JOURNAL OF PHILOSOPHICAL RESEARCH

VOLUME 33, 2008

WITTGENSTEIN, CONTEXTUALISM, AND NONSENSE:

A REPLY TO HANS-JOHANN GLOCK

EDMUND DAIN
UNIVERSITY OF CHICAGO

ABSTRACT: What nonsense might be, and what Wittgenstein

thought that nonsense might be, are two of the central questions
in the current debate between those—such as Cora Diamond,
James Conant and Michael Kremer—who favour a “resolute”
approach to Wittgenstein’s work, and those—such as P. M. S.
Hacker and Hans-Johann Glock—who instead favour a more
“traditional” approach. What answer we give to these ques-
tions will determine the nature and force of his criticisms of
traditional philosophy, and so the very shape Wittgenstein’s
work has for us, as well as, to some extent, what the lesson of
the Tractatus might be. My aim in this paper is to provide a
detailed defence of the austere view of nonsense, that lies at
the heart of the resolute approach, against a range of influential
criticisms developed by Hans-Johann Glock and which focus
on Wittgenstein’s contextualism. In so doing, I hope also to
shed some light on the kind of view the austere view is, as
well as how it might relate to certain other crucial aspects of
Wittgenstein’s thought.

I.
C
ontextualism, broadly speaking, is the view that the whole—be it judgement,
thought, proposition, or sentence—has priority in some sense over the individual
parts—be they words or concepts—when it comes to giving an account of meaning,
or of understanding. Some version of that view lies behind Frege’s context-principle,
which is reformulated by Wittgenstein in the Tractatus and quoted approvingly
in his Philosophical Investigations (as well as appearing in various formats and
at several places in between, e.g., Wittgenstein 1993: 54; 1978: 44 and 63; 1975:
102 EDMUND DAIN

58–59; 1969: 5). In this paper, I discuss a number of influential criticisms levelled
at the austere view of nonsense (“austerity”)1 by Hans-Johann Glock, in his paper
“All Kinds of Nonsense” (2004), and which focus in large part on Wittgenstein’s
contextualism. I shall argue (among other things) that Glock fails to present a sound
case against the austere view.
Glock, like P. M. S. Hacker, rejects the austere view of nonsense both substan-
tially, as offering the correct account of nonsense, and exegetically too, as offering
the correct account of Wittgenstein’s view of nonsense, either in the Tractatus or
in the later work.2 In short, for Glock, Wittgenstein’s contextualism, neither in the
Tractatus nor in the Investigations, will serve to justify attributing to Wittgenstein
an austere view of nonsense. Wittgenstein’s later (weak or “non-restrictive”)
contextualism is, Glock suggests, substantially plausible, but “militates” against
the austere view. And Wittgenstein’s earlier (strong, “restrictive”) version, Glock
argues, though it could be used to provide an argument in support of austerity, not
only relies on a notion of meaning unavailable to “resolute”3 readers (i.e., those
very readers who would most want to attribute to Wittgenstein an austere view),
but is also plain wrong. These points are supplemented, too, by a number of others
that Glock makes, and that are aimed at undermining any independent plausibility
that the austere view might be thought to have, exegetically or otherwise. I shall
want to dispute much of what Glock writes, but before turning to his criticisms, I
want first to note one key difference between Glock’s view and a view which is in
many respects very similar to Glock’s, that of P. M. S. Hacker.
Glock identifies the following two strands to the austere view of nonsense:
(1) The privation view: nonsense only ever arises from privation, from our
failure to assign to the words (in that context and to date) a meaning; there
is no such thing as “positive” or “substantial” nonsense, consisting of words
which do in such contexts have meanings.
(2) Nonsense monism: all nonsense is logically equivalent; there is only one
logical kind, or there are no logically different kinds, of nonsense (Glock
2004: 222).
Against austerity, both Glock and Hacker maintain that there can be such a thing as
what Glock terms “combinatorial nonsense”—nonsense, that is, that results from
the logico-syntactically illegitimate combination of otherwise meaningful words
(Glock 2004: 223). Hence, both reject (1), the privation view. Some nonsense
results not from a simple failure to assign meanings to the words, but from using
words in ways that are prohibited by, or that violate, the logico-syntactical rules
for their correct use. Since in a nonsense-sentence, so the story goes, the words
are being misused, they do not there stand for a meaning, but they are meaningful
nonetheless in that they do have a rule-governed use in the language.4
But whereas Hacker nevertheless maintains, along with followers of an austere
view, that nonsense “no more comes in kinds than it comes in degrees” (Hacker
2000: 365), and so affirms (2), Glock, on the contrary (and as the title of his paper
suggests), takes their shared stand on the existence of combinatorial nonsense to
WITTGENSTEIN, CONTEXTUALISM, AND NONSENSE 103

