So I promise I’m not going to do this for the whole book, because that would be extraordinarily tedious, but having just finished the first chapter of Brandom’s Making It Explicit I wanted to put up some very quick remarks on the issue of anti-naturalism. The main piece of text I’m going to address here is part 5 of section IV of chapter 1 – pages 42- 46 of the Harvard University Press paperback. But first a background-establishing passage from earlier in the chapter.

On pages 35-36, Brandom is explaining and endorsing an approach to the understanding of normativity that relies on the idea of sanctions:

The fundamental strategy pursued by such a theory is a promising one. As here elaborated, it involves three distinguishable commitments. First, Kant’s distinction between acting according to a rule and acting according to a conception of a rule is taken to express an important insight about the special way in which we are normative creatures. [Brandom is drawing the distinction between norms and laws of nature here.] Second, the pragmatist regress-of-rules argument is taken to show that in order to make use of this insight, it is necessary that the sort of normative attitude that Kant takes to play an essential mediating role in our government by norms be understood as involving implicit acknowledgement of norms in practice. Specifically, it is necessary to make sense of the idea of practically taking or treating performances as correct or incorrect. Third, taking or treating performances as correct or incorrect, approving or disapproving of them in practice, is explained in terms of positive and negative sanctions, rewards and punishments. This tripartite structure is endorsed and pursued in the rest of this work. There are reasons not to be happy with the regularist way of working it out that has just been sketched, however.

Without going into what the regularist way of working this schema out is, let’s flip forward to Brandom’s section 5, Normative Sanctions. Here Brandom writes:

Defining normative attitudes in terms of dispositions to apply sanctions does not by itself reduce the normative to the non-normative – it just trades off one sort of norm for another…. Benefit and harm, desirable and undesirable, are concepts that also have normative senses. Indeed, these senses would seem to be primary, so that some sort of reductive hypothesis would be needed to naturalize them. To turn the retributive story about normative attitudes into a naturalistic one, an account might for instance understand what is good (and so rewarding) in terms of what is desirable, what is desirable in terms of what is desired, and what is desired in terms of what is pursued.

Commitment to such a reduction is optional.

Let me note in passing to note how simplistic this example of naturalistic explanation is. (It’s just an example, obviously; this isn’t doing much work for Brandom – the anti-naturalist arguments (if such they are) come later; but I want to note that this isn’t the sort of thing that a lot of people would mean by ‘naturalistic explanation’. Naturalism has the resources of all the natural and social sciences at its disposal – naturalistic accounts of normativity can be vastly more complicated than this.)

Brandom then discusses a fictional case study (in typical analytic philosophy style, he uses imaginary primitive tribes as the standard illustrative example: some seriously problematic cultural politics are behind this disciplinary habit, but it’s hardly specific to Brandom), in which “to enter a particular hut one is obliged to display a leaf from a certain sort of tree” (43). How are sanctions implemented, here? One way is that “one who violates the norm is beaten with sticks”. Beating someone with sticks is “describable in nonnormative terms”. Alternatively, “one who violates the norms is not permitted to attend the weekly festival. In such a case, the normative significance of transgression is itself specified in normative terms (of what is appropriate, of [what] the transgressor is entitled to do).”

Brandom goes on to write that:

In the cases so far imagined, these webs of norms linked by internal sanctions are anchored, as each chain of definitional dependence terminates in some normative status that is definable independently, by external sanctions specified in nonnormative terms [e.g. beating someone with sticks]. Even this restriction can be relaxed. The consequences of an assessment of a performance as correct or incorrect with respect to one norm may extend no further than other assessments of correctness, with respect to other norms… Such an interpretation would not support any reduction of normative status to nonnormatively specifiable dispositions, whether to perform or to assess, whether individual or communal. (44)

Now this is where I want to protest, in a somewhat conditional way. I possibly agree with the letter of this passage, depending on how one interprets the content of the last sentence – but if this argument is meant to be an anti-naturalistic one (and this is sort of my operating assumption – one of the reasons I’m reading Making It Explicit is that I want to understand properly what Pete means when he refers to ‘Brandomian anti-naturalism’) it simply doesn’t work, and seems (again if this is its intention) to be motivated by a strange confusion. (Strange I mean given how sophisticated and impressive the rest of Brandom’s text is.)

I don’t doubt that this argument is a legitimate critique of some philosophical positions. If, for example, a theorist is claiming that ‘normative’ sanctions must ultimately be referred to ‘non-normative’ sanctions (like beating people with sticks) in order to explain every aspect of normative sanctions’ content – if a theorist is claiming this, it is obviously an adequate rebuttal to demonstrate the possibility of there being at least some self-enclosed webs of normative sanctions, where further normative activity is itself the content (or partial content) of the sanctions, ad infinitum. It’s useful to note that this can be the case. But this is not a rebuttal of naturalism. The naturalistic claim – or at least the naturalistic claim I’m familiar with and want to promote – is that normative practices are themselves natural phenomena (as are the norms implicit in them), irrespective of whether the consequences of transgressing norms are ‘physical’ or ‘cultural’ (because culture, too, is a natural thing, according to a naturalistic position). The naturalistic position is therefore untouched by arguments of this kind.

