Brandom Thesis Plan

January 27, 2011

[I’m going to be talking from now on as if I’m writing a thesis on Brandom – this may very well of course not actually be the case, but I can’t be bothered to write ‘thesis-length-document’ every time I refer to the thing, so I ask you to bear with me.]

My own interest in writing on Brandom is related to social-theoretic (ultimately political-economic) analysis of contemporary society. Writing about Brandom is obviously drawing a rather long bow in order to address that issue, and I think one of the principal challenges that will be involved in writing this document will be doing so in a way that maintains its argumentative unity while also drawing some of the connections I want to between different disciplinary spaces. This shouldn’t be impossible – it’s obvious, just googling, that plenty of people have written about neopragmatism in relation to the meta-theoretical aspects of social science, so I should be able to slot into that discursive space without causing anyone to wig out too much. Still, in an ideal world I want the document to achieve a number of different things. At a first schematic pass:

1) Situate Brandom within a tradition that includes not just pragmatism and neo-pragmatism but also the canonical texts of the social theoretic tradition.
2) Articulate some of the problems that the social-theoretic tradition has classically faced, and position Brandom’s work as providing resources that can help resolve some of these issues.
3) Give an exposition of the structure and – in many (though not close to all) areas – detail of Brandom’s system. (This is a contribution to the interpretive debate around Brandom’s work – and will be the core of the document.)
4) [Maybe – if it fits] Argue that Brandom’s normative pragmatics is more dissociable from his linguistic philosophy than it is often taken to be (including, I think, by Brandom).
5) Show how the Brandomian resources expounded in the core of the document can resolve some of the problems set up in (2).

This is all subject to change – it may be too ambitious, and I’m mostly just trying to get my head around the general conceptual space the document needs to slot into, and its possible structure. If this is the general plan, though, there are several areas of reading I need to dig down into.

– The social theoretic canon. As I’ve said, I’m already ploughing through this in the out-of-sight blog boiler room, so this isn’t too vexing.
– The philosophical pragmatist tradition – 19th century to present day. This is a bit more troublesome, since I’ve not read most of the key figures here. But this is something I will have to address eventually, evidently.
– Brandom. The man has written heaps, and I’m obviously going to have to read lots and lots of it.
– Brandom-related stuff. The literature here will I think be acceptably small. Brandom’s a significant and frequently-discussed figure, but it’s not as if he’s Kant, and his major work was only published in 1994 – there’s only so much that can have been written on him.
– Stuff referred to in, and necessary to understand the context of, all of the above.

That sounds about right, I think. Not sure how I’ll approach all this as yet – I’m just articulating the ballpark of what needs to be done. I may yet heavily modify the plan.


With that articulated, I’m going to call an Official Brandom Break on the blog. I seriously need to attend to other things for a while. This stuff’s like crack.

Brandom-related Reading List

January 27, 2011

In my last post I finally got far enough into my discussion of Making It Explicit to give a first-pass articulation of what I take to be the core of Brandom’s argument about the origins of normativity. There’s a lot of exegesis still to do, but this seems like a good point to take a breather.

In case it isn’t obvious, all this Brandom stuff is meant to contribute to Part One of my megalomaniacal larger intellectual project – that is, this is meant to be part of an empirically adequate theory of social practice. Before I started writing on Brandom at length, I had the (in retrospect rather overoptimistic) idea that I could articulate my interpretation of Brandom in the format of an academic paper. Since it’s taken me circa 20,000 words just to get to the main argument, that may not be realistic. So I’m now thinking that I’ll try to get a scholarship at some point, and write this thing up as a doctoral thesis if at all possible. It may very well prove not
be possible, of course – but even if it isn’t that’s roughly the length of the piece of writing I think I need to be aiming for here, and so I want to start thinking about the end result in something like those terms.

The purpose of this post is just to serve as a repository for a list of texts that I probably ought to read if I really want to write this up in a long-form and academic way. I’ll drop more texts in as I run across them or they occur to me. If anyone has any suggestions, those would of course be more than welcome.


A list of Brandom’s publications is available at Brandom’s website here:

Books about Brandom:

Weiss, Bernhard & Wanderer, Jeremy (eds.) Reading Brandom: On Making It Explicit, Routledge, 2010.

Wanderer, Jeremy, Robert Brandom, McGill Queens University Press, 2008.

More Heath:

Heath, Joseph, Following the Rules: Practical Reasoning and Deontic Constraint, Oxford University Press, 2008.

I see that Jon Cogburn is teaching a course that involves a bunch of Brandom and the following two works:

Okrent, Mark. 2007. Rational Animals: The Teleological Roots of Intentionality. Columbus: Ohio University Press.

Macdonald, Graham and David Papieau. Editors. 2006. Teleosemantics. Oxford: Oxford University Press.

Much more to follow I’m sure.

The Origins of Normativity

January 26, 2011

In my last post I discussed the first half of Joseph Heath’s paper Brandom on the Sources of Normativity [pdf]. I argued that Heath has significantly misunderstood the nature of Brandom’s argument in Making it Explicit. Heath’s misunderstanding is, I argued, attributable to two things. First, Brandom’s rather confusing discussion of “internal sanctions” in Chapter 1 of MIE. Second, Heath’s unwillingness to accept the idea that pretty much the whole of MIE‘s argument is required in order to cash out the content of some of the book’s early categories. As I said at the end of this earlier post, Heath is frustrated that an account of the nature of normative practices – an account of what makes such practices distinctively normative – is not forthcoming in the book’s first chapter. Without such an account, Heath argues, the apparatus of MIE is built on non-naturalistic axiomatic foundations in a way that removes much of its potential explanatory power. Were Heath right in his analysis of the structure of MIE‘s argument, this would indeed be a serious problem. My claim, however – which I aim to begin to justify in this post – is that MIE‘s argument as a whole functions as precisely the explanation of what makes normative practices normative, that Heath thinks is missing from the book’s first chapter.

Heath’s misunderstanding is particularly striking because in the second half of his paper he presents his own ‘alternative’ to Brandom’s account, quoting a section from the end of Chapter 1 of MIE, and arguing that the ideas expressed in this passage could be repurposed as part of a more adequate account of the origins of normativity. Heath – a little oddly – doesn’t see that this is precisely what this section of MIE – along with the much later sections it foreshadows – is already doing. Heath’s ‘alternative’ is thus a much simplified version of Brandom’s own argument. For this reason, Heath’s discussion provides a useful way in to Brandom’s more complex account – serving both as a first-pass articulation of the argument, and a point of comparison that allows us to see the motivating factors behind MIE‘s greater complexity.

I will quote an important section of Heath’s paper at some length. Heath has just been expressing a desire for “a simple behavioural account of what it is to follow a norm”:

One strategy for developing such an account is suggested by the parallel that Brandom often draws between interpreting behaviour as intentional and interpreting practices as norm-governed. In adopting the “intentional stance” toward an organism, one is in effect choosing to use a certain vocabulary in describing its actions. The core elements of this vocabulary are the concepts of belief and desire. Thus in adopting the intentional stance one is choosing to explain the organism’s actions as goal-directed, and as guided by some set of representations pertaining to the achievement of this goal.

A theory that purports to explain intentionality in terms of the ascription of these states remains incomplete, however, insofar as it gives no account of the system that does the ascribing. One is inclined to think that any system which can adopt an intentional stance toward another must itself also be an intentional system. Brandom therefore distinguishes between “simple” intentional systems and “interpreting” intentional systems. But where does this “interpreting” intentionality come from? If one posits some further intentional system, whose ascriptions constitute the intentionality of the interpreting system, then a regress has been initiated. What is needed, in order to resolve this regress in a satisfactory manner, is some kind of source for all this intentionality. What is needed is some type of original intentionality.