constitute grounds for rejecting (2), or nonsense monism. Thus, Glock writes, “There
are many kinds of nonsense, and one of them results from the illicit combination of
meaningful words.”5 For Glock, combinatorial nonsense does amount to something
logically distinct from mere privation, whereas for Hacker it does not.
The reason for this difference is, I think, quite simple, and it is not to be traced
(say) to Glock’s holding a stronger, more robust (or, some would say, more hon-
est) understanding of what combinatorial nonsense consists in than does Hacker.
Rather, the difference is over what would constitute a logical difference between
nonsense-sentences. For Hacker, such a difference would have to be one in the
end-product, as it were, in the sense of the resultant whole. And for there to be
differences in that sense, one would first have to hold a genuinely substantial view
of nonsense (in which the individual words of a nonsense-sentence do have their
ordinary meanings and together express a logically incoherent thought), which
Hacker, of course, does not. For Glock, on the other hand, a logical difference here
is rather one in the cause—in what makes the string of signs nonsense. Hence, on
that way of counting, strings of signs that are nonsense because we have failed to
give them a meaning are a logically distinct kind of nonsense from strings of signs
that violate logical syntax, say. Thus, for Glock, points (1) and (2) are much more
closely related than they are for Hacker; where for Hacker, one can consistently
maintain (2) whilst rejecting (1), for Glock, rejecting (1) would require one also to
reject (2). The difference, then, between Glock and Hacker here is largely termi-
nological. It is that difference that allows them to say such seemingly contrasting
things about the idea of there being different kinds of nonsense, whilst neverthe-
less saying such similar things on many related substantial points. Although much
of their discussion is aimed at the idea of substantial nonsense (and so logically
distinct kinds of nonsense primarily in Hacker’s sense), I take it that followers of
the austere view would reject the notion that there can be logical differences be-
tween nonsense-sentences in either sense (since giving up on (2), in either Glock
or Hacker’s sense, would require one to give up on (1) also).6

II.
That is all that I want to say about the differences between Glock’s view and
Hacker’s. I want to turn instead now to Glock’s criticisms of austerity, especially
in relation to the Tractatus, beginning with his focus on the appeal, among resolute
readers, to Wittgenstein’s contextualism, to Wittgenstein’s reformulation there of
Frege’s context-principle.
That principle appears, in different guises, four times in Frege’s Foundations of
Arithmetic: in the introduction, as one of three guiding principles—“never to ask
for the meaning of a word in isolation, but only in the context of a proposition”; at
§60—“we ought always to keep before our eyes a complete proposition. Only in
a proposition have the words really a meaning. . . . It is enough if the proposition
taken as a whole has a sense; it is this that confers on its parts also their content”;
at §62—“it is only in the context of a proposition that words have any meaning”;
104 EDMUND DAIN

and in conclusion at §106—“we must never try to define the meaning of a word in
isolation, but only as it is used in the context of a proposition” (Frege 1980: x, §60,
§62, and §106). In the Tractatus, the principle becomes: “Only propositions have
sense; only in the nexus of a proposition does a name have meaning” (TLP 3.3),
and “An expression has meaning only in a proposition” (TLP 3.314).7
Although Wittgenstein himself, as Glock emphasizes, does not actually use the
context-principle to justify an austere view of nonsense, it has nevertheless been
used that way by resolute readers of the Tractatus.8 But there are many different
interpretations of the context-principle, and of what Frege and Wittgenstein might
have meant by it at different points in their writings. Glock distinguishes between
strong, “restrictive” contextualism, and weak, or “non-restrictive” contextualism.
Whatever exactly Glock means by each of these terms, it is clear that, for him,
the former, strong version, requires at least that words only have a meaning when
actually used in a proposition,9 while the latter, weak version, “is compatible with
the idea that individual words can mean something without actually occurring in
a proposition”—rather, they must only be capable of occurring in a proposition;
they must have been given a (rule-governed) use in the language.10 With these two
versions of contextualism come two different notions of what it is for a word to
“occur in the context of a proposition”—a narrower and a broader interpretation
of what that context must be exactly. In the former, strong sense, the context is that
of an actual proposition, of which the word must be part. In the latter, weak sense,
the context is rather that of propositions more generally—a word must only have
a role in them, and need not actually be employed in that role at any one time for
it to be said to have a meaning.
Then Glock’s argument is this: if the contextualism of the Tractatus is to be
taken to offer support for the austere view of nonsense, or evidence that Wittgenstein
held such a view there, then it must be taken in the strong sense—since if it is only
by virtue of a general possibility of occurrence in propositions that a word has a
meaning, as the weak version would have it, then words occurring in a nonsense-
sentence may still have a meaning even though they do not there actually occur
within a genuine proposition. In one sense, the stronger sense, of the phrase, we
would then be able to ask after the meaning of a word taken in isolation, outside
the (immediate) context of a proposition, and so too therefore in the context of a
nonsense-sentence. Taken in the strong sense required by austerity, however, Glock
maintains, contextualism is plain wrong: words can and do have meaning outside
the context of a proposition—for instance, numbers on pages, names used in greet-
ing or as labels on jars, entries in dictionaries, and so on. Thus, if Wittgenstein
did hold this view in the Tractatus, then, Glock thinks, Wittgenstein was simply
mistaken. And while that mistake might, at least in part, be accounted for, such an
account will have to go by way of an appeal to certain technical, picture-theoretic
commitments on Wittgenstein’s behalf, such as his extraordinary notion of “mean-
ing,” and which are simply not open to resolute readers or readings to appeal to.
So while there is some evidence that Wittgenstein did hold strong contextualism
WITTGENSTEIN, CONTEXTUALISM, AND NONSENSE 105

