Wittgenstein and Brandom: Affinities and

Divergences

SIMON BL ACKBURN

A B S T R AC T WO RK T YPE

It is not difficult to find both affinities and divergences in the work of Article
Wittgenstein and Brandom but this particular text explores several key issues
beyond first impressions and reveals hidden divergences in supposed A R T I C L E H I S T O RY

similarities and occasionally less profound dissimilarities where their Received:

philosophies seem to differ radically. Both Wittgenstein and Brandom (as well 3–January–2018
as Dewey), while agreeing that representations cannot be taken to be primitive, Accepted:
would not approve of Rorty’s drive to jettison the very idea of representation 30–March–2018
along with that of truth. Wittgenstein, on the other hand, “is averse to any
A R T I C L E L A N G U AG E
attempt at general, explanatory, theories of how language works” while
“Brandom is much more positive about the possibility of general, systematic English
and explanatory theories of meaning.” The divergence about the builders of
K E Y WO R D S
Philosophical Investigations § 2 and their language use is traced back to
Wittgenstein’s Builders
Wittgenstein’s being essentially a worldly pragmatist, while Brandom's practice
Language Games
is essentially intralinguistic. In the second part, the text takes issue with
Practice
Brandom's theory of making explicit, and in the final sections, the paper
Expressivism
questions the supposed contrast between expressivism and the ‘motley of
Realism
language' and takes a look at Brandom's strive for a normative realism that
distinguishes him from other expressivists.
© Studia Humanitatis – Universidad de Salamanca 2019

S. Blackburn (✉)
 Disputatio. Philosophical Research Bulletin

University of Cambridge Vol. 8, No. 9, Jun. 2019, pp. 0-00
e-mail: swb24@[Link] ISSN: 2254-0601 | [Link]

© The author(s) 2019. This work, published by Disputatio [[Link]], is an Open Access article distributed under the terms of the Creative Commons
License [BY–NC–ND]. The copy, distribution and public communication of this work will be according to the copyright notice
([Link] For inquiries and permissions, please email: (✉) boletin@[Link].
2 | S I MON B L A C K B UR N

Wittgenstein and Brandom: Affinities and

Divergences

SIMON BLACKBURN

§1. Embodiment, Action, and Causation

T
HERE IS NO DIFFICULTY about seeing broad affinities in the work of these
two remarkable philosophers. Brandom himself pays handsome tribute
to (the later) Wittgenstein as the towering figure behind the
‘conceptual sea change’ of replacing concern with semantics by concern with
pragmatics: the insistence on seeing our linguistic capacities in the context of
human activities and practices (Brandom 2011, p. 160). It is his emphasis on the
use of terms as (generally speaking) a key to their meaning that justifies calling the
later Wittgenstein a pragmatist, and Brandom is of course happy to present
himself as following the tradition as it descended through Sellars, Quine, and
Rorty. Wittgenstein liked Goethe’s saying “Im Anfang war die Tat” (in the
beginning was the deed) and Brandom’s Locke lectures came out under the title
Between Saying and Doing. But although according to some kind of priority to
practice over semantics suggests a particular tradition, it is far from defining one.
Virtually everyone would suppose that if it were not for human activities that put
them to use, words would mean nothing. Any distinctive approach down this road
needs to specify both how our activities are to be described, if they are to give life
to our words, and how the life of our words is to be described, if activities account
for them.
Against this background affinity between Brandom and Wittgenstein it is also
easy to identify at least one divergence between them. Whereas the later
Wittgenstein is averse to any attempt at general, explanatory, theories of how
language works, preferring instead to insist upon the kaleidoscope of different
things we might find ourselves saying in different contexts, Brandom is much
more positive about the possibility of general, systematic and explanatory
theories of meaning, and takes himself to have provided such a theory, or at least
to have paved the way for such a theory to follow. Brandom’s theoretical
ambitions include providing ‘transcendental’ argument that any language must

Disputatio 8, no. 9 (2019): pp. 00-00

 WI T T GE NS T E I N A ND B R A NDOM: A F F I NI T I E S A ND DI V E R GE NC E S | 3

show a certain logical structure; Wittgenstein not only avoids, but counsels against
any such ambition. For Brandom language has an essential core, a “downtown”,
but for Wittgenstein it is like an old city, a ‘maze of little streets and squares’, and
has none (Wittgenstein 1953, § 18).
On this second question a useful landmark might be a number of negative
claims, whose most flamboyant expression comes not in Brandom himself, but in
the hostile semantic eliminativist writings of Richard Rorty. These counsel us to
avoid describing the powers of words by using any typically semantic vocabulary:
the vocabulary of representation, reference, or truth. This ban is not an
immediate consequence of the pragmatist emphasis on practice. After all, our
practices include innumerable activities such as charting the coastline, making a
timetable, drawing up a menu, or erecting signposts, and these charts, timetables,
menus and signposts are quite naturally described as representing, describing, or
signifying other things: the shape of the coastline, the times of the trains, the
lunch offerings, or the way to some destination. And truth is in the offing, since
it is vital to us that these things can be right about their subject matter, or wrong.
It is of course one thing to avoid taking reference, representation, and truth
as primitives, as unmoved movers in our theory of mind and language, but
another thing altogether to refuse to allow them any place at all. So Rorty’s
eliminativism is going to need some special motivation. In many parts of language
there is indeed such a motivation: many authors, including Wittgenstein, have
supposed that in selected areas philosophical understanding requires getting rid
of the idea that every word refers to a correlated thing. But the intended width
or scope of this suspicion is very different in different pragmatists. While for
Rorty, it seems intended to apply everywhere and always, this is not at all the case
in Wittgenstein and (in spite of his dedicating his Locke lectures to Rorty) it is
repudiated in what Brandom presents as his own ‘analytic pragmatism’. For that
is at least mainly the project of rehabilitating reference and representation on a
pragmatist footing.
This allies Brandom with Wittgenstein, for in my view Rorty misunderstood
the main pragmatist attitude to representation. For Wittgenstein, as for Dewey,
and Peirce, it is not that the very ideas of semantic representation and truth that
must be jettisoned. The only claims are that we cannot take representation as
primitive, nor as useful everywhere and without question. Dewey for example
wrote that:
The basic fallacy in representative realism is that while it actually depends
upon the inferential phase of enquiry, it fails to interpret the immediate quality
and the related idea in terms of their functions in inquiry. On the contrary it