The most mechanical way of resolving this regress would be to abandon the “stance stance,” and posit an intentional system whose intentionality is not inherited from some further system. John Searle uses such an argument to defend his view that some system must possess intrinsic intentionality. Brandom, however, wants to maintain allegiance to the stance stance. He rejects Searle’s move, choosing instead to argue that original intentionality can found, roughly, in groups where agents each ascribe intentional states to one another. He states this claim as follows:

The key to this account is that an interpretation of this sort must interpret community members as taking or treating each other in practice as adopting intentionally contentful commitments and other normative statuses. If the practices attributed to the community by the theorist have the right structure, then according to that interpretation, the community members’ practical attitudes institute normative statuses and confer intentional content on them; according to the interpretation, the intentional contentfulness of their states and performances is the product of their own activity, not that of the theorist interpreting that activity. Insofar as their intentionality is derivative – because the normative significance of their states is instituted by the attitudes adopted toward them – their intentionality derives from each other, not from outside the community. On this line, only communities, not individuals, can be interpreted as having original intentionality (61).

I would like to argue that this solution to the problem of the “origins of intentionality” provides a blueprint for a solution to the problem of the “origins of normativity.”

That last sentence is of course right – because Brandom’s solution to the problem of the origins of intentionality is intended to provide a solution to the problem of the origins of normativity: that’s what it’s for. Still, putting this interpretive problem aside, Heath’s discussion provides a useful summary of the issues at stake. In subsequent posts I aim to much expand on this analysis – but for now I want to make a few scattered and very brief remarks.

1) Brandom is here appropriating – in I think an exceptionally clever way – Dennett’s idea of the ‘intentional stance’. Dennett himself – like Heath – doesn’t seem to follow Brandom’s argument (I will try to deal with Dennett’s own interpretation of and response to MIE in a later post) – but be that as it may: Dennett’s idea is that there is nothing to being an intentional system (‘being an intentional system’ can be treated for our purposes as the same thing as ‘engaging in normative practices’) beyond being legitimately taken to be an intentional system by an interpreting agent. The problem then becomes – what is the status of the interpreter? How does this interpreter acquire their own (normative, intentional) interpretive capacity?

2) This argument and problem is, of course, parallel to that which we encountered in our discussion of regularism. According to Brandom’s normative phenomenalism, a practice is normative (it can be done right or wrong) if it is properly taken as normative (by, again, an interpreting agent). The question then (again) becomes, what makes the interpreting practice of normative attribution normative?

3) Brandom’s solution to this problem is to permit or acknowledge the existence of a potentially infinite chain of ‘takings as’. If we stop our chain of ‘takings as’ somewhere, and insist that we can in principle go no further in our chain of justifications, then wherever we choose to stop this chain is the location where practice becomes identical with normative guideline, and the distinction between natural law and moral law disappears (and this prompts the Ethical/Political problem of regularism which we ran into repeatedly in our previous set of posts). Brandom, however, insists that any such ‘grounding’ of the chain of interpretations is always in principle open to contestation, and thus the grounding practice is open to critique as falsely grounding, because divergent from real normative guidelines (justified by some other inferential chain). Because any grounding is always in principle open to such contestation, the ethical/political problem of regularism cannot get any purchase on Brandom’s meta-theoretical apparatus (even if it can be used as part of a critique of any given normative framework advanced within the terms of that apparatus.)

4) Yet because Brandom’s apparatus still does permit (and indeed requires) the grounding of justificatory chains in material inferences – moments where we say ‘this is simply what we do’ – Brandom’s apparatus is not vulnerable to the charge of vicious regress associated with regulism either. We can ground justificatory chains in actual practice – but only in those practices which the inferential web of normative attitudes we inhabit implicitly takes to be legitimate as material inferences – and this judgement is itself always in principle open to contestation.

5) This means that the ability of any material inference to serve as grounding normative practice must itself be granted (and thus instituted) (even if only implicitly, by the absence of contestation) by the practices of a broader community of normatively-sanctioning agents. And each of these agents’ own normative practices are themselves instituted as normatively sanctioning by similar (implicit or explicit) acts of taking-as. No practice is therefore intrinsically normative… and yet the normativity of the practices that attribute – and thereby (according to Brandom’s phenomenalism) institute – normativity to other practices can itself be understood as instituted in this way. In other words, Brandom’s system allows us to see normativity as instituted by (naturalistically analysable) social practices (and nothing else), without falling into the trap of regularism. And the way it does this is having ‘original normativity’ an emergent attribute of communities of mutually-recognising practice (as these communities are themselves recognised as such in practice), rather than any particular set of individual practices. This structure of Brandom’s argument allows him to close the explanatory circle – allowing him legitimately to claim both that “it’s norms all the way down” (because one never reaches an explanatory point at which any given regularity of practice can be equated with a normative guideline without an active (even if implicit) normative judgement permitting this identification) and that his account is a fully naturalistic one, explaining normativity entirely in terms of naturalistically analysable social practices (which it does).

I find Brandom’s argument completely compelling – it seems to me to be just brilliant and, in a way, wonderfully simple (despite its relative complexity, compared with other common accounts of the origins of normativity). In future posts I’ll try to elaborate on different aspects of this argument, and go into more detail about the various moves I’ve only gestured at above.

Joseph Heath on Brandom

January 24, 2011

I ended my last post by quoting from Joseph Heath’s paper Brandom on the Sources of Normativity [pdf]. Googling, I see that Heath is a professor of philosophy at the University of Toronto – his work seems to be partly focused on synthesising a Habermasian theory of communicative action with the rational choice discourse of modern economics. Heath also writes popular books, such as 2002’s The Efficient Society: Why Canada is as Close to Utopia as it Gets and 2005’s The Rebel Sell. I thought I’d use Heath’s paper as a jumping-off point for continuing my discussion of Brandom. This post is just notes on the first half of Heath’s paper.

Heath writes that Brandom

wants to argue that the concept of a social norm – a rule that determines, implicitly or explicitly, whether an action is correct or incorrect – can serve as a primitive concept in the development of a general theory of meaning.

But I don’t think this is right – Brandom’s theory is, in part, an account of the origin of norms – norms or rules are not ‘primitive’ for Brandom in any strong sense (though the reflexive nature of the argument means that norms are, at a number of points in the argument, treated as if they are primitive – but that’s because the full content of these concepts as Brandom understands them can’t be cashed out until later in the work).

In the paper’s introduction Heath complains that

Most of the detail in Making it Explicit consists in Brandom’s attempt to show that, if one is willing to take the concept of a social norm as given, one can similarly generate an account of inference, truth, reference, and ultimately representation. However, the mechanism used to achieve the latter is still relatively unexplored, and fraught with technical difficulties. In order to make it worthwhile to iron out the kinks in this mechanism, Brandom must provide the reader with some reason to think that it is somehow more plausible to take normativity, rather than representation, as a primitive. The first chapter of Making it Explicit, however, manifestly fails to achieve this.

As I say, this in my opinion fails to give adequate credit to the ‘eating its own tail’ quality of Brandom’s argument. Although Brandom’s argument of course builds sequentially, and I think the book is well structured to this end, the justification for a number of moves that are made near the start of the book can’t adequately be articulated until the argument as a whole is complete. Heath seems to find this vexing or implausible, but I don’t see any reason why such an argumentative structure is intrinsically problematic.

In sections I and II of the paper, Heath gives a useful summary of Brandom’s discussion of regulism and regularism, which I won’t summarise yet again.

In section III the paper starts going more badly wrong. Heath stumbles over Brandom’s discussion of ‘external’ and ‘internal’ sanctions, suggesting that –

Brandom does not explain exactly how having a fully internalized set of sanctions moves us away from the regularist account.

and that

he does not say much to allay the concern that a system of norms could not be sustained by internal sanctions alone

I’m sympathetic to these objections, since this is basically where I came in in my discussion of Brandom – I too initially baulked at the discussion of internal sanctions in Chapter One of MIE. I still think this section of MIE is one of the least satisfactory, but I no longer think this is a problem with Brandom’s general theoretical apparatus – I think it’s a problem in the articulation of that apparatus. I think that really Brandom is interested here (as often elsewhere) in the possibility of a nonnormative descritive metalanguage – when he says that it is ‘norms all the way down’, he means that our methods of articulation of conceptual content (which includes the descriptive propositions of a naturalistic account of norms) have to be normative (and thus part of a set of normative practices). But the ‘internal’ / ‘external’ sanctions distinction is a really bad way of articulating that point, if this point is indeed what Brandom’s after here. In fact the distinction between ‘internal’ and ‘external’ sanctions as articulated here is, I think, strictly speaking incoherent given that a large part of Brandom’s apparatus is devoted to an account of how causal chains can be mapped inferentially. No causal chain (to use Brandom’s example, beating people with sticks) is, on Brandom’s account, intrinsically external to a web of inferential connections, just as no causal chain is intrinsically internal. The question is simply how we map causal chains inferentially (or don’t), and this is a question that can only be answered from a social-perspectival position of such mapping (and acts of ‘taking-as’). Beating people with sticks can be an internal sanction, just as a mark on a scorecard can be an internal sanction – what matters for the distinction is how that sanction is treated in practice (or, strictly, how that sanction is properly treated in practice). I think Brandom neglects all of this because he hasn’t gotten around to ariculating these subtleties early in Chapter 1 – but I agree with Heath that this renders Brandom’s discussion of internal sanctions extremely problematic.