Reaching for a parallel here, one could perhaps compare the claim that consciousness is a natural phenomenon, emerging out of specific biological (mostly brain) structures in interaction with a broader environment. This claim doesn’t of course stand or fall on whether any given activity of consciousness can be related to the activity of nerve-endings of the eye, skin, etc., which we already know are natural things [with the apparent implication that maybe brain-states aren’t?]. Similarly, Brandom’s argument seems oddly to presuppose that normative sanctions are not, as normative sanctions, natural, or susceptible to naturalistic explanation – but that if they can be connected to activities we all already know to be natural (like beating people with sticks), this will serve to demonstrate the naturalness of normativity. But this is not how naturalistic arguments work.

I’m open to the possibility that I’m simply misreading Brandom here. As I say, this seems like a strange confusion for Brandom to be participating in, given how sophisticated and impressive Making It Explicit is in other respects. And I think the strategic intention of passages like the ones I’m quoting is probably to avoid having to bother with debates about naturalism at all – this isn’t Brandom’s area of interest. Still, this is the only candidate I’ve come across for an anti-naturalistic argument in Making It Explicit so far – I wanted to post on it, and explain why it doesn’t successfully perform this function (if indeed it’s even meant to), before moving into the depths of inferential semantics.


So, I’ve been having a conversation, over at ktismatics, with Pete Wolfendale of deontologistics. This post is a continuation of (my side of) that conversation. The basic issue is the status of norms – Pete adheres to a ‘Brandomian anti-naturalism’; I think that norms are natural phenomena. What are the differences between our positions, and why do I think Pete’s wrong?

Pete has various things going on in his discussion of normativity, I think. The first is a methodological insistence on the primacy of norms in terms of rational discursive claims. I’m basically 100% on board with this. Any judgement we make is going to be guided by norms of judgement – one can’t rid oneself of normativity in order to attain an objective understanding which can then be used to ground normativity – the idea’s incoherent. Objectivity is acquired via commitment to certain norms. If this is part of the thrust of the Brandomian emphasis on inferentialism, I’m all for it.

But Pete is making a stronger claim than this, I think. The basic naturalistic claim (the claim I’m arguing for) is that norms are emergent properties of natural phenomena (where ‘natural’ contrasts with ‘supernatural’ and ‘metaphysical’, rather than with ‘social’ – social and cultural phenomena are natural phenomena, in this sense of the term). Pete thinks there’s something more to norms than this. What could it be?

Here are some extracts from Pete’s long and dense post Dissecting Norms. I’m skipping over a lot of significant content in order to get to stuff that I regard as core, which is in Pete’s section 5, Rules, Norms and Practices. (I’ve already more or less articulated this argument in the ktismatics thread, but when I did so I hadn’t read much of Pete’s work, and didn’t say it right, so this articulation supersedes that one!) Here’s Pete, using Wittgenstein and Brandom:

arguments about how to interpret a rule ultimately regress to arguments about how to make explicit the norms implicit within our practices. These arguments then appeal to facts about what has been done to make claims about what should be done. The irreducibility of norms to practices consists in the fact that the former facts always underdetermine the latter. This is not to say that there are no good reasons for picking one interpretation over another, but rather than this process of interpretation is inherently open ended. There will always be facts which give us reason to exclude possible interpretations, but there will always be points at which we have to make decisions about how to apply a rule by making appeals to some selective reading of what has previously been done.

Pete goes on to give the example of case law, and writes that case law…

…is a different kind of rational process than the process of description which goes on in the sciences. Even though experimental evidence always underdetermines which scientific theory is right, we can always develop more tests, and get more evidence. The world is fully determinate, and it will always help us choose between theories as long as we can work out the right question. Interpretation on the other hand does not have the kind of plenitude of evidence which is bequeathed to description. For instance, in interpreting the intentions of the writers of the US Constitution, we will always be confronted at points where we have to make a best guess, by marshalling an interpretational narrative gleaned from the historical facts available.