Disputatio 8, no. 9 (2019): pp. 00-00

4 | S I MON B L A C K B UR N

views representative power as an inherent property of sensations and ideas as

such, treating them as “representations” in and of themselves. Dualism or
bifurcation of mental and physical existence is a necessary result, presented,
however, not as a result but as a given fact...psychological or mental existences
which are then endowed with the miraculous power of standing for and pointing
to existences of a different order (Dewey 1968, p. 514–515)
The ‘inferential phase of enquiry’ here refers to the activities setting the scene
in which an observation, be it of a thing or of a feature of a situation, becomes
involved in inference, enquiry, and absorption into our view of the world. The
complaint is that by making representative power a miraculous, self–standing
property of ideas or concepts or any ‘thing’ either in the world or the mind, we
simply generate mystery. Then, since physical things like inscriptions and sounds
evidently do not have these magical powers, we add the false idea of a dualism of
mind, where there must reside things such as ideas or concepts that do have such
powers, and the inert world where nothing does. It is the intrinsicness of semantic
properties to some thing or other that is the target, not the propriety of semantic
terminology itself. A similar message is found in Wittgenstein: we should
remember his swift destruction of the idea that we can only obey the order to
bring a red flower by first imagining a red flower, and then using what we imagine
as a pattern to follow (Wittgenstein 1964, p. 3). Here the imagined flower is
supposed to give us the redness of a real flower directly, whereas no word can do
so.
A rejection of semantic ‘intrinsicalism’ has wide implications. It means that
we can no longer unthinkingly suppose that linguistic form is a straightforward
guide to reality, nor even a guide to what we are supposing reality to contain. The
subject matter of thought can no longer simply be assumed to be things
correlated straightforwardly with words, or states of affairs conceived of as
structures of things. Brandom thus applauds three of Rorty’s targets. They are
semantic atomism, which means taking words in abstraction from sentential or
other contexts, the idea of intrinsic semantic powers, and semantic nominalism,
or the modelling of all meaning on naming. But although Brandom does not
himself go the way that Rorty did, wrongly drawing eliminativist consequences
from the repudiation of these errors, he is not always forthright in condemning
the eliminativist turn either (Brandom 2013, p. 95–97).
If it is odd to find the intensely serious Dewey, whose favourite words included
‘inquiry’ and ‘education’ saluted as leading the way to Rorty’s ironic, post–
modern world, it is equally strange to suppose that the Wittgenstein of
“philosophy leaves everything as it is” should be in the business of exorcising any

Disputatio 8, no. 9 (2019): pp. 00-00

 WI T T GE NS T E I N A ND B R A NDOM: A F F I NI T I E S A ND DI V E R GE NC E S | 5

family of terms that have perfectly satisfactory and useful everyday functions, even
if philosophers have made a mess of identifying what those uses are. Perhaps
Wittgenstein could just about get tarred with the eliminativist brush because of
the first thirty or so paragraphs of the Philosophical Investigations, which could
perhaps be read as an attack on the “Augustinian” picture of words representing
things, root and branch. But closer inspection shows that Wittgenstein is
advancing no such embargo. Right at the outset, in paragraphs §2 and §3
Wittgenstein conceded that ‘Augustine, we might say, does describe a system of
communication’ and one that might be appropriate for a ‘narrowly
circumscribed region’ of our own language. The system that conforms to
Augustine’s description is illustrated by the communication between builder A
and assistant B, whereby A calls out ‘block’, ‘slab’ ‘pillar’ and so on, and B brings
such an object to him. This is, Wittgenstein says, a language ‘more primitive than
ours’ but he allows that it is a language nonetheless, and we might be reminded
of Karl von Frisch’s famous discovery of the language or system of signalling of
honeybees.
Brandom refuses to allow that these builders are using even a primitive
‘language game’. He argues that there is a bright line drawn only when we have
practices of inference and assertion, and since these builders make no inferences
and need not be regarded as asserting anything, only as signalling or calling for
things, they are on the wrong side of the line (Brandom 2008, p. 42). Wittgenstein
differs: he explicitly tells us that we are not to be troubled that the primitive
language game consists only of orders (§18). Neither does he deny that we can
use semantic descriptions of their terms: “Of course, one can reduce the
description of the use of the word ‘slab’ to the statement that this word signifies
this object” (§10), but he warns us that such a description is apt to conceal
important differences: In §7 and §8 he imagines augmenting the builders’
repertoire with colour terms, numerals, and indexicals referring to place. His
evident concern is to insist that there are differences of function here. The
discussion culminates in §12 with the well–known comparison between words and
the various handles in a locomotive, so that in §13 he announces that “When we
say: ‘Every word in the language signifies something’ we have so far said nothing
whatever; unless we have explained exactly what distinction we wish to make”. The
concern, in short, is with pluralities of function, not with it being somehow wrong
to say that, for instance the word ‘slab’ in the primitive language, let alone in our
own more elaborate language, signifies (refers to/represents) slabs. After all, if
you “say nothing whatever”, then you do not say anything wrong, either.
We return to the shared concern with pluralities of function later. First,
however, how great is the potential divergence between Brandom and