Heath adds:

There is also a problem explaining how a community could ever induct new members into a set of practices, if the sanctions governing those practices were all internal. One would need to show that a practice which is “norms all the way down” is also learnable. It is a conspicuous feature of the way that we initiate children, for example, into our practices, that we rely quite heavily on sanctions to signal approval and disapproval. Often these sanctions involve cooperating, or witholding cooperation – parents are always saying things like “no, I won’t pass it to you until you ask nicely for it.” These are external sanctions. It is difficult to imagine that we could do without these sorts of sanctions entirely, or that the practices in which they are used are not “genuinely” normative on that account

I don’t think this is right – saying “no, I won’t pass it to you until you ask nicely for it” can be an internal sanction too, since the parent’s ethical disapproval can be interpreted by the infant in normative terms, and this may be precisely where the force of the sanction lies. Nevertheless, I think Heath has a general point in this area. Training operates by generating new normative practices. Those new practices can be generated via sanctions. Those sanctions can be “internal”. But they can also be “external” in the sense that they form no part of the ‘student’s’ implicit set of inferential mappings. If a sanction presents as wholly arbitrary (i.e. not a sanction at all from the perspective of the student, but merely an inferentially inexplicable event) then it is ‘external’ to the student’s inferential chain. As soon as the student is capable of adopting a subject-position from within which the sanction functions as an actual (normative) sanction, however, the sanction is now ‘internal’ to the set of inferential connections pragmatically inhabitable by the student (even if the student doesn’t endorse that set of practices).

I need to address all this is more detail in a post on ‘Training’. For now I just want to register that I’m sympathetic to Heath’s dubiousness about the discussion of internal sanctions in MIE Chapter 1, but I think Heath is reading too much into the discussion, and misreading it as a result. Heath regards this discussion as foundational for Brandom’s system, and therefore woefully unsatisfactory. I regard it as a first pass at articulating a set of points that Brandom will elaborate in greater – and much more compelling – detail throughout the rest of the work, and that indeed can only be properly elaborated once more has been established in the way of theoretical resources.

In section IV Heath continues with his reading based on his interpretation of the ‘internal sanctions’ section. He suggests that Brandom

is not so much trying to provide an account of the action-theoretic underpinnings of normativity as he is offering a principled refusal to provide such an account.

This is really the only place to go if you regard Brandom’s discussion of internal sanctions as foundational for his system, rather than as a preliminary pass at a much more elaborate argument. And this view also results in seeing Brandom as much more strongly ‘anti-naturalistic’ than I think his work justifies. Heath continues:

The drive to “explain” normativity in terms of something more basic is, according to Brandom, guided by a desire (ultimately misplaced) to make norms “naturalistically respectable”.

Which is also to say that if (on Heath’s read) we adopt Brandom’s explanatory strategy we are also giving up the aim of rendering norms naturalistically respectable. (I of course disagree with this reading of Brandom.)

Heath then contrasts purely descriptive with normative vocabulary. He argues that Brandom’s case for his system ultimately rests on the fact that starting with norms as primitive allows us to account for description, but starting with description as primitive does not allow us to account for norms (because you can’t get “ought” from “is”). This is indeed a crucial part of Brandom’s argument. But (as I’ve already said) Heath misses the reflexive element of Brandom’s account. Heath refers to the greater explanatory scope made available (according to Brandom) by starting with normative practices, using Brandom’s term “expressive completeness”, and refers us to page 641 of MIE. But Brandom’s term “expressive completeness” does not just signify explanatory scope, but also signifies the self-embedding of a philosophical analysis, such that it is able to account for its own conditions of possibility. As Brandom writes on that page:

One of the criteria of adequacy that has guided the project from the outset is that it be possible to elaborate the model of discursive practice to the point where it is characterized by just this sort of expressive completeness. This means that the model reconstructs the expressive resources needed to describe the model itself. By means of these logical resources, the theory of discursive practices becomes expressively available to those to whom it applies. What is required is just that the scorekeeping practices that confer conceptual contents on the fundamental sorts of explicitating vocabulary used in stating the theory and specifying the content-conferring discursive scorekeeping practices in the first place be themselves specified within the terms of the theory.

This specification, I claim, involves for Brandom not just reference to primitively normative practices, but to the naturalistically describable practices by which those practices (which are, in fact, the very same practices) are taken to be normative (and so on down the line of ‘taking-as’) such that the circle of normative commitments and empirical description is capable of being closed in both directions. The other day (at David Roden’s Enemy Industry blog, specifically this post – I should have linked to this before, I now realise – apologies), I referred to Geoffrey Bennington’s description of Derrida’s theoretical intent: “the empirical is the transcendental of the transcendental (of the empirical)”. I see a similar (and, unlike – imo – Derrida’s, successful) explanatory loop at work in Brandom’s system. Norms can be understood in terms of empirical, naturalistically analysable practices. But naturalistically analysable practices can only be described from within a thoroughgoingly normative social-perspectival location and set of inferential commitments. The empirical practices that institute this social-perspectival location can in turn be analysed naturalistically. But such analysis can, of course, itself only be carried out from within a normative social-perspectival location. And so on and so on. This circle – a ‘virtuous’ rather than a ‘vicious’ one – keeps on spinning round, and ensures that there is no ‘foundational’ point for Brandom’s system in a strong sense – although for us normative practice is foundational, since we would not be subjects at all if we did not participate in and experience such practice (while we can perfectly well be subjects without a commitment to naturalistic modes of scientific analysis). I will go on to discuss the naturalistic understanding of the ‘brute fact’ of the (biologically evolved) social capacities that can institute normativity in a later post. For now I just want to make the general point about Brandom’s explanatory circle.

On my read of Brandom’s system, a strong naturalism (though one that includes rather than excludes analysis of social practice) is compatible with Brandom’s irreducibly normative pragmatics in the manner just described (the irreducibility simply comes from the impossibility of escaping the embededness of our naturalistic accounts in a social-perspectival set of normative commitments, not from a general opposition to naturalistic modes of explanation w/r/t normativity). On Heath’s read of Brandom, however, Brandom’s system has to be opposed to a strong naturalism. For Heath we can be either naturalists or Brandomians, but we cannot be both. Since Heath – reasonably enough – thinks that there are very compelling reasons for a commitment to naturalism (I of course agree), Brandom’s theoretical apparatus seems unappealing, and Heath’s read of the minutiae of Brandom’s system is correspondingly unsympathetic.


In section V of his paper, Heath’s emphasis shifts, and he begins to elaborate his alternative to Brandom’s perceived anti-naturalism. This is where the paper becomes most problematic: because Heath has mistaken the nature of Brandom’s account of the origins of normativity, when he comes to propose his ‘alternative’ Heath in fact proposes a much simplified version of Brandom’s own argument. I will address this in my next post.

I’ve now laid out in reasonable detail a set of theoretical problems that Brandom’s version of normative phenomenalism needs to resolve if it is to fulfil the explanatory tasks demanded of it. In the upcoming series of Brandom-related posts I will begin to try to explain exactly what Brandom’s solution to these problems is. Brandom’s solution is an unusually complicated one, even by the standards of analytic philosophy, so this will probably take a few passes, involving first a rough-and-ready account, then more details as we move further in to the complexities of Brandom’s system. I don’t expect this discussion to be complete for some time, and I’m going to need to take a break from Brandom posts soon to turn to off-line obligations – apologies if I leave all this hanging for a while.