in the Tractatus, that evidence itself, Glock argues, counts against his having held
an austere view of nonsense there.
On the other hand, however, Glock continues, there is good reason perhaps not to
attribute such a (strong) view to Wittgenstein at all, or, at least, there is good reason
to see Wittgenstein as already, in the Tractatus, moving away from that strong view
towards the weak version Glock thinks is to be found in the Investigations, since
strong contextualism is at odds with certain other elements of the Tractatus, namely,
its compositionalism—the idea, expressed for instance at Tractatus 4.024–4.03,11
that the sense of a sentence is in some sense dependent upon, or built up out of, or
arrived at by reflection upon the meanings of its constituent parts, the individual
words (together with the structure of their arrangement).12 Much better perhaps,
then, Glock suggests, to attribute to Wittgenstein in the Tractatus the weak, non-
restrictive form of contextualism,13 thus dissolving any sense of inconsistency but,
with it, too, anything that could constitute a contextual justification for the austere
view of nonsense.
Glock, then, presents those, primarily resolute readers,14 who either favour an
austere view of nonsense substantially or who find such a view in Wittgenstein
(or, worse still, both), with something of a Scylla and Charybdis between which to
navigate. On the one hand, strong contextualism would provide evidence for the
austere view of nonsense, but at the expense of both plausibility and internal (to the
Tractatus) consistency—with the only reasonable explanation of either the latter
incoherence or the former error going by way of notions unavailable to resolute
readings of Wittgenstein. On the other hand, weak contextualism would restore
both plausibility and consistency to the Tractatus view, but at the expense of any
justification for an austere view of nonsense.
So far, the argument, if correct, would entitle Glock to conclude only that (ex-
egetically or otherwise) austerity receives no support from the contextualism of the
Tractatus, not that the austere view is either wrong or not Wittgenstein’s. But Glock
goes further: for Glock, this argument serves also to undermine the preferred reading
of Tractatus 5.473 and 5.4733 as explicitly stating an austere view of nonsense, as
well as diminishing any independent appeal austerity might have had. And, Glock
continues, the case only gets worse for austerity when we turn to Wittgenstein’s
later work, where Wittgenstein’s uncompromisingly non-technical use of the word
“meaning” renders strong contextualism untenable, and whose philosophic proce-
dure anyway assumes the existence of combinatorial nonsense.15

III.
Glock, then, argues for two overarching conclusions: first, the exegetical claim,
that the austere view is not Wittgenstein’s view of nonsense, early or later; and
second, the substantial claim, that the austere view is plain wrong as a view of
nonsense. Glock’s arguments for those conclusions can be broken down into the
following steps. First, Glock’s exegetical case proceeds by way of the following
four steps:
106 EDMUND DAIN

The Exegetical Argument:

(1) The attribution to Wittgenstein (early and later) of an austere view of non-
sense relies on two kinds of evidence:
a. Apparent statements of the view—e.g. TLP 5.473–5.4733 and PI §500;
b. Wittgenstein’s contextualism—e.g. TLP 3.3 and PI §49.
(2) Wittgenstein’s contextualism, (1)b, will not support the attribution to him
of an austere view of nonsense.
(3) The austere view lacks independent (and substantial) plausibility.
(4) Hence, without the support offered by Wittgenstein’s contextualism (1)b, the
evidence of (1)a is more plausibly read as supporting a different conception
of nonsense.
Glock’s claim, then, is that the evidence of (1)b will not justify the attribution to
Wittgenstein of an austere view of nonsense, and that the evidence of (1)a alone
will not then suffice either.
Why won’t Wittgenstein’s contextualism provide an argument for his having
held an austere view? Glock’s argument for (2) consists of the following points:
The Case for (2):
(i) Wittgenstein’s contextualism must be either such that a word can only
be said to have a meaning when it is actually being used in a proposition
(strong contextualism) or such that a word can have a meaning outside
the context of a proposition (weak contextualism).16
(ii) Only strong contextualism can provide an argument for the austere view
(since if a word can have a meaning outside a genuine proposition, as
weak contextualism allows, it can also have a meaning in a nonsense-
sentence).
(iii) Later Wittgenstein’s contextualism is weak contextualism.
(iv) Therefore, later Wittgenstein’s contextualism will not support the attribu-
tion to him of an austere view of nonsense.
(v) If the contextualism of the Tractatus is weak contextualism, then it will
not support the attribution of an austere view of nonsense to Wittgenstein
there.
(vi) If the contextualism of the Tractatus is strong contextualism, then it
requires (for substantial plausibility) a theoretical notion of meaning.
(p→q)
(vii) A theoretical notion of meaning is incompatible with the austere view
of nonsense (since austerely-conceived nonsense “cannot constitute a
theory”) (Glock 2004: 227).
(q→~r)
WITTGENSTEIN, CONTEXTUALISM, AND NONSENSE 107

