditions if we understand truth in model theoretic terms.
Now, maybe that’s not a problem. Maybe we want to construe truth differently as something like warranted assertibility. But this seems to commit us to some kind of anti-realism since it seems to debar the claim that some sentences have verification-transcendent truth conditions.
I guess my wider worry with Brandom, is that taking normativity as basic is difficult to square with naturalism and places restrictions on the class of possible cognitive systems at seem prima facie very strong.
Best wishes
David
March 21, 2011 at 3:32 pm
Hi Guys,
I’m just going to pop in briefly to recommend some relevant literature. First, Brandom’s most concise (and brilliant) discussion of his views on truth and the ways it can be used as a semantic notion (i.e., in model-theoretic truth conditional semantics of the Davidsonian and possible worlds kinds) is to be found in his essay ‘Why Truth is not Important in Philosophy’, which is essay number 6 in Reason in Philosophy. Chapter 5 of MIE comes a close second to this.
On the point about model-theory, it’s pretty inert in relation to Brandom’s system insofar as he rejects model-theoretic semantics in favour of something more proof-theoretic, specifically, his inferential semantics (MIE) and its formal counterpart in his incompatibility semantics (chapters 4&5 of Between Saying and Doing). I also seriously recommend Jaroslav Peregrin’s work on inferential semantics, which extends Brandom’s ideas in a slightly different formal direction (focusing more directly upon proof-theoretic notions of inference rather than incompatibility), which deals quite directly with the logical problems that supposedly plague proof-theoretic semantics, such as Godellian/Tarskian completeness problems.
On the point about warranted assertaibility, Brandom does indeed adopt something like this, influenced as he is by Dummett, but he goes to great lengths in chapter 8 of MIE (summarised in chapter 6 of Articulating Reasons) to show that his model is not susceptible to the objections that plague Dummett’s (and Crispin Wright’s) models. This is done in two ways: i) by maintaining a distinction between two different normative statuses that sentences/propositions can have – commitment and entitlement – and differentiating between the content of claims such as ‘P’ and ‘P is warranted’ on this basis, and ii) by developing a perspectival model of truth and knowledge that enables him to underwrite a difference between what we take to be the case and what is the case, which he calls ‘objectivity’, although it is really only the intersubjective necessary condition of objectivity, which can then be built on top of it by supplementation with Sellars’ account of perception.
I’d really recommend talking a deeper look into the details of Brandom’s semantic picture. It’s often dismissed by people who think it can’t deal with problems that he explicitly addresses, without taking a look at how he aims to address them. It’s also a very logical successor to both Davidson’s and Dennett’s interpretational approaches to intentionality, but it provides a much richer account of semantic content than Davidson’s relatively bare disquotational approach can muster, while staying true to the Quinean pragmatic insights that inspired it (as well as to the parallel insights of Wittgenstein, Sellars, and Dummett).
Best,
Pete
March 21, 2011 at 4:08 pm
Hi Pete,
That’s a most helpful recommendation. I’m just re-reading the Putnam material on Model Theory (takes me back to undergraduate phil-maths days)as a prelude to getting back to reading Brandom. I’ll definitely look up that article and respond at some point. cheers as always.
March 22, 2011 at 12:17 am
Wow, this is a great conversation, thanks guys! I’m rushing now – apologies – but to be honest I can’t think of a thing to add to Reid and Pete’s articulation of the Brandomiam position.
March 22, 2011 at 12:29 pm
Ok, sorry to be late to the conversation :-(. These remarks will be almost completely superfluous: basically, as I said earlier, I agree with Reid and Pete, and their comments probably articulate the Brandomian case better than I will. For the sake of working this stuff through myself, though – I think this is especially important:
Yes. For Brandom semantics can be fully cashed out in terms of pragmatics – and MIE aims to make that argument in detail, starting with normative social practices, constructing an inferentialist semantics using the pragmatist resources, and then using the inferentialist semantics to generate a concept of objective reference (which legitimates truth-talk). I haven’t worked through Brandom’s linguistic philosophy in anything like the detail I need to, but as Reid says, one of the things that makes Brandom’s work so impressive is that MIE aims (convincingly, to my mind) to have actually done the work: not just “it’s possible to explain objective reference in inferentialist terms”, but also “here’s how”.