Disputatio 8, no. 9 (2019): pp. 00-00

6 | S I MON B L A C K B UR N

Wittgenstein here? Elements that Brandom considers crucial to language and

thought are missing from Wittgenstein’s primitive language: there is no mention
of social–deontic norms as a foundation for anything worth calling meaning, and
there is no network of inferential practices to cement the builders’ meanings in
place. Nevertheless, the signalling behaviour of builders A and B, like that of Karl
von Frisch’s bees, is an essential part of a practice and technique —the practice and
technique of building together with slabs, beams and pillars in the one case, and
of having fellow members of a hive coordinating in the activity of gathering
pollen in the other case. Brandom is within his rights to query whether in
primitive and inflexible cases we should talk of communication, and within his
rights to draw a line below which the builders and the bees fall, but an important
point remains. Wittgenstein’s notion of a practice is essentially worldly, identifying
a function of embodied, practical creatures coping with their environment.
Brandom’s by contrast is essentially intralinguistic: it is the practice of users of
language, witnessed in their own inferences and the norms to which they hold
themselves and others. For Wittgenstein a notion of reference, or of a primitive
ancestor of reference, swims into view when we have the builders’ successful
communication as part of their technique of working together with beams, slabs
and the rest. The builders are intelligently engaged with the elements of their trade,
and for Wittgenstein this is critically important. For Brandom, insofar as he
remains an inferentialist, intelligence is exercised intralinguistically, in the
movement from one commitment, expressed by a sentence, to another, rather
than in engagement with things. Reference, for Brandom, either is, or at least is
nearer to being, a syntactic or proof–theoretic notion, visible in the patterns of
inference to which users are committed, than any relationship of a word to a part
of the world.
We might sum up this difference by saying that Wittgenstein was
fundamentally an engineer, not an inferentialist.
The builders are not shown deploying norms and normative language. But
there is, of course, space for characteristic normative behaviour to enter their
primitive world —after getting beams by saying ‘beam’ builder A is likely to be
annoyed if B brings what A regards as the wrong item such as a slab instead of a
beam, just as a teacher introducing numerals may feel like giving up on the learner
who cannot get the sequence right. But this is going to be true whenever someone
disappoints expectations or fails to do his part in the joint exercise of a technique.
Normative reactions bubble up whenever some cooperative behaviour is
expected but not delivered: arriving with a slab when signalled ‘beam’, saying
‘six’ when prompted to continue the number series beyond four, or in general
mislabelling and mistaking things. Indeed, as Gideon Rosen has emphasized,

Disputatio 8, no. 9 (2019): pp. 00-00

 WI T T GE NS T E I N A ND B R A NDOM: A F F I NI T I E S A ND DI V E R GE NC E S | 7

normative reactions can be found not only when we are supposedly cooperating,
but whenever we have actions of any kind. Spitting is a behaviour that gives rise
to normative reactions, but is not itself constituted by an element of normativity
Rosen (2001, p. 622).1
Alongside Wilfrid Sellars, Brandom certainly makes room for entrances into
language, and exits from it. But at least for the most part he seems to conceive of
these in purely causal terms: our engagements with our environment are just
matters of differential causal relations, comparable with the changes in a piece
of iron wrought by damp or magnetism. Intelligence on this view is only exercised
inside language, as we give content to words by the networks of inferences we
make. It is this that motivates Brandom’s wish to build a viable notion of reference
out of materials visible only within the linguistic habits of subjects, such as having
their inferences shaped by relations between pronouns and their antecedents,
giving us a proof–theoretic or syntactic foothold on a notion of co–reference, and
thence, perhaps on reference itself.2 While Brandom also brings to bear our own
commitments as we describe subjects in ‘de re’ terms, it must be that anaphoric
relations are his primary ingredients, since otherwise we simply have it that we,
ourselves able to refer, can interpret others as doing the same. In any pragmatist
programme that disallows a fundamental or primitive place for reference and
representation, our own abilities to refer to things cannot be the fundamental
unmoved mover.3
Brandom is not in the end as pure an inferentialist as I have so far implied.
In Chapter 6 of Between Saying and Doing rather different materials are introduced
and welcomed. Here Brandom addresses the worry that in his approach to
language “all that is in play is words and their use” and he imagines a critic

1
Rosen argues that this throws into doubt the very meaning of the debated question ‘whether meaning is
normative’, and I agree with him.
2
There is an ambiguity in this. Brandom cannot evade results showing that if any set models some
sentences then so does any other set of the same cardinality, implying that reference to one thing rather
than another cannot be conjured from the proof–theoretic structure of any text, however large. All he
can claim is a syntactic fix on whether an expression is grammatically referential. See Arif Ahmed,
“Quine” in C. Misak (ed.), Oxford Handbook of American Philosophy, Oxford: Oxford University Press,
2008, p. 321, n. 18. See also section 2.3.2
3
There is another similar question in the offing here. Brandom seems to give an account of what it is to
have a commitment in terms of what it is to attribute a commitment to another. Although there are
behavioural expressions of this, such as the imposition of punishments when what is interpreted as error
comes about, the threat of a regress still looms. A ‘Martian’ who sees behaviour but not commitments
cannot see disappointed commitments either, even if he can see punishing and sanctioning behaviour.
See Anandi Hattiangadi ‘Making it Implicit: Brandom on Rule Following’ Philosophy and
Phenomenological Research, Vol. 66, No. 2 (Mar., 2003), pp. 419–431.

Disputatio 8, no. 9 (2019): pp. 00-00

8 | S I MON B L A C K B UR N

complaining that “if the world is left out of the story, what justification could there
be for saying that meaning has not been?” (Brandom 2008, p. 177). His answer to
the complaint is that the practices that establish semantic relations between words
and the world are themselves essentially world–involving. It is the “practical
involvement with objects exhibited by a sentient creature dealing skilfully with its
world” that brings about, or even constitutes, the fact that the terms of its
language have semantic relations with the elements of that world. The practical
involvement is further described as “an open–ended sequence of feedback
governed performances” —here one might think, for instance, of the kind of
alert responsiveness to every movement of its intended prey that can be witnessed
in the behaviour of a hunting predator. Brandom also goes on to insist that we
often cannot think of practices without “referring to the actual objects
incorporated in them in different ways” (Brandom 2008, p. 179). This is surely
right, but it raises the question of why the joint practice of the builders should
not similarly be described in terms of asking for, delivering, and in the process
referring to, beams, slabs and bricks, in spite of there being no inferential
relations obtaining between elements of their linguistic repertoire. It also raises
the question of whether it is wise to dissociate the essential nature of linguistic
practice from the essential nature of the object–related practices and activities
that make up the major part of the lives of the users of the language.
I suspect that Brandom thinks that the worldly side, the ‘open–ended
sequence of feedback–governed performances’ does not itself take us outside the
‘space of causes’ (it is an elaboration of what Huw Price calls ‘e–representation’,
or causal sensitivity to the environment). On this view the hunting animal is in
effect no different from a heat–seeking missile; a poor candidate for intelligence
although similarly going through a sequence of feedback governed
performances. But if that is an objection, one needs to ask what the words
‘sentient’ and ‘skillful’ are doing in specifying the intended range of cases. Unlike
the iron passively rusting, the hunting animal and we ourselves are indeed
sentient and skillful, but as a result our transactions with things are not confined
to passive reactions to the causal impacts the environment generates. Our
sentience and our goal–directed natures come into play. On both the input and
output side we can be, amongst other things, educated, quick, acute, industrious,
subtle, discriminating and in many other ways variable for better or worse, and
notably we are in command of what we look for and find. This is where our
embodied skills in dealing with our world are shown. We are not merely
differently responsive to whole causal fields, for our intentionality, our attention
and our actions are selectively directed at particular elements of such fields, and