(viii) Strong contextualism is incompatible with the austere view of non-

sense.
(p→~r)
(ix) If the contextualism of the Tractatus is strong contextualism, then it will
not support the attribution of an austere view of nonsense to Wittgenstein
there. (From (viii).)
(x) Therefore, early Wittgenstein’s contextualism will not support the attribu-
tion to him of an austere view of nonsense. (From (v) and (ix).)
(xi) Neither early nor later Wittgenstein’s contextualism will provide sup-
port for attributing to him an austere view of nonsense. (From (iv) and
(x).)17
Point (3) is then established by the argument that, unless strong contextualism is
correct, the prohibition on words having a meaning in nonsense-sentences will
violate the “privation”18 element of austerity—that is, Glock suggests, if words
can have meaning outside the (immediate) context of a proposition, but not in a
nonsense-sentence, their lack of meaning in the nonsense-sentence will then in part
be a result of their context, and not simply of our failure to give them a meaning.19
And Glock’s case for point (4) then proceeds by noting other passages (and from a
variety of sources) which Glock claims allow for “combinatorial nonsense”—non-
sense, that is, which combines meaningful words20 in illegitimate ways—and by
way of the additional point that the austere view (specifically, its “monism”21) is
incompatible with Wittgenstein’s later philosophical method, and his use there of
reductio ad absurdum arguments.22
Finally, Glock’s exegetical case is supplemented by the following substantial
objections, which largely rehearse features of the former (exegetical) argument:
The Substantial Argument:
(A) The austere view must be either independently plausible or receive support
from (either strong or weak) contextualism.
(B) Weak contextualism will not support the austere view.
(C) Strong contextualism is wrong (since it relies on an incorrect notion of
meaning and since it conflicts with compositionalism23), so will not support
the austere view.
(D) The austere view lacks independent plausibility.
(E) The austere view neither receives support from contextualism nor is inde-
pendently plausible, and therefore is incorrect.24

IV.
In the following pages, I will want to dispute a number of Glock’s claims, focus-
sing especially on Glock’s case for (2), and his discussion of contextualism, which
108 EDMUND DAIN

forms, as it were, the lynchpin of his argument against the austere view. I want
to begin with Glock’s claim, repeated both in his exegetical and in his substantial
arguments against the austere view, that only strong contextualism is strong enough
to secure the case for the austere view of nonsense. That claim is made in step (ii),
as I have numbered them, but before I turn to that step, I want to say something
about Glock’s understanding of the austere view itself, and the assumptions that
might seem to underlie his criticisms of austerity.
The suspicion here is that Glock quite simply gets the austere view wrong, that
he misunderstands the kind of view that it is, instead treating it in a way appropriate
to a very different kind of view. Glock, that is, treats the austere view of nonsense
not as the primarily negative view that it is, a view predominantly of what nonsense
is not, but on the contrary as involving a positive thesis about nonsense, and this
is shown by his conception of the kind of justificatory demand that the austere
view has to meet.25 Thus, that suspicion is borne out, I think, by Glock’s treating
the austere view as consisting of two basic premises or assumptions—as Glock
calls them, the privation view and nonsense-monism—and then treating those two
premises as themselves standing in need of positive philosophical justification and
argument, as themselves part of a general theory of sense or meaning.
That mistake in large part comes from Glock’s ignoring, or not taking sufficiently
seriously, how the austere view arises, namely, out of the rejection of the very idea
of a theory of sense, which might seem to be able to give us a more substantial
conception of what nonsense is over-and-above our ordinary conception, insofar
as there is one, and out of the rejection of the idea that Wittgenstein himself was
genuinely concerned, at any period, to put forward a theory of sense or of mean-
ing.26 By not taking seriously enough the possibility of austerity’s arising out of that
rejection, Glock ends up treating the austere view as itself a positive philosophical
theory about what nonsense is, and not as a simple affirmation of the commonsense
notion of nonsense arising when our words fail to say anything, coupled with the
rejection of the idea that there is any other, more sophisticated and philosophical,
kind of nonsense over and above the kind involved in that ordinary notion.
What I want to suggest here, then, is that at the very beginning of his paper, in
his setting-up of the issues at stake, Glock mispresents the kind of view that the
austere view is. In so doing, he closes off from consideration certain arguments
in favour of the view and lays down a demand for an altogether different kind of
argument, justifying austerity as part of a theory of sense or meaning and not as
standing outside such theories. Glock excludes from the picture the possibility that
the austere view might be the outcome of the kind of rejection that it is in fact an
outcome of, at least in the sense in which Conant and Diamond talk of an austere
view of nonsense; instead Glock thinks that the austere view, its two “premises,”
and the arguments in favour of them, must themselves be part of a theory of sense
or of meaning.
Contrary to Glock, the conception of nonsense at work in the austere view,
and invoked by Wittgenstein in his use of that word, is just our ordinary notion
WITTGENSTEIN, CONTEXTUALISM, AND NONSENSE 109