Now – this distinction, between interpretation and description, is in my opinion unsustainable at the level of philosophical generality at which Pete’s argument has to be pitched in order to work. There are a host of differences between case law and scientific investigation, but none of those differences (of institutional structure, object of discursive focus, etc.) can sustain the philosophical weight of the distinction between description and interpretation that Pete is using here as part of his attempt to expectionalise normativity. In fact, this distinction is incompatible with the methodological primacy of normative pragmatics and interpretation that Pete elsewhere seems to be using in his system. The whole point of the Brandomian/Wittgensteinian apparatus that Pete is deploying, as I understand it [NB: Please note that I’ve only just started reading Brandom, so my use of ‘Brandom’ here should generally be taken as meaning something close to ‘Brandom as inferred by me, largely from Pete’s presentations of his views’], is to insist that description is also a normative activity – in fact, is first and foremost something that can, in principle, be made explicit as interpretation.

Carrying on with section 5 of Pete’s Dissecting Norms post – Pete writes:

The reason I have been claiming that norms do not really exist, or that they are pseudo-beings (here), is precisely because the truth of interpretations of them cannot be assessed in objective terms. This indicates that there is no real, fully determinate thing underlying our talk about a given norm. This isn’t to say that talk about norms is not a matter of truth, simply that the kind of truth in question is not objective truth.

How does this understanding of objective truth, and the corresponding insistence on a “determinate” thing underlying and undergirding empirical description – how does this insistence square with a Brandomian methodological priority given to inferentialism over referentialism? What is the epistemological mechanism by which certainty of the determinacy of the objective object of knowledge is acquired, independent of the unconcludable inferentialist game of giving and asking for reasons? [NB: I think Pete’s answers to these questions probably involve a transcendental deduction of a quite substantive non-empirically acquired metaphysical realism – but I’ll cross that bridge if and when I come to it…]

Perhaps more to the point, this distinction just isn’t right as a way of distinguishing scientific investigation from (say) case law. Ktismatics, in the thread, pointed out the large extent to which scientific endeavour does not assume a determinate object of analysis, but is instead often probabilistic; I would add that science is in any case presenting hypotheses that function as models with greater or lesser predictive and explanatory power. This latter endeavour is a very different one from making metaphysical claims about the fundamental (necessary) structure of reality: science’s authority comes from the modesty of the status of its claims – the fact that they can never be anything more than “best guess”, and therefore must be meta-theoretically silent on the determinacy or otherwise of the empirical phenomena they relate to. [And, for that matter, there’s always the possibility of acquiring additional evidence about what the Founding Fathers intended, for example – just like in the ‘descriptive’ case.]

Further, and still more to the point, scientific investigation is fundamentally interpretive. Scientists acquire data (which is in itself a fully normatively guided activity: granting specific aspects of the non-human world the social ability to be referred to as authorities in a distinctively scientific way, is itself an extremely complex social, normative, interpretive practice) and then interpret the data, as part of a large-scale complexly institutionalised discursive activity. Pete seems to be implicitly relying here on a version of the myth of the given, whereby supreme court justices are (rightly) aware that the activities they professionally engage in involve ongoing rational contestation of rival claims and interpretations, but scientists can read their conclusions off the natural world without engaging in interpretation, open-ended rational discourse, disputation, or all of the other social practices of rational discourse that are here being used to distinguish the making of norms from the discovery of scientific facts.

My point here isn’t meant to be an anti-scientific one, obviously. There are of course lots of non-philosophical ways in which the interpretation/discovery distinction can hold up just fine. But the point of a meta-discursive theoretical apparatus of the kind that Pete seems to be deploying is that it operates at a level of generality and abstraction where these distinctions drop away, and reference to an objective world can (rightly, I think) be understood as of a piece with other more overtly-interpretive discursive practices in which the ability to refer to an objective world, and our understanding of the nature of that world, is constructed out of an ongoing discursive pragmatics. (A discursive pragmatics that can then be analysed naturalistically, using the scientific resources it itself makes available, I’m arguing.)

Now in a way all this is slightly beside the point. The real focus of Pete’s argument in this section is to distinguish practices from norms. The reason I’m talking philosophy of science, however hastily, is that I think the (imo faulty) philosophy of science that Pete is relying on in the passages I’ve quoted above – a philosophy of science whereby the object of scientific investigation is taken to be necessarily determinate, and in which scientific results are therefore, by implication at least, ultimately excludable from the ongoing process of interpretive negotiation and contestation that is otherwise part of the basis of Pete’s Brandomian deontologistics – … this philosophy of science is necessary in order for Pete then to assert the determinacy of the empirical world, in contrast to always-empirically-underdetermined norms – which is what gives him his anti-naturalism re: normativity.

Pete’s argument here is Wittgensteinian. We have a given set of practices that constitute the following of a norm – a canonical example from the Investigations is somebody who has been taught how to following the “+2” rule, and has demonstrated their mastery of the rule by writing out the number series 2, 4, 6, 8, 10, etc. Wittgenstein writes:

Now we get the pupil to continue a series (say +2) beyond 1000—and he writes 1000, 1004, 1008, 1012.