Disputatio 8, no. 9 (2019): pp. 00-00

 WI T T GE NS T E I N A ND B R A NDOM: A F F I NI T I E S A ND DI V E R GE NC E S | 9

this is where reference gets a foothold.4

An intelligent exercise such as builder B’s search for a missing beam, is a
practice designed to increase the probability of the occurrence of an experience
enabling B to say that it is in some particular place X, and increasing towards
certainty the probability that if she does say that, she is right. It is not merely a
question of extralinguistic causation; it is rather that no intralinguistic, purely
inferential movements of B’s mind would substitute it. Brandom is fond of saying
that ‘experience’ is not one of his words, but I am not sure this is an advantage at
this point. It is the experienced builder who is particularly skillful at looking,
recognizing, selecting, the required tools and objects needed for the task at hand.
5

Wittgenstein has no sympathy at all with a Sellarsian dualism in which the

‘space of causes’ is distinct from the ‘space of reasons’. When in §85 of
Philosophical Investigations he considers the case in which we are directed by a
signpost, he is not concerned to contrast being caused by it to go in a certain way
with being justified in so going. On the contrary he is more concerned to identify
these, as we unthinkingly react in the way we have become used to, which is
typically the right way or the intended way. When a person is guided by a signpost
we can say that the signpost causes his going one way rather than another, but it
is also true that, habituated as he is, it gives him a reason to go one way rather
than another. On seeing it he becomes justified in supposing that his goal lies
this way or that, just as a trained map reader who can interpret the cartographer’s
symbols is both caused by the presence of a symbol for, say, a pond, to expect a
pond in the landscape, and justified in so expecting. 6
As Wittgenstein says, doubts could arise about whether we are interpreting the
signpost rightly, but often enough they do not, and he dismissively adds that
whether they do or not is “no longer a philosophical proposition, but an
empirical one”. In fact, the whole thrust of the “rule–following considerations” is
surely that, although there is no hidden, magical state of the mind, nor a state of
affairs in Frege’s third world or Begriffshimmel that guarantees, in the sense of

4
It is also, tellingly, where the interpretation of non–linguistic creatures as nevertheless believing and
desiring gets a foothold. The case is wonderfully made in Jonathan Bennett, Linguistic Behaviour,
Cambridge: Cambridge University Press, 1976, Ch. 2.
5
The case that Brandom uses too thin a conception of practice was argued in Steven Levine, ‘Brandom's
Pragmatism’ in Transactions of the Charles S. Peirce society, Vol. 48, No. 2 (Spring 2012), pp. 125–140.
6
Thinking that the reason–giving, justificatory power of the process of observation derives only from the
beliefs to which it gives rise, so that its observation is on all fours, epistemically, with any other causal
process such as a knock on the head, leads to the disaster of a coherence theory of truth.

Disputatio 8, no. 9 (2019): pp. 00-00

10 | S I MON B L A C K B UR N

logically forcing, the correct uptake, the common dispositions of mankind

stepping in to close any gap that this might be thought to leave. The thrust could
be described as making the ‘space of causes’ into the space of reasons, just as
Hume supposes that our natural bent interprets a world of regularities as a world
of causal necessities, or for that matter makes a world of natural happenings
describable in moral terms. The vocabulary of causation differs from the
vocabulary of reasons, because the second involves an element of appraisal that
the first lacks, but their reference is the same. The metaphor of disjoint spaces,
with nothing in one that is in the other, is a calamity, and we need instead
something like “the space of intelligent practice” which smoothly amalgamates
doings and reasonings.7

§ 2 Functional Pluralism
As we have seen, Investigations §12 foreshadows the way that throughout his later
work Wittgenstein is intensely concerned to destroy the idea that all words or
sentences function in essentially the same way. In a different paper I have given
a list of the cases in which he puts this idea of functional pluralism to work
(Blackburn 2010). Examples include modal and logical language, ethical and
normative language, mathematical sentences, psychological ascriptions, religious
commitments, attributions of knowledge, and no doubt others. In all these cases
Wittgenstein thinks that we miss the plurality of functions, and hence fall into
philosophical confusion, precisely because the “clothing of our language makes
everything alike” (Wittgenstein 1953, II xi, p. 224). Expressivists about the ethical,
the modal, and other things, have thus saluted Wittgenstein, alongside Ramsey,
Ayer, Ryle and others as both a pioneer and a companion–in–arms.
In his Locke lectures and afterwards, Brandom approaches this by developing
the idea of a pragmatically mediated relation between vocabularies. This comes
about when we first look at what we need to do in order to count as saying what
some vocabulary V lets us overtly express. This gives us what Brandom calls
practice/vocabulary or PV sufficiency. We then switch attention to the ways in
which the practice may be described, enabling us to identify a vocabulary V¢ in

7
The amalgamation suggests the solution to the “rule–following considerations”. The problem they are
supposed to pose is that natural dispositions are not in themselves “normative”, nor infinite in extent, so
that you cannot identify the infinitely extensive, normative rules that are being followed, by means of
them. Seeing that it is we ourselves who are confident of the ways our dispositions extend, and that it is
we ourselves who implement and insist upon norms, enable us to deflect such worries. The Martian
(footnote 9) might be unable to interpret sanctioning behaviour as expressing an attitude to a particular
broken rule, but amongst ourselves we easily and naturally do so.