of having failed to say anything by our words, of not having given to our words
a meaning, in that context and to date.27 That notion brings with it the idea that
we could give those words a meaning—that there is nothing, as it were, internal
or essential to that combination that renders it nonsensical. We could, for any
combination of signs, assign to them a meaning such that the whole would make
sense, and so, ultimately at least, if such a combination does not make sense it
will be because of our failure to make just such an assignment. That idea, Glock
(slightly grudgingly) acknowledges, is at least trivially true,28 but he assumes that
the austere view must be meant in some other, stronger, sense than that trivial one,
and not that it simply denies the existence of any further kind of nonsense over
and above this trivial kind.
Glock’s claim, then, that the austere view requires the support of contextual-
ism, and indeed his whole case against the austere view, seems to rest on a much
more basic error about the kind of view the austere view is and about the kind of
justification it requires. I shall want to argue, however, that even were we to follow
Glock in his characterization of the issues at hand, still his arguments do not present
a sound case against the austere view, exegetically or otherwise.
I want to turn back, then, to the claim of (ii), that only strong, restrictive
contextualism could possibly provide an argument for the austere view. Thus, it
is strong contextualism, Glock writes, that “provides the crucial premise for the
following argument”:
P1 A word (name) has meaning only in the context of a proposition.
P2 A proposition is a sentence with a sense.
C No component of a sequence of signs that lacks a sense can have a
meaning. (Glock 2004: 225)
But whether or not strong contextualism is the version that is required depends, in part
at least, also on what other versions are available, and whether they too are capable of
ruling out enough for the austere view. Hence, Glock’s claim in (ii) in turn depends
on the claim of step (i), that Wittgenstein’s contextualism must either be strong or
else it must be weak, and so must maintain instead that words can have a meaning
outside the immediate context of a proposition. I want to raise a question about the
distinction Glock draws here between strong and weak contextualism, and about his
claim that the former strong version is what is required if Wittgenstein’s contextual-
ism is to provide evidence of his having held an austere view of nonsense, or in order
for a substantial case for the austere view to be built upon contextualism.
Those last claims, about what is required of contextualism (whether exegetically
or substantially) by an austere view of nonsense, depend upon the weak view being
taken in a particular way, i.e., as holding a word has a meaning if it can be used
in a proposition—if, that is, it has a role in propositions generally. That version or
interpretation would leave open the possibility of Glock’s combinatorial nonsense
(resulting from prohibited combinations), and also of substantial nonsense (result-
ing from illegitimate or prohibited combinations, and expressing an incoherent
110 EDMUND DAIN

sense). Hence, if that view is the only alternative to the strong view, then it looks
very much as if the strong view is what is required in order to make the case for
austerity. But the contrast that Glock presents between strong contextualism and
the weak variety is actually this: on the strong view, no word has a meaning except
when it is actually being used in a proposition; on the weak view, words can mean
something “without actually occurring in propositions.” Although Glock clearly
associates the latter position with the view that he attributes to the later Wittgen-
stein, and which Glock himself endorses (i.e., the view that a word must only be
capable of occurring in a proposition—in the sense that it has been given some
rule-governed use in the language—in order for it to have a meaning), accepting
this view is not the only way of denying the strong view.
There are, in effect, two ways of describing the terrain here. Either the weak view
is just the view that words can at least sometimes mean something without actually
occurring at that moment as a component of a proposition; and in that case, there
will be many different ways in which one might hold such a view, many different
views which might all nevertheless count equally as “weak” in that sense. On this
way of counting, for instance, simply acknowledging the existence of exceptions
to strong contextualism (as, e.g., Rupert Read 2000: 77 might be taken to suggest)
would be sufficient for one’s view to count as weak. If we describe the terrain
in this way, strong contextualism is not required to secure the case for austerity
because there may be versions of the weak view, which are not yet as weak as
Glock’s version of that view, but which would still provide support to austerity. Or
we might describe the weak view as Glock’s favoured view, that to have a meaning
a word must only be capable of occurring in a proposition. In that case, however,
the strong and weak versions no longer between them cover all the ground there is
to be had: there will be scope for a variety of positions in between those two, and
neither strong nor weak. Hence, again, the strong view would not—not clearly at
any rate—be required to make the case for austerity.
What Glock wants is something from each of these descriptions: from the first,
he wants the idea that the two positions between them take up all the territory avail-
able, so that if one’s form of contextualism is not one then it must be the other;
and he wants from the second the association of the weak view with the version
of contextualism that Glock favours, which effectively closes off that position to
the austere view and its followers. Needless to say, he cannot actually have it both
ways. The point here, then, is just this: that if rejecting the strong view does not
automatically lead to one’s adopting Glock’s own view, then it is not at all clear
that the strong view is, as Glock says, what is required in order for contextualism
to provide support for austerity. Crucially, it would not then be clear that Glock’s
objections to the strong view are so much as relevant to the exegetical or substantial
plausibility of the austere view of nonsense.29
Clearly, the strong view would support austerity (even if, as Glock notes,
Wittgenstein himself does not actually use it explicitly to provide an argument for
austerity); and it is a view that, for instance, Diamond ascribes to the Frege of The
WITTGENSTEIN, CONTEXTUALISM, AND NONSENSE 111