To recap the last series of posts very quickly and schematically, Brandom’s naturalistic account of the basis of normativity needs to resolve two problems characteristic of ‘regularism’. The ‘Epistemological’ problem can be resolved (at least in a preliminary way) by giving a regularist account of the interpretive acts by which we pick out which regularities are generative of normative standards. This solution, however, throws us into the arms of the ‘Ethical/Political’ problem, in that it seems to give us no basis for criticising this interpretive regularity as wrong. This in turn seems to evacuate one of the most crucial features of normative demands – their possible divergence from actual practice – leading to the conclusion that such a ‘regularist’ position does not give an account of genuine normativity at all.

Here is Brandom explaining the dilemma and the outlines of his solution in the Conclusion to Making It Explicit (I haven’t yet discussed a number of the terms Brandom is using here, but the general thrust of his remarks should be clear) [I’ve also thrown in a few square-bracketted comments to explain what I take Brandom to be doing in several early remarks, but it’s the conclusion of the quoted passage that’s relevant to the argument of this post]:

Norms (in the sense of normative statuses) are not objects in the causal order. Natural science, eschewing categories of social practice…

[notice that Brandom is characterising the ‘naturalistic’ position he opposes as one that restricts its descriptive categories to those made available by the disciplinary space of the natural sciences – Brandom’s own naturalism is distinguished from this position by also including (naturalistically understood) categories of social practice – DL]

…will never run across commitments in its cataloging of the furniture of the world; they are not by themselves causally efficacious – any more than strikes or outs are in baseball. Nonetheless, according to the account presented here, there are norms, and their existence is neither supernatural nor mysterious. Normative statuses are domesticated by being understood in terms of normative attitudes…

[ – a naturalistically analysable category of social analysis – DL]

…which are in the causal order. What is causally efficacious is our practically taking or treating ourselves and each other as having commitments (acknowledging and attributing commitments) – just as what is causally efficacious is umpires and players dealing with each other in a way that can be described as taking the score to include so many strikes and outs.

It must then be asked how such an apparently reductive story about norms as instituted by social practices can be understood to be compatible with an insistence on the irreducibly normative character of the metalanguage in which norm-instituting social practices are specified.

[again, this is Brandom fighting his fight within the analytic discursive space over whether a non-normative meta-language is in principle possible (it isn’t). When Brandom says that our language of analysis is intrinsically normative, this should not be misunderstood as a claim that normativity itself is unaccountable for in naturalistic terms – except in a highly specific sense related to the possibility of such am non-normative meta-language.]

Here is the short answer: The work done by talk of deontic statuses cannot be done by talk of deontic attitudes actually adopted or relinquished, nor of regularities exhibited by such adopting and relinquishing, nor of dispositions to adopt and relinquish such attitudes. Talk of deontic statuses can in general be traded only for talk of proprieties governing the adoption and alteration of deontic attitudes – proprieties implicit in social scorekeeping practices. (MIE p. 626)

Rephrasing and simplifying this to put it in the explanatory order I’ve adopted in these posts, Brandom here first says that he aims to explain norms in terms of social practices. He then confronts the problem (which I characterised above as the ethical/political problem of regularism) that doing so seems to evacuate normativity in the very move with which we ‘explain’ it. [Actually of course this isn’t what Brandom says in the passage above – he talks about the possibility of a non-normative meta-language; but the regularism issue is a large part of what’s motivating that discussion. I’ll try to address this in a later post.] Brandom then resolves this problem by, apparently paradoxically, saying that what counts in explaining norms is not what social actions (“deontic attitudes”) we actually take, but rather what social actions we ought to take. As Brandom puts it on the next page, his account

incorporates a phenomenalist approach to norms, but it is a normative phenomenalism, explaining having a certain normative status in effect as being properly taken to have it. (627)

In other words, Brandom re-introduces normativity in the very theoretical move that is apparently intended to explain it in naturalistic terms: his solution to the regularism problem is to say that we should use proprieties of practice, rather than anything even close to regularities of practice, as our basic explanatory building block.

This is an extremely confusing (/frustrating) move – and Brandom is aware of the problem:

At this point it can easily look as though the account of normative statuses as instituted by social practices is marching around in an unproductive circle (at best, unilluminating; at worst viciously circular and incoherent). For clearly the prior question arises once more: What is the relation between normative specifications of practices and nonnormative specifications of behaviour? (627)

It is Brandom’s answer to this question that’s so hideously complicated. In a word, Brandom’s answer is the entirety of Making It Explicit‘s explanatory apparatus. Brandom begins his phenomenalist account with normative attitudes, meaning not just attitudes that generate norms, but also attitudes that we ourselves (normatively) take as conforming to proprieties of practice. Then by the time Brandom has laid out his entire explanatory apparatus, he has given us an account of how this attribution of proprieties of practice to normative attitudes (i.e. the social process by which those proprieties of practice are in the first place generated) should be understood.

Described this way, there’s no reason for us not to think that Brandom’s account is viciously circular. My claim (and Brandom’s) is that the circular nature of this account is in fact ‘virtuous’ rather than vicious. Crucial to the explanatory achievement of Brandom’s project is the claim that his apparatus ‘precipitates out’ an analysis of how objective (and, as a subset of objective, naturalistic) description can be generated by our ‘original’ normative practices. This description is then capable of naturalistically accounting for the practices that themselves generate our capacity for objective description: as Brandom says, the practices that in fact institute norms are still part of the (natural) causal order. However, our account cannot begin with non-normatively understood practices, if it is to generate our capacity to explain normative statuses naturalistically. And this is connected to the fact that any explanation must be accomplished by social creatures (in this case members of the species homo sapiens sapiens) who are already engaged in normative practices, to be capable of attempting such explanation at all.

I realise that the above few paragraphs are pretty much just words at the moment – I need to do a huge amount of unpacking if I’m to explain what I take Brandom’s account to actually achieve. However, I think it’s important to give at least a sketch of the general ballpark of the explanation I take Brandom to be attempting – since it at least gives a sense of the kind of account we’re aiming for here.

One reason for doing this can perhaps be illustrated by very quickly referring to the discussion of Making It Explicit laid out in Joesph Heath’s paper Brandom on the Sources of Normativity [pdf]. Heath complains that

One of the most unsatisfactory sections of Robert Brandom’s very complex and difficult book, Making It Explicit, is, unfortunately, the very first chapter…. one expects Brandom to show that the concept of ‘social norm’ that he rests his analysis of language on can, in turn, be cashed out in terms of some simpler set of action-theoretic or behavioural concepts. And since preference for a pragmatic order of explanation is what motivates Brandom’s whole project, it would not be unreasonable to expect an analysis of the action-theoretic primitives to appear front and centre at the beginning of the book. Furthermore, Brandom starts out in the first chapter sounding as if he is going to supply just such an analysis. Thus the absence of any conclusive argument or analysis on this score comes as something of a surprise. Many readers of Making it Explicit finish the first chapter not quite knowing whether Brandom chose to omit the argument, whether he made an argument, but a very weak one, or whether he chose to defer the burden of proof until chapter eight.

My suggestion is the one that Heath later refers to as Brandom’s ‘Hegelian’

tendency to think that philosophical claims cannot be evaluated punctually, so to speak, but must be accepted and rejected at the level of whole “systems”.

I believe that Brandom can plausibly claim to require the entire theoretical apparatus of his system (subject to the important proviso I introduced in my post on Brandom’s Linguistic Exceptionalism) in order to give an adequate account of the nature of this basic theoretical building block.

The task of my next series of posts (which as I say, will probably be a while in coming) will be to explain in more detail why that is so, and what Brandom’s account actually consists in.

Here is the Wittgenstein passage that I’ve been circling around in the last set of posts.

217. “How am I able to obey a rule?” – if this is not a question about causes, then it is about the justification for my following the rule in the way I do.

If I have exhausted the justifications I have reached bedrock, and my spade is turned. Then I am inclined to say: “This is simply what I do.”

In my last post, I suggested that the ‘Epistemological’ regularist problem could be resolved within a broadly regularist framework by accounting for the interpretive acts of ‘seeing-as’ that are required in order to pick out a given regularity, and then to decide what counts as conformity with that regularity, in regularist terms. (I.e. our own act of interpretation is itself a regularity of practice.) I said that this move has the problem that it reproduces, within our solution to the regularist epistemological problem, a moment of the ‘paradoxical’ dynamic that led us to reject regulism. That is, it apparently gives the interpreting agent complete sovereign power over what counts as a binding norm – and this appears to evacuate an important aspect of the meaning of normativity.