Disputatio 8, no. 9 (2019): pp. 00-00

 WI T T GE NS T E I N A ND B R A NDOM: A F F I NI T I E S A ND DI V E R GE NC E S | 11

which we can specify the doings that are themselves PV sufficient, and this gives
us an inverse notion of vocabulary–practice sufficiency, or VP sufficiency. The
interesting thing is that V¢ may be different in various ways from V, although there
is this pragmatically mediated relationship between them. Typically the idea is
that V is the richer vocabulary, using notions like conditionals, necessity and
possible worlds, or normative concepts, whereas V¢ does without them, in spite of
being able to specify the practices that are sufficient to equip a speaker for
mastery of V. This abstract description is taken to apply to the shape of
expressivist treatments of, for instance, ethical language. The expressivist
describes the practical doings of people, such as grading, choosing, preferring,
planning, or endorsing, and argues that although these practices can be
described without using normative terms, they equip people for deploying overtly
normative language. Or, they permit us, interpreting them, to do so ourselves
using normative terms: “they think that they ought to prefer B to A” for example,
even if they use no such words. Normative language makes explicit what is
embryonically there before or underneath its arrival.
Brandom gives a number of examples of this relationship and the work it can
do.

(1) Sellars’s approach to our saying that something looks one way or another
provides one example. Brandom endorses Sellars’s view that to
understand such a saying you must be able to say and think that things
are one way or another, and then what is involved in the ‘looks’ locution
is just a hedging or partial withdrawing that ‘evinces the reliable
differential disposition to respond to something by claiming that it is
[green], while withholding the endorsement of that claim’ (Brandom
2015, p. 126). All that is required then to understand ‘looks’ claims (V)
can be described in terms of being able to understand ‘is’ claims, and
nervousness about endorsing them, and these capacities are describable
in the public language of objects, their properties, and our caution about
them, V¢.

(2) In his most famous paper dealing with semantic issues Gilbert Ryle talks
of the commitment of one who is prepared to infer B from A. (Ryle
1950). He talks of the relation between being so disposed, and saying
things like “If A, then B”. A person might have inferential practices, and
say things like “A so B”, without using the conditional form. But when he
says “A so B” we can enquire

Disputatio 8, no. 9 (2019): pp. 00-00

12 | S I MON B L A C K B UR N

whether the conclusion, for example that tomorrow is Tuesday, is legitimately

drawn from the premiss, for instance that today is Monday. And to ask whether
the conclusion is legitimately drawn from the premiss is to raise the question
whether it is true that, if today is Monday, tomorrow is Tuesday...what we have
been taught, if we have been taught it, is in the first instance to argue “p so q” or
else “not–q so not–p”... To accept the conditional is not making a report on any
inference or a comment on any inference. Nor is it recommending, exhorting,
confessing, requesting, or commanding anything (p. 330).

The conditional is a kind of travel warrant (an “inference ticket”):

accepting it is accepting a license to make the inference, and perhaps as
well as a commitment to reject the position of those who refuse to make
it.
Here the conditional form belongs to Brandom’s upper–class vocabulary,
V. Our dispositions to move from A to B and perhaps to grade such
dispositions and even to argue about them make up a practice P. This
practice can be followed and described without actually using any
conditional forms, in a vocabulary V¢. But the practice of arguing from
premiss to conclusion equips anyone for mastery of the V vocabulary: she
needs no more in order to be able to understand the inference ticket,
and to proffer it when need be. Brandom puts this by saying that the
practice described by the lower–class vocabulary can be “algorithmically
elaborated” into a practice of actually using the upper–class, V
vocabulary.

(3) A similar relationship obtains between inferential practices and their

expression in modal and causal vocabularies. Brandom does not
generally acknowledge Hume as an influence, but for Hume exposure to
a world of regular events gives rise to the functional change whereby
observers find themselves following inferential routes (and we might add
endorsing such routes, to the point of being mystified by people who do
not follow them) and this is enough to equip them to deploy the language
of causality with its new implications of necessary connections between
events. The idea can be extended to cover modality in general (here we
might especially compare the Wittgenstein of Zettel §299 where the
inexorability of the rules of multiplication are diagnosed in terms of the
inexorability of our attitude “not merely towards the technique of
calculating, but also towards innumerable related practices”).

Disputatio 8, no. 9 (2019): pp. 00-00

 WI T T GE NS T E I N A ND B R A NDOM: A F F I NI T I E S A ND DI V E R GE NC E S | 13

(4) In the case of indexicality Brandom holds that “in spite of the semantic
irreducibility of indexical to non–indexical vocabulary, it is possible to
say, entirely in non–indexical terms, what one must do in order to be
deploying indexical vocabulary correctly: to be saying essentially and
irreducibly indexical things” (Brandom 2008, p. 25). The rules he offers
are straightforward: if a speaker s at time t and place <x, y, z> wants to say
that P holds of <x, y, z, t, s> it is correct to say “P holds of me, here, now”
and conversely if at the same place and time she asserts “P holds of me,
here, now” she is committed to P holding at <x, y, z, t, s>. The non–
indexical vocabulary of reference to persons, places and times stands as
V¢ to the indexical vocabulary V. 8

(5) About normativity Brandom defends what he calls a Kant–Sellars thesis:

...the claim that in order to apply or deploy ordinary, empirical, descriptive

vocabulary, including observational vocabulary —and hence, in order to deploy
any vocabulary whatsoever— one must be able to do everything needed to
introduce normative vocabulary (Brandom 2008, p. 110).

The idea is that deploying any vocabulary involves making commitments

and the very idea of a commitment implies the possibility of defence,
criticism, withdrawal, and endorsement: the practices that equip one to
use the notion of what you ought to say or must infer or are incorrect to
claim.