Foundations of Arithmetic (Diamond 1995: 73–93). Frege’s view there, Diamond

notes, does not allow for the meaningful use of, e.g., proper names in isolation,
as in greetings (Diamond 1995: 80). But Diamond also adapts that contextualism
to Frege’s later treatment of sentences as complete names, and that view would,
Diamond says, allow for the meaningful use of proper names in such cases.30 So
Diamond clearly does not take herself, or Frege, to whom she also ascribes an
austere view of nonsense, to be committed to the strong view. Rather, what contex-
tualism must rule out, if it is to be taken as providing support for an austere view
of nonsense, is simply this much: as Diamond puts it, “senseless whole and parts
with content.”31 Thus, Diamond writes of Frege:
[H]e does not merely mean that a word has meaning if it contributes to
the sense of any sentence in which it occurs, in accordance with general
rules; that is, he is not saying that it is the general possibility a word has
of contributing to sense that confers meaning on it. That would allow for
the possibility of a senseless sentence composed of words which had had
content conferred on them by general rules. But what he actually says . . .
is that it is through the sense of the whole that the parts get their content,
and if this means anything at all, it must rule out the combination: senseless
whole and parts with content.32
That combination is the bare minimum that contextualism must exclude if it is
to provide a case for austerity. And ruling out that much requires, according to
Diamond, ruling out one way of taking Frege’s principle—the way favoured by
Glock. But that need not (though it might actually) result in one’s taking the strong
view as Glock describes it.
There is, then, a question-mark over whether the contrast Glock presents between
the strong and weak views is quite as straightforward as he seems to suggest, and,
given that, there is then a further question-mark over whether what would be required
by austerity is that strong view at all. Although strong, restrictive contextualism might
be used to provide an argument for austerity, there is a deal of scope for less restric-
tive versions (such as that involved in Diamond’s discussion above), and which are
not yet as weak as Glock’s own favoured version, but which would still be capable
of excluding the possibility of, e.g., Glock’s combinatorial nonsense. That scope for
different varieties of contextualism, which would nevertheless still be potent enough
to provide an argument for the austere view of nonsense, may also serve to undermine
the import of Glock’s objections to strong contextualism, if those objections hinge
on features of that view absent from those less restrictive versions.
My claim, then, is that Glock misrepresents the contrast between strong and
weak contextualism, and that in doing so, he exaggerates how much must be
excluded by contextualism for it to be compatible with, or provide a case for, the
austere view of nonsense. Despite that, however, the contextualism to be found in
the Tractatus does indeed look very much like strong contextualism, that words
only have meaning when actually used in a proposition, in much the same way
that its ontological counterpart in the Tractatus seems to rule out the idea of a
112 EDMUND DAIN

simple object (“thing”) occurring not in some state of affairs. Wittgenstein asserts
the parallel as follows:
2.0122 Things are independent in so far as they can occur in all possible
situations, but this form of independence is a form of connexion with
states of affairs, a form of dependence. (It is impossible for words to
appear in two different roles: by themselves, and in propositions.)
If that were the case, and the version of contextualism at work in the Tractatus is
indeed the strong version as this remark suggests, then notwithstanding the objec-
tions to Glock already posted, his criticisms of the austere view, at least in relation
to the Tractatus, might seem to hold good (by combining steps (vi) to (ix) in the
above argument with the claim that Wittgenstein’s Tractarian contextualism is
indeed strong). I shall want to argue now that even if the version of contextualism
to be found in the Tractatus is, as it seems to be, strong or restrictive, still Glock’s
arguments do not amount to a sound objection to attributing the austere view of
nonsense to Wittgenstein there.