However, I have already criticised (in my post from the other day, Two Wittgensteinian Arguments) Wittgenstein’s own application of this ‘sceptical paradox’ in his private language argument: just as Wittgenstein’s private language argument doesn’t in itself justify Wittgenstein’s conclusion, so the problem of sovereign interpretive power doesn’t in itself evacuate normativity from our account. I have argued that Wittgenstein’s private language paradox could in principle be responded to not with a “here we cannot talk about following a rule”, but rather with a “this is simply what I do”. And the same is true of the sovereign power of the interpreting agent in our solution to the epistemological regularist problem. Why can we not say “this is simply what I do” and grant that, although one could in principle redetermine binding norms moment by moment through a sovereign interpretive act, in fact one doesn’t? Genuinely normative interpretation, on this account, would be “simply what I do”.

There is in fact, I believe, nothing wrong with making this move. However, this is where the difference between Wittgenstein’s therapeutic quietism and Brandom’s emphasis on explanation rears its head. For the quietist Wittgensteinian, saying “this is simply what I do” puts an end to the discussion. For the Brandomian, however, saying “this is simply what I do” may put an end to the discussion – for it may draw attention to a material inference that is, so to speak, ‘axiomatic’ within our given set of social interactions – but it may also provide the occasion for an examination of just what it is I do, and why. The process of explicitation means that no “this is simply what I do” is intrinsically the end of the conversation. And because “what I do” is for this reason never tacitly transcendent to theoretical consciousness (as Wittgenstein’s own normative generative principles of interpretive acts in fact are), we may always potentially be faced with the need to explain exactly what it is that we do, and why. It is in this scenario that pointing to a simple (non-transcendent to theoretical consciousness) regularity of practice seems like insufficient justification for an interpretive act – and this is where the ethical/political problem of regularism (which insists that any given regularity of practice cannot automatically be taken as generative of normative principles, on pain of both category error and ethical monstrosity) seems to bite.

A lot of these issues cannot be fully addressed until I have unfolded additional aspects of Brandom’s system – specifically, his discussions of deontic scorekeeping and (especially) of the social practice of asking for and giving reasons. These analytic tools are necessary before we can give an account of how real practices can become part of (or be excluded from) inferential chains of justification, in a way that can make a practice itself a justification, without seeing that justification as intrinsic to the practice. And this discussion needs also to give an account of the circumstances in which, for Brandom, saying “this is simply what I do” can itself function as a reason – I will try to address this perhaps slightly fraught issue under the heading ‘the reason of no reason’ at a later date.

All this needs to wait, however. The main thing I want to do in this post is just draw attention to the way in which Brandom’s commitment to explicitation and explanatory argument prevents him from making the Wittgensteinian “this is simply what I do” move in response to the problem of individual sovereignty over the act of interpretive seeing-as that serves as our (preliminary) solution to the epistemological problem of regularism. Brandom needs a more robust response to this problem than “this is simply what I do”. The following series of posts aims to begin to explain what that response is.

Regularism Revisited

January 21, 2011

I’ve now unpacked a fair few of the Wittgensteinian issues that are important to the early stages of Making It Explicit‘s argument – so I want to return to the issue of regularism, to expand on the problems that Brandom’s practice-theoretic account of the origins of normativity needs to resolve.

In the earlier post Regulism and Regularism I distinguished two problems that Brandom identifies with regularism as a philosophical position – the Epistemological and the Ethical/Political. The Epistemological problem is also a ‘seeing-as’ problem: granted (for the sake of argument) that normative standards can simply be identified with a given regularity of practice, such that conformity to the regularity is obeying the norm, and deviation from the regularity is going against the norm, it is possible to find multiple regularities in the same set of practices, such that those practices cannot themselves dictate which regularity we take as identical with a normative standard. This is related to, but not quite the same as, the Wittgensteinian rule-following problem of being able to interpret any regularity (given a chosen regularity) as compatible with any future practice, provided our interpretation of the generative principles underlying the empirical regularity is rococo enough. Really, then, there are two epistemological problems for regularism: the problem of selecting which regularity counts as the norm-determining regularity; and the problem of deciding which future practices would count as compatible with that regularity, once selected. Each of these questions involve an act of interpretation, or at least of taking-as (taking-as-the-important-regularity; taking-as-compatible-with-that-regularity) – and the practices under analysis cannot themselves determine what interpretive act, or act of taking-as, we use as our prism for selecting a regularity and determining what counts as conformity to it.

This set of issues is central to what Brandom means when he says that his position is not a naturalistic one. I already discussed this point briefly in my post on Sanctions, but I think we are now in a position to cash out in more detail the sense in which Brandom’s philosophy is ‘non-naturalistic’. When Brandom criticises ‘naturalism’, he generally has in mind a position that believes that simple description of a set of practices is capable of communicating the same conceptual content as the articulation of a normative demand. Such ‘simple description’ (on the ‘naturalistic’ account that Brandom opposes) would be non-normative – it would not be influenced by the describing subject’s own social location or commitments, but rather would be part of a (probably physicalist) metalanguage that can a-normatively record the objective states of affairs obtaining in the material world.

Brandom opposes such a picture on two counts. In the first place (for reasons connected to but not fully covered by the issue under discussion here) Brandom does not believe that there could ever be such a thing as a non-normatively descriptive language (short reason why: truth is also a norm). I’ll expand on this point in more detail in future posts. In the second place, for the reasons discussed above, Brandom does not believe that any descriptive language (even if such a perfect a-normative descriptive language could in principle exist) would be able to determine which regularities out of those observable would be determining of normative standards; nor would it be able to determine what future compatibility or incompatibility with such standards would involve. For this reason, the idea of a ‘naturalistic’ account of what norms are established by regularities of practice is, Brandom believes, in principle impossible.

It’s important to keep distinct these two (related) critiques of the possibility of a ‘naturalistic’ account of norms – I’ll try to separate them out further, and give a fuller account of each, in future posts. For now I just want to emphasise that Brandom’s target, in his critique of ‘naturalistic’ accounts of normativity, is narrower than I think it is often taken to be. Brandom’s critique is essentially directed at a quite specific set of positions – one that involves a fantasy about the possibility of a non-normative metalanguage. I am going to argue that Brandom’s own position is in fact a naturalistic one as that term is commonly used, and that one of the things Making It Explicit does is demonstrate how a naturalistic account of the origins of normativity is possible that does not fall into the theoretical traps Brandom here identifies. Brandom’s argument will aim to demonstrate that a set of interpretive and intrinsically normative acts of ‘seeing-as’ are required in order to establish reference to any given set of objective regularities in the first place; but that these interpretive, normative acts of ‘seeing-as’ can themselves then be understood in the naturalistic terms made available by such objective reference. The argument’s more complicated than that, but I want to at least gesture at the general ball-park of the position being defended.

Backing up a little – we’ve covered the first set of problems with the ‘regularist’ position. An act of interpretation is required in order to choose which regularity counts as a normative guideline. Which act of interpretation – and thus which regularity – we select is not itself determined by the regularity under examination – therefore the regularity itself cannot be determining of normative standards. This argument, in a way, parallels the ‘regulist’ one, in that it hinges on our ability to concoct multiple rules compatible with the same empirical phenomenon.

There is a solution to this problem within a broadly regularist framework. We can say that the act of interpretation is itself guided by a regularity of practice (which, indeed, it will be) – we can then acknowledge the ‘seeing-as’ difficulty while incorporating the normative act of seeing-as within an account that understands normativity in terms of regularity of practice. Of course, we then are faced with the parallel difficulty of how we choose which regular practice of interpretation counts as the one that ‘correctly’ picks out the regularity of practice determining of normative standards. And if we pick a second set of regular interpretive acts that make that decision for us, the problem again presents itself one stage of interpretation further down. Nevertheless, we can imagine (at least for the sake of argument) that these difficulties are resolvable – for instance, by fiating an regularity of interpretive practice that also, in its interpretive practice, validates that regularity as the correct one to pick out. Would this be an acceptible conclusion from a Brandomian point of view?