In the cases I have talked of we find practice that equips you for moving to the
upper–class vocabulary V, the one which makes explicit the kinds of commitment
that were in fact already embryonically present as we used the apparently more
economical or less demanding vocabulary V¢. This is my own interpretation of
the matter, since I find Brandom’s own way of putting the relationship, in terms
of PV sufficiency, a little obscure. Consider, for instance the case of normativity.
As we have seen, expressivists are certainly going to hold that the practices of
holding attitudes, endorsing choices, insisting on one thing or another are
enough to equip you to use normative language. But they are not wedded to the

8
Brandom’s view about this is contested by John McDowell, ‘Comments on Lecture One’, in Philosophical
Topics, Vol. 36, No. 2, (2008), p. 53


Disputatio 8, no. 9 (2019): pp. 00-00

14 | S I MON B L A C K B UR N

thought that these abilities are by themselves sufficient to enable you to use that
language. Once you have normative language at your disposal you can happily
say such things as “if lying is wrong, then getting your little brother to lie for you
is wrong”, and the notorious Frege–Geach problem is precisely that the ability to
parse such sayings is not given by the simple possession of attitudes and
preferences. Simple possession of practical stances might be evinced, it seems,
simply through practical attitudes, and if these need voicing a simple “Boo–
Hooray” language might perform the task. So even showing that these practices
equip you for understanding the indirect context is not straightforward: it is the
enterprise of crossing what I christened Frege’s abyss, and while I think the
crossing can be made less mysterious than it has seemed to many to be, the
opponents of expressivism continue to deny it. Brandom mentions the Frege–
Geach problem but his discussion, it seems to me, does not demonstrate how he
(or Sellars, with whom he associates himself) actually solve it, or whether they
even think that it is a problem to be solved.9 We should notice as well that the
notion of an algorithm does not in its natural use relate practices, so when
Brandom talks of one practice being algorithmically elaborated into another he
presumably has in mind that any transformation of the underlying vocabulary V¢
into the fully–fledged upper–class vocabulary V with its inclusion of the
inferential practices we get when we put normative statements in indirect
contexts, is in some sense a mechanical or computable transformation. But it is
difficult to see what this could mean if it is not to imply a full classical reduction
of V to V¢. Algorithms require deductive relationships.
Another aspect of this relationship between vocabularies needs noticing.
Suppose that V is the vocabulary of the “manifest image” or ordinary life with its
abundant reference to people, places, landmarks, spatial relations, middle–sized
dry goods, and so on. It is one thing to say, as Sellars and Brandom do, that an
essential ingredient in using that language is the ability to make inferences, to
adopt commitments and to behave normatively in appraising, accepting, or
rejecting such commitments. But it is quite another thing to imagine a weaker
vocabulary, V¢ that is anything like sufficient to describe mastery of V. As
Brandom emphasizes the use of V will typically involve a capacity to make

9
The Frege–Geach problem is not indexed in Between Saying and Doing. In Perspectives on Pragmatism
Brandom directs us to Section IV of chapter 3, but that section does not offer a clear approach to the problem.
However Brandom could certainly follow the kind of approach I have sometimes suggested. If we think of
having an attitude in terms of a commitment to endorse or reject other attitudes, and we have second–order
attitudes, supposing that some such commitments require (or rule out) others, then we have the materials to
hand to solve the Frege–Geach problem. Allan Gibbard’s similar approach goes via the notion of plans for
acceptances and revisions of plans.

Disputatio 8, no. 9 (2019): pp. 00-00

 WI T T GE NS T E I N A ND B R A NDOM: A F F I NI T I E S A ND DI V E R GE NC E S | 15

abundant and automatic material inferences. So consider as an example “if a

house has foundations, it cannot be hoisted onto a truck”. Surely there is no
prospect at all of a weaker vocabulary describing a set of inferential capacities
that is sufficient to give anyone mastery of this. You can only make such an
inference through knowing about houses, foundations, abilities, spatial
displacement, and trucks. We can, I accept, isolate things you must be able to do
in order to achieve this mastery, but that is describing necessary structural skills,
not sufficient practices describable in a weaker vocabulary that eschews reference
to houses and the rest.10 Skills with modality and normativity may well be
describable in a vocabulary free of these specific references, but the full–scale
inferential practices cannot be.

§ 3. Expressivism and the Motley of Language

When we try to compare Brandom’s approach with Wittgenstein’s functional
pluralism, the ground becomes a little slippery. Brandom tells us:

I endorse a sophisticated expressivism with regard to logical, modal, and normative

vocabulary. This is quite a different line of thought from that motivating contemporary
expressivist treatment of normative vocabulary, for instance in Gibbard and Blackburn. I
call my version a ‘sophisticated’ expressivism to mark the fact that the expressive role to be
shared by both classical and modal logical vocabulary is one possible role picked out from
a structured space of possibilities (Brandom 2011, p. 206–207).

It is unclear to me quite what this final sentence is suggesting, since it does not
seem to introduce a contrast to which Gibbard and myself (or Ayer, Stevenson,
or Hare) were blind, as we offered motivations for expressive treatments of
normative vocabulary, such as a desire to protect naturalism, a concern for
understanding the motivational nature of normative commitments, or worries
about supervenience. Instead it turns attention to something Brandom wishes to
say about classical and modal logical vocabulary, but although his expressivism
certainly deserves admiration for its scope, and its detection of logical, modal and
normative claims as universally present in any assertive practice, it remains
obscure whether it is this generality, or some other feature, that increases its
sophistication. 11

10
This point is also made in John Macfarlane, ‘Brandom’s Demarcation of Logic’, Philosophical Topics, Vol 36,
2008, p. 61.
11
One factor that thickens the plot here is that Brandom is (rightly) an unqualified admirer of Huw Price’s
‘subject naturalism’, but Price sees his subject naturalism as very much at one with the expressivist approaches