We say to him: “Look what you’ve done!”—He doesn’t understand. We say: “You were meant to add two: look how you began the series!”—He answers: “Yes, isn’t it right? I thought that was how I was meant to do it.”——Or suppose he pointed to the series and said: “But I went on in the same way.”—It would now be no use to say: “But can’t you see….?”—and repeat the old examples and explanations.—In such a case we might say, perhaps: It comes natural to this person to understand our order with our explanations as we should understand the order: “Add 2 up to 1000, 4 up to 2000, 6 up to 3000 and so on.”

Wittgenstein’s point here is part of a larger argument about regresses of interpretation – the rule for interpreting a rule must itself be interpreted, etc. etc. (resulting in the requirement for a pragmatic non-interpretational mode of rule-following). The way Pete is using arguments of this kind, however, is to draw attention to what he takes to be a special property of norms (or, really, I guess, to norms’ technical non-existence, and thus inability to have properties). Norms, Pete is arguing, are underdetermined by practice. Norms therefore cannot be identified with practice. Norms therefore cannot be understood naturalistically.

Assuming we are using ‘practice’ in a suitably broad way (such that our descriptive framework also takes into account the biological features of the human animal that enable and channel social practices, etc.), I think it’s clear where Pete’s argument falls down. Let me telegraph sufficiently: the following sentence is the point of this post, and everything else has just been a clearing of ground in order to make more explicit where it’s coming from:

Pete’s argument falls down because norms are no more underdetermined by practices than are practices themselves.

It’s a simple as that really, but just to elaborate: Pete seems to be assuming a fully determinate empirical reality that can be absolutely known in its determinacy, empirically. I think this is based on a faulty philosophy of science – it is in principle impossible to acquire such knowledge (my asserting this is not, of course, the same thing as my asserting that no empirical knowledge can ever be acquired – the dispute is over what constitutes ‘knowledge’, not over whether knowledge (including non-scientific knowledge) is possible). Pete then seems (reasonably) to deny this epistemological access with regard to the future practices that are intrinsically involved (given a Brandomian framework of open-ended rational discourse and rationally-motivated action) in the constitution of present norms as norms. Pete then seems to argue that this lack of absolute knowledge about future practices, is really a feature of the intrinsic indeterminacy of norms, versus the intrinsic determinacy of practices, even though the epistemological issue that’s doing the work here is in fact one related to practices, as the Wittgenstein argument makes clear: the problem in the Investigations passage quoted above appears when the pupils’ practices are different from those of the teacher; absent such possible divergence, what are the grounds for differentiating norms from practices at all?

That could all have been more felicitously phrased. But I hope the basic point has been communicated: I don’t see what legitimate grounds Pete can have for differentiating norms from practices (in a very general sense) in the way he does – which is also to say, I don’t see what compelling arguments Pete has against naturalism w/r/t normativity.


Now in the conversation at ktismatics I kept apologising for not having read enough of Pete’s work – and apologies are still due: I’ve now read the relevant posts from Pete’s very useful important posts page, but there’s still lots to read – notably Pete’s Essay On Transcendental Realism, which after all is what ktismatics was discussing to begin with. And then of course there’s Brandom. I obviously think it’s legitimate to analyse Pete’s work without having acquired the relevant Brandom background; but equally obviously, knowledge of Brandom’s system would be useful here. My excuse is simply that Making It Explicit is enormous, and I didn’t want to drop the conversation for the time required to read it. [I should also probably note, by the by, that there seem to have been a number of other blogospheric arguments about normativity raging recently. In case it’s unclear, this isn’t meant to be an intervention into those debates.]

Anyway, questions of due diligence notwithstanding, I still think this post is an adequate pass at the relevant section of Pete’s Dissecting Norms argument. I’ve already made clear that Pete has no obligation to respond to my remarks on his work, and let me reiterate that – Pete’s already been extremely diligent in responding to my comments in the ktismatics thread. But if you’re reading, Pete, and feel like replying, my question basically boils down to: what’s your response to the sentence of this post I’ve put in bold type? If you agree with it, how do you see it as compatible with your argument re: the status of norms? If you disagree with it, why?



It occurs to me that most of the post above is actually superfluous with regard to the basic point, which my original comment in the ktismatics thread possibly expresses more clearly. To re-iterate one more time, then:

Given that the under-determination of norms by practices comes (according to Pete’s argument (as I understand it)) from the fact that we can’t fully know in advance which future practices will be agreed to be compatible with these norms; and given that this indeterminacy is by definition also an indeterminacy of future practices; why is this indeterminacy taken as telling us anything specifically useful w/r/t norms as opposed to practices?

Hopefully I’ve now said that in enough different ways that it should be clear what my objection is.]