V.
Glock’s argument here, then, begins with the claim that strong contextualism,
at least in relation to our ordinary use of the word “meaning,” is plain wrong.
Hence, in order to have any substantial plausibility at all as a view, this form of
contextualism requires a theoretical notion of meaning. And that, Glock claims, is
at odds with austerity and the resolute programme.
Glock’s argument here goes by way of the simple fact that, in ordinary thought
and talk, words can and do very often have a meaning outside the context (in the
strict sense at least) of a proposition.33 So, for instance, Glock gives two examples.
First, the following list of words:
to be to abide
to have to arise
to do to awake
Second, a dictionary entry:
nonsense n 1 a: words or language having no meaning or conveying no intel-
ligible ideas b(1): language, conduct, or an idea that is absurd or contrary to
good sense (2): an instance or absurd action 2 a: things of no importance or
value: trifles b: affected or impudent conduct. (Glock 2004: 226)
In the first example, Glock notes, the words are not part of a proposition, but nor
are they simply meaningless: rather, the left-hand column lists the auxiliary verbs,
and the right-hand column the first of the irregular verbs of the English language.
Of the second case, Glock writes: “It would be absurd to maintain that the words
printed in bold at the beginning of dictionary entries are meaningless, all the more
so since the text that follows specifies what they mean.”34
WITTGENSTEIN, CONTEXTUALISM, AND NONSENSE 113

Glock may well be right that, in one sense of the word “meaning,” that claim
would be absurd, i.e., obviously false. But that point, one might reply, fails to engage
with the strong context-principle (as invoked by Frege and the early Wittgenstein),
precisely because the sense of “meaning” appealed to, or assumed there, is a quite
different one. So, as Diamond writes:
You may use the word “meaning” in any way you like, but nothing that
logically can be a characteristic of a word in isolation can help to explain its
meaning in the sense of “meaning” in which what a sentence says depends
on the meanings of its working parts.35
In that sense of the word “meaning”—described broadly enough to encapsulate
Frege’s different conceptions as well as that of the Tractatus—it clearly would not
be absurd to say of “nonsense,” as it appears in bold in the dictionary entry cited
above, that it is meaningless; though the text of the definition does indeed specify
the various roles that word can play as a working part of a proposition, it does not
there actually fulfil any one of those roles. And similar things might be said of
Glock’s first example, too. For instance, some of these “verbs” can have very dif-
ferent propositional roles (we talk of a “to do” list, for example, or exclaim “What
a to do!”) and they each could be given others, but they do not play any such role
at all in Glock’s list.
That response, however, plays nicely into Glock’s hands. For Glock argues
that such an objection is not open to resolute readers of the Tractatus—i.e., those
who would most want to find an austere view of nonsense there—since it relies
on adopting a theoretical notion of meaning as against Glock’s ordinary use of
the term (step (vi)), and since such a notion is itself incompatible with the austere
view of nonsense (step (vii)).
Glock’s claim here is just this: if one thinks of the Tractatus as consisting—in
whole or in large part—of “plain” (i.e., austerely conceived) nonsense, then one can-
not also claim to find at work there a theoretical notion of meaning since, as Glock
writes, “such nonsense cannot constitute a theory” (Glock 2004: 227). That much,
at least, is surely true: nonsense, however conceived, cannot constitute a theory,
but it is less clear why it should follow from this that resolute readers cannot find
in the Tractatus a technical notion of meaning.36 That, on the contrary, seems to be
part of the backdrop against which any reading of the Tractatus must situate itself.
What is clear is that if resolute readers wish to discard the Tractarian statement of
the context-principle as elucidatory nonsense at the climax of the book (and not
all will wish to do so), then they will not also be able to rely on it as substantial
evidence for the austere view.37 If they do that, they might still want to rely on it as
exegetical evidence, as forming part of the Tractarian ladder which one climbs up
but must kick away afterwards, and if so, some story will be needed of how that is
so much as possible. Still, such a story may not be as hard to find as might at first be
thought, since the austere view does allow for all kinds of other differences between
nonsense-sentences—differences not logical, but, say, psychological or aesthetic,
for instance—and which may suffice to provide such a story. It would, however, be
114 EDMUND DAIN

fair to say that the burden here would lie with those resolute readers who followed
this route (even if alternative—standard—readings are likely to themselves require
a parallel story of their own). Nevertheless, it simply is not clear that no such story
is possible, and Glock provides no reason for thinking it to be. Hence, even if we
grant that strong contextualism is required, and if we grant too that that relies on a
technical notion of meaning, that provides no clear-cut case against attributing to
Wittgenstein in the Tractatus an austere view of nonsense.