Disputatio 8, no. 9 (2019): pp. 00-00

16 | S I MON B L A C K B UR N

Light may be cast on this by turning to the adjective ‘analytic’ in what

Brandom describes as his own analytic pragmatism. This marks Brandom’s idea
of rehabilitating semantic vocabulary by seeing it as a philosophical latecomer
that is only to be ushered in via careful attention to inferential practices that serve
to introduce it. So, as we have seen, reference is not conceived as a self–standing
intelligible relation between words and things but as a piece of vocabulary whose
utility derives partly from inferential practices associated with anaphora, and
partly from a description of the social–deontic commitments of people offering
de re descriptions of each other. Similarly truth, for Brandom is best explicated
through the prosentential theory of Dorothy Grover and others. The
prosentential theory, like Horwich’s deflationism, is presented as a descendant
of Ramsey’s redundancy theory of truth, and each of them has the consequence
that truth cannot be a substantive or robust or metaphysically heavyweight
property or relation, bringing with it the kind of suspect “authoritarianism” that
opened it to Rorty’s attack, and that might make it a proper target for
eliminativism. On these deflationist views, truth and reference are on the one
hand too small to bring suspect baggage with them and on the other hand too
catholic for anyone to worry about whether they properly consort with the upper–
class, V vocabularies of conditionals, normative judgments, or modal claims.
Were it not for his desire to distinguish himself from it, I should have said that
this shows Brandom aiming at the goal of quasi–realism, the enterprise of
justifying and explaining the reason why all claims wear the same “everyday
clothing” as empirical, or observational or other paradigm naturalistically
legitimate claims.
If we read Wittgenstein as he sometimes asks that we read him, as a
philosophical quietist, content with the higgledy–piggledy layout of our linguistic
and cognitive city, and resistant not only to classical projects of analysis but to any
idea of privileging one vocabulary over another, then Brandom will not seem like
a true descendant of his.12 But when we look at Wittgenstein’s own practice we
see something rather different. Wittgenstein may have thought that
philosophical enterprises were motivated largely or entirely by mistakes about
language. But he also thought that it was intensely difficult to achieve a
perspicuous view within which puzzles and difficulties would no longer appear.
His functional pluralism was an essential tool in dislodging our mistakes and
approaching the perspicuous view. The ‘motley’ of language is not just the
uninteresting fact that we talk about a lot of different things and say a lot of

of Gibbard and myself. See Perspectives on Pragmatism, p. 191.

12
This is McDowell’s reaction, in the paper cited in footnote 8.

Disputatio 8, no. 9 (2019): pp. 00-00

 WI T T GE NS T E I N A ND B R A NDOM: A F F I NI T I E S A ND DI V E R GE NC E S | 17

different things about them. It is a motley of underlying activities and states, and
different functional roles for the expression of those activities and states. This is
what cements his close affinity with the project that expressivists have developed
in most of the areas we have covered.

§ 4. Realism?
However, one twist in Brandom’s analytic pragmatism sets him on a somewhat
different course. Perhaps surprisingly Brandom is eager to present the upshot of
his pragmatism as modal or normative realism. His excavations show that the
upper–class languages for instance of norms and possibilities are realistic: and this
is something he apparently identifies with the commitment to there being
objective normative and modal facts (Brandom 2015, p. 195). So whereas others,
including myself, have typically used expressivism as an alternative to realism, for
Brandom it is a route to realism. This is something that prompts Huw Price, for
instance, to complain that Brandom stands much further from Rorty or
Wittgenstein than he seems to recognize Price 2011).
The issues here are not straightforward, mainly because the word ‘realism’ is
now almost useless as a label for any identifiable ‘ism’.13 At the very least we will
want to distinguish what I shall call full–blooded realism from anything that gives
us the trappings or husk of realism without the substance. Full–blooded modal
realism is nicely described as follows:

“The world is everything that is the case”, as Wittgenstein famously said. As theoretical
speculators we might hope (as an ideal) to grasp the world by coming to know everything
that is the case. But what role in this enterprise is played by our modal beliefs —our beliefs
expressible by modal constructions? What is the point of having such beliefs and the means
to express them? Central to modal realism is the view that having modal beliefs has exactly
the same kind of point as having non–modal belief about, say, cabbages or kings. Just as
someone lacking beliefs about cabbages or kings would lack beliefs about everything that is
the case, so too would one who lacked modal concepts and beliefs deploying them
(McFetdridge 1990, p. 140).

I do not think that this is a position that Brandom wants to endorse. Changing
from talking about cabbages and kings to talking about possibilities is not simply
sauntering to a different part of our surrounding landscape. Rather his view that

13
It is, after all, over thirty years since Arthur Fine complained that “the realist programme has degenerated by
now to the point where it is quite beyond salvaging”. Arthur Fine “Unnatural Attitudes: Realist and
Instrumentalist Attachments to Science”, Mind, Vol. 95, No. 378 (Apr., 1986), p 149.

Disputatio 8, no. 9 (2019): pp. 00-00

18 | S I MON B L A C K B UR N

there are objective modal facts surely has to be taken in the light of his
deflationism coupled with his own expressivism about the modal (unlike Huw
Price, Brandom does not present himself as an expressivist about cabbages and
kings). It is to be something like an “elaboration” of the fact that we are fully
committed to various inferential practices, including those expressed by
counterfactuals denying that the inferences in question are good because of how
we happen to be (this is how independence and objectivity are parsed). In other
words, just as the figure I christened the quasi–realist gets to say things like “there
are duties, and facts about duties, and these facts are often independent of how
we happen to think”, or “there is a possibility that p but not–q which it would be
wrong to ignore” so Brandom’s realist follows the same course to get to the same
kind of commitment.
Does this deserve calling realism? Does the pragmatist, Ramseyan,
Wittgensteinian, or Sellarsian revolution so quickly eat its children? It is many
years since it was first suggested to me that quasi–realism might better be called
queasy realism.14 But while it can be done, I still find it misleading. Expressivism
does not take us to McFetridge’s realism as described above. It does not open the
door to any quasi–geographical, metaphysical imaginings. It does not open the
door to an “outside–in” epistemology whereby we are somehow sensitive to the
facts about modality. It does not deal in aspects of reality about which it might be
difficult to know anything, or about which our concern would seem entirely
optional and rather puzzling. And it was certainly not at all congenial to “full–
blooded” modal realists such as David Lewis.15 Perhaps it is nearer to the kind of
“non–metaphysical” realism promoted by such writers as Parfit or Dworkin in the
case of ethics, although that hoped to strut the title whilst avoiding any kind of
engagement with metaphysics, epistemology, or semantic theory at all.
Because reference and truth are, for Brandom, deflated, minimalist, or quasi–
syntactic notions, so is the notion of a fact. So if, protected by deflationism, and
having a good story about why we use concepts like those of obligation,
correctness, necessity and possibility, we find ourselves talking happily enough of
truth, reference, description, representation, and facts in these domains, it is
hard not to feel that we have been given something less than we might have
expected, a realism with the core removed, a husk rather than the real thing that
McFetridge describes. We might say of this “realism” what Wittgenstein said about
every word signifying something, that until we know what contrasts and

14
By the late Bob Hargrave.
15
I know this at first hand, having tried unsuccessfully to convince Lewis of the virtues of expressivism in the
nineteen–eighties, both at Oxford and at Princeton.