I need to take some care here, because the position just sketched is in fact very close to that which Brandom ultimately adopts – I am going to end up arguing that with sufficiently complex amendments, this position can indeed be seen as a Brandomian one. However, it is important to understand why such amendments are necessary, and why many versions of this position are unacceptable. This issue hinges on the Ethical/Political objection to regularism.

Let’s say, as an example, that we decide that the regularity of practice generative of binding normative standards is the most common set of sanctions at work in a society. Or, let us say (if we wanted to complicate the picture marginally) that we decide it is the most respected set of sanctions, where degree of respect can itself be analysed by looking at regularities of practice associated with approval and disapproval within the social space under examination. Any number of different versions of this basic approach can be imagined, but they all have in common the fact that (putting the epistemological objection to one side) we are identifying binding or legitimate norms with some regularity of practice within the social sphere.

The problem here is quite a simple one: It feels like an absolutely essential component of normative or ethical demands that they be capable of differentiation from actual practice. If we identify binding ethical demands with some regularity of practice, it seems as if we are identifying, in however complex a way, what ought to be with what is. We are not just transgressing the Humean injunction that one can’t get ought from is. More importantly, we seem to be opening ourselves to the apparent ethical monstrosity that what ought to be can be determined by the manipulation of actual practice. This move seems to open whatever ethical theory or ethical attitudes we assemble from such a starting point to an intense vulnerability to power. If, for example, we identify normative or ethical attitudes with the most common (or, for example, the most esteemed) regularity of practice in a given society (even global society), we are permitting that society to determine not just what people do, but what they ought to do. The limit case here is often taken to be a (rather heavily fantastised) idea of a totalitarian state, in which an entire society adopts a monstrous ethical-political set of beliefs and practices. In such a scenario, we want to believe, it does not matter that truly ethical beliefs and practice are opposed by the entire society in question – what’s truly right is still truly right, and what’s truly wrong is still truly wrong. (Orwell’s 1984 is often the go-to text for the articulation of this theoretical dilemma, with the final submission of the protagonist’s resistance even in his innermost subjectivity to the injunctions of the totalitarian state taken as exemplary of the ethical-political reason for requiring a stronger foundation for our norms than social-perspectival practice. A reading of 1984 is used to challenge a pragmatist understanding of the origins of normativity (in this case Richard Rorty’s) by James Conant in the Brandom-edited volume Rorty and his Critics. Rorty gives, to my mind, a rather good response; and I see Brandom’s work as, among other things, concerned with the greater elaboration of the position Rorty articulates there.)

It seems of great importance, then, not to understand normative demands in terms of regularities of practice. We can locate two extremes of this difficulty: on the one hand the identification (just discussed) of binding norms with the regularities of practice of society at large (however understood). On the other hand, we could choose to grant the interpretive act of seeing-as (that must form a component of any solution to the epistemological problem of regularism) the ability to determine which regularities of practice – however unusual or atypical – we take as generative of binding norms. The difficulty then becomes that we have apparently granted the agent who makes this interpretive action of ‘seeing-as’ the ability to choose whichever possible regularity of practice they like as normatively binding – and this ability seems to be identical with the ability to choose whatever norms they wish as binding upon them. We then find ourselves in a situation that precisely parallels the ‘sovereign self-determination’ moment of the rule-following paradox associated with regulism. That is to say, we have made the move from explicit rules to rules implicit in practice, but only by reinstating an act of interpretation at the moment of identification of those rules. Even if we then make this act of interpretation an implicit practice, by saying “this is simply what we do” in response to the regulist dilemma associated with interpretive acts, we have still granted the (perhaps ‘blindly’) interpreting agent absolute sovereignty over which regularities count as generative of binding norms. (We have, as it were, pushed the ‘sovereign power’ moment of the regulist paradox back into implicit practice.) Which is to say, we have apparently granted this agent absolute sovereignty over the normative demands that impinge upon them. We could call this latter scenario the ‘Nietzschean’ possibility – the supposed ability of social actors to create their own normative frameworks through sheer act of will in their interpretation of society and history. This contrasts with the ‘totalitarian’ possibility in which those with power over a society physically and psychologically enforce a set of practices that then automatically determine what is (supposedly) ethically admirable. On the one hand these two possibilities seem at opposite poles – but they are, in fact, different manifestations of the same ‘regularist’ dilemma. (This conceptual closeness may be symptomatic of a real social closeness between the political-ethical ideal of the self-willed norm-generating individual, and fascist/totalitarian politics – though obviously I’m not aiming to discuss that now.)

We are faced, then, with a dilemma. How to articulate an account of the emergence of binding norms from social practice that does not resolve itself into one of the ethically unpalatable (not to say epistemologically implausible) positions associated with ‘regularism’? How to avoid collapsing what should be into what is, while still giving a naturalistic account of normativity? I will begin to try to elaborate Brandom’s answer to this question in my next set of posts.

Wittgenstein Bugbears

January 18, 2011

Okay, a few more remarks on Wittgenstein before I start to try to get to the point of this digression.

A common way of interpreting Wittgenstein runs more or less as follows: W articulates his ‘rule-following paradox’. This paradox shows that there are multiple ways in which any individual can interpret any given rule. Yet we interpret rules correctly – how is this possible? Well, obviously we are members of “forms of life” – social, collective modes of behaviour which impart to us the correct rules. In this interpretation, “forms of life” (or a similar phrase) provide the basic explanatory ground in Wittgenstein’s account – the place where rules have, so to speak, their ontological footing. Wittgenstein is thus taken to open the Cartesian preoccupations of much philosophy onto the social: social forms of life, rather than (say) the synthesising activity of the transcendental subject, are taken to provide the source of our normative frameworks. And this is, indeed, more or less right, as far as it goes.

If this is how we interpret Wittgenstein, it is extremely tempting to attempt to extend Wittgenstein’s (“sadly underdeveloped”) discussions of ‘forms of life’ into a more robust social theory. David Bloor’s Wittgenstein: A Social Theory of Knowledge can be seen as engaged in this endeavour; as, arguably, can Pierre Bourdieu (I will hopefully get around to contrasting Bourdieu with Brandom later in this series of posts). And, indeed, this is ultimately what I aim to be doing here myself. However, a couple of provisos need to be borne in mind.

First, as I think Henry Staten has brilliantly demonstrated in his comparative study of Wittgenstein and Derrida, it is a mistake to see Wittgenstein as solely preoccupied with the problem of how individuals conform to a rule. It is true that Wittgenstein, in the Investigations and related works, returns obsessively to the moments when an individual deviates from another’s or a community’s normative practice, or when the ‘correct’ normative practice is unclear. However, as Staten argues, this is because Wittgenstein is preoccupied with those liminal moments in which it is unclear what it would mean to follow a specific rule – and Wittgenstein does not just resolve these moments by providing a philosophical account of how a rule can, in fact, be followed after all: Wittgenstein is just as interested in the improvisation of practice and unpredictability of creative behaviour that is capable of constituting new, alternative rules, or (perhaps still more importantly) of not adhering to any communal rule – even a newly created one – at all. The conformity of an individual’s practice to an already-collectively-accepted rule is one direction in which these liminal moments can go; but it is not privileged, in Wittgenstein’s text, over the alternative directions just enumerated. For Wittgenstein, the liminal moments in which no guide to behaviour can be found are not problems that necessarily have to be resolved in favour of conformity to a communal “form of life”. In fact Wittgenstein’s own philosophical practice is centrally concerned with rediscovering such liminal moments, breaking up the frozen sea of habitual practice in order to immerse ourselves once again in the waters of doubt. Wittgenstein is at least as concerned with breaking the grip of habitual social practice, as he is with discovering such practice as the legitimate ground for normative guidelines. Those who understand Wittgenstein as principally a theorist of conformity to communally-established rules (which most Wittgensteinians, in one way or another, do) miss absolutely central parts of his philosophical and ethical endeavour.