It seems to me that David’s core unease about Brandom’s project is expressed here:
I think this is wrong – I agree that Brandom is committed to cashing out truth-talk in terms pretty close to ‘warranted assertibility’ (though his account is obviously complex) – but I don’t think this commits him to an anti-realism, because Brandom’s account of implicit norms can make us (implicitly) committed to things that no one actually thinks – and, in principle, to things that no one will ever think. Brandom’s concept of objectivity is structural: because we’re aware of the possibility of a gap between what a person believes and what they should believe (from someone else’s point of view, and/or given the first person’s implicit commitments), we can leverage this idea to generate a generic concept of “that which we ought to believe but don’t”, or even (I guess) “that which an entity capable of possessing knowledge and reasoning capacity we may never in fact possess ought to believe”. This latter speculative direction isn’t of much interest to Brandom, I don’t think – but the point is that Brandom can have truth criteria that are not limited by the actual beliefs or verification-capacities of a specific community, even though those criteria are generated by the community. (Also, the possibility of verification is not the only kind of legitimation for belief available – Brandom’s work is metatheoretically agnostic as to what counts as a good reason – so you could be a Brandomian verificationist, but that’s not required by the metatheoretical system.) So an object of reference isn’t understood as what we think it is (or can test it as) – it’s understood as what we ought to think it is (even, in principle, if we don’t and perhaps even can’t know that we ought to think this). Brandom would argue (and I think he’s right), that there’s nothing more to the concept of truth than some suitably elaborated version of that set of normative demands. But I don’t see this as anti-realist in any way.
On this –
– I’m sympathetic to this impulse, since this was my first reaction to Brandom too. Nevertheless, I’m now convinced it’s wrong. One of the main things I’m trying to do in this series of posts is explain how Brandom’s philosophy is strongly compatible with naturalism. The issue is that although in a certain sense Brandom thinks that normativity is basic, he only thinks it is basic ‘for us’. We are intrinsically normative creatures, because if we weren’t normative we wouldn’t be talking and thinking like this at all. Language is an intrinsically normative endeavour, for Brandom (and as Reid says, he gives (what are to my mind) good reasons for thinking this). So Brandom is very hostile to the project of attempting to construct a non-normative meta-language into which normative vocabulary could be translated. Norms are required to get any language or thought off the ground at all (for Brandom). This doesn’t mean, however, that once we’ve got our objectively descriptive language-games up and running, we can’t analyse naturalistically the social practices that generate those norms in the first place. So Brandom has it both ways: in one sense normativity is irreducible (we can’t get outside of norms, because we’re normative through and through), and in another sense norms are (potentially) analysable in naturalistic terms (natural processes generate the norms that make us, as sapient creatures). A lot of misinterpretation of Brandom results from confusing these two different ‘levels’ of his argument (imo). I should add that MIE is also compatible with non-naturalistic understandings of the relevant social practices – but that’s because the work doesn’t attempt the scientific (or non-scientific) analysis of the specific practices that generate Brandom’s general classes of practice (that question is simply outside the scope of the book), not because it’s aimed in a ‘non-naturalistic’ direction.
As I said in the original post, though, I’m not good at understanding the objections to this aspect of Brandom’s work. I share Reid’s impulse: how do you make sense of the concept of truth without construing it in normative terms? This is something that seems so intuitive to me that I suspect I’m not getting at what vexes people. And of course all this is very compressed.
April 22, 2017 at 6:40 am
[…] do with a society’s deciding what people should or should not do; in other words, it’s normative. There are many who try to downplay the relationship between law and morality, including this […]