Disputatio 8, no. 9 (2019): pp. 00-00

 WI T T GE NS T E I N A ND B R A NDOM: A F F I NI T I E S A ND DI V E R GE NC E S | 19

implications are in play, it tells us nothing. It is realism without any “...ism”.

Wittgenstein, I think, would not have been happy to end here. I do not think
he ever showed any inclination to throw out all the contrasts —descriptions versus
rules, statements versus avowals, ethics versus facts— that are so fundamental in
launching his theorizing.16 He does not think that if we look behind the front of
the cab all the handles turn out to do the same kinds of thing. He does not avail
himself of the opportunity that deflationism offers, to throw the very notion of
description into the minimalist pot.17 But, had he done so, then it would still have
been important to see that the real philosophical achievement would lie not in
any realist–sounding things he might end up saying, but in the marvelous journey
that arrived at his saying them. And the same is true of Brandom.
There are many other avenues that could be explored to link and sometimes
contrast the works of Wittgenstein and Brandom. One is that whereas
Wittgenstein is intensely puzzled by things like the inexorability of logic, the
queerness of ethics, or the mysterious nature of self–reference, Brandom is
relatively untroubled, since for him we are always in charge of the ‘social deontic’
norms to which we owe fealty. They are like the rules of chess, which we control.
This is a contrast of philosophical temperament that deserves more exploration
than I can give it here, but one that should interest anyone who is to profit from
these two major figures. 18

16
Although philosophy leaving everything as it is, may be a close cousin of the revolution eating its children.
17
I mooted this strategy for Wittgenstein in “Wittgenstein, Wright, Rorty and Minimalism” Mind, 1998, pp. 167–
8.
18
I owe thanks to Robert Kraut, and especially to Arif Ahmed, for acute and helpful comments on an earlier
draft.

Disputatio 8, no. 9 (2019): pp. 00-00

20 | S I MON B L A C K B UR N

REFERENCES
BLACKBURN, SIMON (2010). “Wittgenstein’s Irrealism”. In Practical Tortoise Raising
and Other Philosophical Essays, edited by Simon Blackburn. Oxford: Oxford
University Press, pp. 200–20. First published in Wittgenstein: Towards a Re–
evaluation, edited by Rudolf Haller & Johannes Brandl. Vienna: Hölder–
Pichler–Tempski, 1990. doi:
[Link]
BRANDOM, ROBERT B. (2008). Between Saying and Doing. Oxford: Oxford University
Press.
BRANDOM, ROBERT B. (2011). Perspectives on Pragmatism. Cambridge, Mass.:
Harvard University Press.
BRANDOM, ROBERT B. (2013). “Global anti–representationalism?”. In Expressivism,
Pragmatism and Representationalism, edited by Huw Price. Cambridge:
Cambridge University Press. doi:
[Link]
BRANDOM, ROBERT B. (2015). From Empiricism to Expressivism. Cambridge, Mass.:
Harvard University Press.
DEWEY, JOHN (1968). “Logic: the Theory of Inquiry”. In Last Works, edited by Jo
Ann Boydston. Carbondale: Southern Illinois University Press.
MCFETRIDGE, IAN (1990), Logical Necessity and Other Essays. London: The
Aristotelian Society.
PRICE, HUW (2011). “One Cheer for Representationalism”. In Naturalism without
Mirrors. Oxford: Oxford University Press, pp. 304–321.
ROSEN, GIDEON (2001). “Brandom on Modality, Normativity and Intentionality”.

Philosophy and Phenomenological Research 63, no. 3: pp. 611–623. doi:
[Link]
RYLE, GILBERT (1950). “’If’, ‘So’, and ‘Because’”. In Philosophical Analysis, edited
by Max Black. Cornell: Cornell University Press. pp. 323–341.
WITTGENSTEIN, LUDWIG (1953). Philosophical Investigations. Oxford: Blackwell.
WITTGENSTEIN, LUDWIG (1964). The Blue and Brown Books. Oxford: Blackwell.

Disputatio 8, no. 9 (2019): pp. 00-00

 WI T T GE NS T E I N A ND B R A NDOM: A F F I NI T I E S A ND DI V E R GE NC E S | 21

N OTES ON CONTRIBUTOR
Simon Blackburn is Distinguished Research Professor in the Department of Philosophy at the University of
North Carolina; Fellow of Trinity College at the University of Cambridge. He began his academic career there
as an undergraduate student in 1962, obtaining a PhD there in 1970. Professor Blackburn is Fellow Emeritus
of Pembroke College Oxford Erskine Fellow at the University of Canterbury, and has been visiting professor
or held a research fellowship at and received honours from more than twenty other universities all over the
world. His main line of interest is questions of ethics, but he has worked successfully in almost any field of
philosophy. He is the author of many books; just to mention the most popular: Think (OUP 1999), Being Good
(OUP 2001) and Truth (Penguin 2005). He published a great wealth of articles, contributions to anthologies
and dictionaries, and proffered many conferences.

P O S TA L A D D R E S S
Faculty of Philosophy, University of Cambridge, Sidgwick Avenue, Cambridge CB3 9DA, UK. e–mail (✉):
swb24@[Link]

HOW TO CITE THIS ARTICLE

Blackburn Simon. (2019). «Wittgenstein and Brandom: Affinities and Divergences». Disputatio.
Philosophical Research Bulletin 8, no. 9: pp. 00–00.

Disputatio 8, no. 9 (2019): pp. 00-00