That’s one issue. The second issue is Wittgenstein’s theoretical quietism, which is central to the Investigations, and which has been adopted in the strong form Wittgenstein advocates (which would rule out most kinds of scientific practice, for example) by more or less no one – no one, at least, who has continued to participate in the discourse of academic philosophy. I would argue that there are two broad categories of responses to Wittgenstein’s quietism in the Wittgenstein-influenced philosophical community: on the one hand, people who largely discard it altogether, except for some general remarks about theoretical modesty and the critique of metaphysics (and I guess this is the camp I’m going to fall in); on the other hand, people who apply it selectively, disregarding Wittgenstein’s most strenuous claims about the inadmissability of explanation or theory, but wheeling on the quietism to close down specific kinds of theoretical discussion, with greater or lesser degrees of justification. I would place John McDowell squarely in this latter camp, for example: Mind and World is, despite McDowell’s avowed quietism, an inescapable explanatory project. Wittgensteinian quietism – and a therapeutic ‘dissolving’ of philosophical perplexities – is, however, McDowell’s excuse for not providing various forms of philosophical explanation that, in my view, the project simply demands if it is to be carried through on its own terms. I agree with Brandom that Making It Explicit provides, not an alternative to McDowell’s category of ‘second nature’, but an explanation of what that category could possibly refer to, if it is to do the philosophical heavy lifting McDowell demands of it. But this is all digression.

Theoretical quietism is central to Wittgenstein’s work in part for ethical/spiritual reasons: there is a Tolstoyan ethic of renunciation of intellectual pride that runs through, and motivates, much of Wittgenstein’s corpus. But there are also some theoretical reasons why quietism is an appealing option for those who gesture – even if only at moments – to such categories as ‘forms of life’ as part of their account of normative standards. In Wittgenstein, as we have seen, the critique of what Brandom calls ‘regulism’ results in the acceptance that “this is simply what we do”. This move, however, places the actual normative guideline that we follow beyond the reach of conscious or theoretical explanation, except as something that presents its demands to us, intuitively. Despite its social-theoretic inclinations, Wittgenstein’s philosophy places norms in a transcendent space – not transcendent of all society, necessarily, but transcendent of our selves in a way that – thanks to Wittgenstein’s quietism – makes them functionally inaccessible and unexaminable. This move is complicated by Wittgenstein’s parallel examination of moments in which we establish alternative normative behaviours – and I will argue (if I remember) in later posts that there is, implicit in these moments of the Investigations, an account of the formation of norms that is strikingly close to Brandom’s anti-quietist analysis. But the main point, for now, is that Wittgenstein’s quietism makes the origin of our norms in principle inaccessible to theoretical consciousness.

Why is this important? It is important because it prevents Wittgenstein himself – though not those Wittgensteinians who wish to take his philosophy in a more explanatory direction – from being vulnerable to the series of theoretical difficulties that Brandom analyses under the heading ‘regularism’. Because Wittgenstein refuses to explain what social behaviours might in fact be constructing the norms that we experience ourselves (as inhabitants of ‘forms of life’) as subject to, Wittgenstein is not vulnerable to the theoretical difficulties that are typically attendant on any such explanation. One might say that Wittgenstein has simply evaded these difficulties – and, of course, he has, albeit in a Tostoyan-renunciation-of-intellectual-hubris kind of way – but that doesn’t make the rejection of quietism any less theoretically problematic. If we aim, unlike Wittgenstein, to actually give a more or less social-theoretic account of how a “form of life” or a communal normative framework functions, we are immediately in the thick of the explanatory problems that Brandom draws attention to in his discussion of regularism. As soon as we reject quietism, we need to figure out our response to these difficulties. If Brandom is right, then that response has to be extremely complicated – much more complicated than most Wittgensteinians seem to anticipate.

The assertion that this is simply what we do, and that at a certain (very early) point, our spade turns, and no more explanation can be sought, is therefore doing double-service in Wittgenstein’s work. On the one hand this move short-circuits the regress of interpretations, rejects ‘regulism’, and establishes Wittgenstein’s pragmatism. On the other hand this same move is used by Wittgenstein to introduce his theoretical quietism, which enables Wittgenstein not to have to confront the difficulties of ‘regularism’. It therefore may seem, in Wittgenstein’s work, that the problems of regulism and regularism are dealt with using the same set of arguments.

However, the two functions that the ‘this is simply what we do’ move serves are logically dissociable – it is idiosyncratic to Wittgenstein’s philosophy that the short-circuit of regulism is also taken to prohibit explanatory intellectual endeavour. Brandom’s concept of ‘material inference’, as we will see, serves to short-circuit the problems of regulism, while still leaving open the possibility of further explanatory tasks – including the explanation of the nature and legitimacy of any given material inference. This process of explicitation – the opposite of Tolstoyan quietism – will lead Brandom on a very long explanatory road. I’m now going to conclude this digression, and attempt to rejoin that road. In the next few posts I’ll try to explore in more depth the problems associated with regularism.

Material Inference

January 18, 2011

All right. I realise I’m going to have to back up and address, at least briefly, the issue of material inference in Brandom if I’m going to be able to make clear what I’m after by contrasting Brandomian explanation with Wittgensteinian quietism.

Brandom derives his category of ‘material inference’ from Wilfrid Sellars. I think Brandom is probably doing some violence to the category as Sellars intended it, but it’s an age since I read Sellars so I wouldn’t want to go to the wall on that. At some point I need to reread Sellars, and also read Brandom’s commentary on Empiricism and the Philosophy of Mind, but this will all have to wait for another day.

Here’s how Brandom introduces the category:

The kind of inference whose correctnesses essentially involve the conceptual contents of its premises and conclusions may be called, following Sellars, “material inference.” (MIE p. 97)

Material inference is to be contrasted with formal inference, wherein inferential structures (e.g. the form of a syllogism) are taken to be valid or invalid independent of the actual substantive content of the variables. The concept of material inference therefore suggests that the differentiation of form and content in, for example, the study of formal logic is, in some deep sense, theoretically unjustifiable – even if it might be pragmatically valuable (which, of course, for Brandom the pragmatist, is more than enough to legitimate a discursive practice). This suggestion occupies a similar theoretical space to Quine’s claim that synthetic and analytic truths are distinguishable only contextually, as those propositions that we are, respectively, willing and unwilling to make vulnerable to the impact of experience in our web of belief. For Brandom, all inference ultimately needs to be explained in terms of material inference – material inference is one of his most fundamental categories.

I’m going to skimp on the discussion of the category here, in the hope of returning to it in more depth at a later date. For now basically all I want to say is that “material inference” serves the function of the Wittgensteinian “this is simply what we do” – it short-circuits a formalist regress. If I make a good material inference then I don’t need an infinite regress of rules for applying rules for applying rules, etc. My inference is good in itself, content and all – indeed the inference cannot be distinguished from the conceptual content it deals with – material inference ‘grounds’ my theoretical chain of explanation.

We now come to the point at which Brandom departs from Wittgenstein. For Wittgenstein, having reached the “this is simply what we do”, we can go no further – our spade is turned. For Brandom, although material inference grounds the activity of inference, preventing the Wittgensteinian regress of interpretations, we can then go on to thematise and question the content of a material inference. Indeed we can, if we wish, begin to apply further elements of the potential ‘infinite regress’ to the material inference, rendering it non-material. This is the process that Brandom calls explicitation. It is a dramatic departure from Wittgensteinian quietism, and it will have large repercussions for Brandom’s system, requiring a substantial theoretical apparatus to deal with the problems it opens up. I will try to begin to look at why in my next post.

[As I’ve already said, I’ll need to return to all this and deal with it much more carefully at a later date – apologies for haste here.]

[This post is the first part of a longer sub-argument, and is not really satisfactory to me – I really need to return to the Investigations and do a proper commentary on the relevant sections that also engages with the debates in the secondary literature. But I don’t want to do that right now, so this post stands as a place-marker for a much more careful possible future analysis.]

I concluded my last post by drawing attention to two different problems Brandom identifies with ‘regularist’ attempts to explain the basis of normativity in naturalistic terms – epistemological and ethical/political. These two problems are brought together – or the extent to which they are aspects of the same theoretical problem is demonstrated – in the set of Wittgensteinian ‘paradoxes’ around rule-following; which in turn connect Brandom’s critique of ‘regularism’ to his critique of ‘regulism’.

I want to devote this group of posts to a brief aside, to address the issue of philosophical and social-theoretic explanation (to which Brandom is committed) as against the Wittgensteinian emphasis on theoretical quietism. As I think we will see, it is this difference in philosophical orientation that breaks apart the critique of regulism from that of regularism – in Wittgenstein’s own texts, these two sets of issues are more tightly bound together than they are in Brandom’s. I want to look briefly at why I think this is so, and what it says about the particular explanatory demands placed on Brandom’s theoretical apparatus.

As we have seen, Brandom’s critique of regulism opens the door to his philosophical pragmatism. The Wittgensteinian rule-following paradoxes demonstrate to Brandom’s satistifaction that there is (in Wittgenstein’s words) “a way of following a rule that is not an interpretation” but that is rather about “following a rule or going against it in actual cases”. This leads Brandom to the conclusion that any philosophical adequate account of normativity needs to be, in part, an account of what it means for rules to be implicit in practice. (Brandom is also committed to developing an account of what it means to make implicit rules explicit in linguistic communication; and then, further, an account of what it means to be able to make explicit in linguistic communication this theoretical account of explicitation itself – i.e. the conditions of possibility of Brandom’s own philosophy. But we are still focussing on the first stage of this long and complicated argument.)

I want to draw attention to two different aspects of Wittgenstein’s rule-following paradoxes, which are sometimes I think taken to be different ways of articulating the same underlying issue, but which are in fact I think rather different in implication (and roughly correspond to Brandom’s distinction between critiques of regulism and critiques of regularism). We could call these, crudely, the move to pragmatism and the move to sociality.


We’ve already looked at the critique of regulism that leads Brandom to a pragmatist position: if we understand norms as explicit rules, then a rule is required in order to make explicit how to follow a rule, and an additional rule is required in order to make explicit how to follow that rule, and so on ad infinitum. The Wittgensteinian argument basically says wtf?, and points out that at some point we simply follow a damn rule, without further thought: in Wittgenstein’s words, we follow a rule ‘blindly’: at some point we reach a rule-following-behaviour at which we simply say ‘this is what we do’.

This argument leads us to rules implicit in practice. But it says nothing about the kinds of practices that are implicitly normative in this way. And, further, it says nothing about how those practices are implicitly normative. These are each additional, and quite different, questions.

Note in particular that nothing in this argument prevents us from concluding that the implicitly normative practices we’re dealing with here are a kind of mental or ideal magic lodged in the depths of the human soul – there is nothing intrinsically naturalistic or social-theoretic about the kind of pragmatism this argument seems to lead us to. [This is partly what justifies Brandom’s close assimilation of the Kantian emphasis on judgement and the Wittgensteinian emphasis on practice, in the first chapter of Making It Explicit: although in MIE Brandom uses Kant as his ‘regulist’ bogeyman, Kant is also a pragmatist of sorts in this precise way – he simply fiats that the synthetic activity of the transcendental subject at some point ‘follows a rule blindly’, and thereby short-circuits the ‘Wittgensteinian’ critique of regulism. We may feel that Kant’s account of the nature of these implicitly normative practices is unenlightening – or indeed mystificatory – but that is a separate issue from the plain move to pragmatism, understood in this very broad way (the practice in question is the activity of the synthesising subject) – a move of which Kant in fact partakes.]

Now – when Wittgenstein makes his critique of regulism, he famously does so by means of a kind of reductio ad absurdum: Wittgenstein points out that we can make any practice conform with a rule, provided we interpret that rule in a creative enough way. This famous argument I think plays two very different roles in Wittgenstein’s work, and it’s worth trying to draw these roles apart. On the one hand, this absurdity is taken to set up the response, on the part of the teacher and the reader, that ‘no, that’s not how you interpret that rule, you’re interpreting the rule for interpreting the rule wrong, here is how you interpret that rule’. This response, which Wittgenstein’s text generates, is part of a longer reductio ad absurdum, whereby this chain of error and response is taken to be capable of continuing indefinitely, through infinite interpretations of interpretations of interpretations, and this apparently vicious regress is then taken as damning evidence of the failure of regulism as a theoretical project. That’s one thing this argument is doing. The discursive lever in this argument is our outrage at the fact that the right rule is not in fact being followed – the argument wouldn’t work if we, the reader, did not share the instructor’s sense that the pupil is doing the wrong thing, and failing to understand the real rule being communicated.

On the other hand, a superficially identical argument elsewhere in Wittgenstein’s text is doing rather different work. Here – as in, for example, the famous ‘private language’ argument – the fact that any number of rules can be made out to conform with a given practice, and any given practice can be taken to conform to a rule (provided we interpret the rule creatively enough) leads to the conclusion not that “this is simply what we do”, but rather “here we cannot talk about following a rule.” For example, when Wittgenstein looks at the possibility of obeying the rules of an entirely private language, he concludes that the subject would be sovereign over the meaning of each token in the language at each moment, and that therefore the stability of meaning – the lack of moment-by-moment absolute control over the content of meaning-units – that is a prerequisite of linguistic communication is absent in the case of a ‘private’ language, therefore it makes no sense to speak of a private language. The difficulty for Wittgenstein is that it’s unclear why the move to “following a rule blindly”, which he makes when dealing with social interaction, cannot also be deployed in the case of private self-interaction. It is possible for the isolated private subject to interpret each symbol whatever way she wants, just as it is possible for the student to do so; but at some point we have to say “this is what we do” – we follow the rule. Even though it seems within our power to re-interpret the rule afresh at every moment, we in fact don’t.

This is what Robert J. Fogelin (in his very useful and insightful book Wittgenstein) calls Wittgenstein’s selective application of a sceptical paradox. The same argumentative set-ups are at different moments in Wittgenstein’s text used to support very different conclusions: at times, the apparent possibility of sovereign control over interpretations, and the subsequent possibility of infinite interpretation of interpretation, etc., is taken to lead us to rules implicit in practice; at other times the apparent possibility of sovereign control over meaning-units is taken as a reductio ad absurdum of the idea that a rule can be followed in this context at all. These are very different arguments and conclusions, and Wittgenstein does not – to my mind at least – fully justify his differential application of each.

That said, it’s clear what’s driving Wittgenstein’s differential applications of his sceptical critiques: he has a particular set of impulses as to how linguistic practice in fact functions, and for many – myself included – those impulses seem plausible. Furthermore, Wittgenstein can, I think, be taken to have justified the move to sociality in the expanded sense of that term that I’ve argued Brandom ultimately intends – whereby the multiple subject-positions in potential communication with one another can be different aspects or moments of the same empirical self, rather than discrete organisms. Thus while I agree with Fogelin that Wittgenstein’s private language argument doesn’t seem to justify the use he makes of it, it does, I think, justify the claim that multiple different accessible subject-positions, with potentially different assessments of the conceptual or normative commitments associated with a symbolic action, are required for the meaningfulness of a symbolic action to any one subject-position – in this sense of ‘public’, language must indeed be public. I think.

Nevertheless, there’s a reason why these sets of issues are more closely bound together or less pristinely distinguished in Wittgenstein’s work than in Brandom’s, and this has to do with different fundamental orientations to philosophy as a task of thought. For Brandom, philosophy is largely an explanatory endeavour – an exploratory one also, to be sure, which leaves room for multiple different approaches to the same basic subject-matter, in the conviction that access to divergent intellectual frameworks is as essential a part of knowledge as it is of wisdom, but nevertheless an endeavour oriented to explaining the origins and functioning of the phenomena under examination. For Wittgenstein, by contrast, and famously, philosophy has nothing to do with explanation. Here is the Investigations‘ paragraph 109:

It was true to say that our considerations could not be scientific ones. It was not of any possible interest to us to find out empirically ‘that, contrary to our preconceived ideas, it is possible to thing such-and-such’ – whatever that may mean. (The conception of thought as a gaseous medium.) And we may not advance any kind of theory. There must not be anything hypothetical in our considerations. We must do away with all explanation, and description alone must take its place. And this description gets its light, that is to say its purpose, from the philosophical problems. These are, of course, not empirical problems; they are solved, rather, by looking into the workings of our language, and that in such a way as to make us recognise those workings: in despite of an urge to misunderstand them. The problems are solved, not by giving new information, but by arranging what we have always known. Philosophy is a battle against the bewitchment of our intelligence by means of language.

I can’t begin to discuss the complexity of commitments involved in this passage in any kind of adequate way here. I want, however, to make some remarks more directly relevant to the broader argument about Brandom of which this post is a part…