VI.
Glock’s exegetical case against the austere view in relation to the Tractatus is
supplemented by the point that strong contextualism is at odds with, or conflicts
with, another view there: namely, the Tractatus’s compositionalism. Glock’s sug-
gestion here is that that conflict undermines the attribution to Wittgenstein in the
Tractatus of strong contextualism, and that it suggests that a substantial case for
the austere view could not be built upon strong contextualism, since, for Glock, the
compositional view is clearly correct. Thus, Glock concludes that the conflict sug-
gests that “the early Wittgenstein did not take TLP 3.3 [the Tractarian reformulation
of Frege’s context-principle] as literally as proponents of the austere view suppose,”
and that strong contextualism must be wrong (Glock 2004: 228 and 226).
Roughly, compositionalism is the view that the sense of a sentence is, in some
sense, determined by the meanings of its constituent parts and the way that those
parts are put together. In the Tractatus, that view gets expressed in the following
remarks:
4.024 To understand a proposition means to know what is the case if it is
true.(One can understand it, therefore, without knowing whether it is
true.)It is understood by anyone who understands its constituents.
4.025 When translating one language into another, we do not proceed by
translating each proposition of the one into a proposition of the other,
but merely by translating the constituents of propositions.
(And the dictionary translates not only substantives, but also verbs,
adjectives, and conjunctions, etc.; and it treats them all in the same
way.)
4.026 The meanings of simple signs (words) must be explained to us if we
are to understand them.
With propositions, however, we make ourselves understood.
4.027 It belongs to the essence of a proposition that it should be able to
communicate a new sense to us.
4.03 A proposition must use old expressions to communicate a new
sense.
The merit of that view, as Wittgenstein’s remarks suggest, is that it seems to
be the only way of explaining our ability to understand new sentences: we do,
that is, understand sentences we have not previously encountered, and the most
WITTGENSTEIN, CONTEXTUALISM, AND NONSENSE 115

plausible—perhaps the only plausible—way of explaining that fact seems to

be that we are familiar with the constituent words and with certain patterns of
combining them.
Though Glock does not expand further on the sense of contradiction between
that view and strong contextualism, reasons for thinking them in conflict are not
hard to find. For where contextualism (of any stripe) asserts the “primacy of the
proposition,” as it were, over its constituent parts, compositionalism on the con-
trary stresses (or seems to stress) the primacy of the individual words over the
proposition.
One way of seeing how those two views might not conflict after all is to ask
what follows from compositionalism; that is, compositionalism itself might be
held to be trivially true, but what is a matter of contention is what that view then
entails. On Glock’s interpretation of it, compositionalism has the consequence that
a word has a meaning independently of any sentence in which it occurs. On another
reading, however, it might be taken to entail instead only that when a component
of one sentence occurs again as a component of another sentence it must have
the same meaning in both occurrences. That second reading would then clearly
be compatible with strong contextualism, but it might seem to leave mysterious
the very feature of natural languages that compositionalism intuitively seems to
be required in order to explain, namely, the fact that we can understand sentences
which we have not previously encountered.38 Does strong contextualism—and the
latter view of the consequences of compositionalism—then leave that fact a matter
of mystery? One explanation of why it does not is given (again) by Diamond in
her discussion of Frege.
Diamond reads Frege as maintaining not only a strong form of contextualism,
but also the compositional view that “we understand a sentence only because we
know the language—know, that is, the general rules fixing the content of expres-
sions in the language.” Thus, Diamond writes:
We need to see how Frege can do both: can mean what he says about the parts
getting their content through the sentence’s having sense, and can recognise
that we grasp what a sentence says via our grasp of general rules determining
the meaning of expressions in the language. (Diamond 1995: 109)
Diamond’s answer, in short, is that we do arrive at the sense of a sentence by means
of attributing content to the parts, but that we proceed to an understanding of the
sense of a sentence by attributing that content only provisionally, conditionally
upon the whole sentence expressing a thought of such-and-such a form. Thus, only
if the sentence as a whole expresses a thought of such-and-such a form will the
parts have the content provisionally assigned to them.
So, for instance, Diamond takes as an example the sentence “Venus is more
massive than Mercury,” and she begins by assuming that there are two kinds of
general linguistic rules. The first enables us “to break down whole sentences into
elements with a syntactic characterization”; the second fixes “the meanings of
proper names, concept expressions and relational expressions of various sorts”
116 EDMUND DAIN

(Diamond 1995: 109). And both kinds of rule apply only conditionally. Now, faced
with an utterance that we have not previously come across, we can apply each kind
of rule in turn. Supposing that “Venus is more massive than Mercury” is such an
example, we might apply the first kind of rule in order to give us a characterization
of what the syntactic structure of the sentence might be—what combination of what
kinds of expressions. So we might take certain pointers—the presence of capital
letters for instance—to signal that what we have here is a proper name, followed
by a relational term, followed by another proper name. But, crucially, we apply
these rules only conditionally—we are, as it were, offering a prognosis, and not a
diagnosis. Diamond writes:
[T]he sentence may be taken to be a two-term relational expression com-
pleted by the proper name “Venus” in the left-hand place and the proper
name “Mercury” in the right-hand place, but only if the thought expressed
by the whole sentence is that the object “Venus” stands for, whatever that
is, has whatever relation it is the relational expression stands for to whatever
object it is “Mercury” stands for. (Diamond 1995: 110)
The sentence will have such a syntactic structure only if the thought it expresses
does actually have a form of this kind. And the same is true of the second kind
of rule: we might know, for instance, that “Venus” is sometimes used as a proper
name to stand for the particular object Venus; but again, that will be borne out only
if the thought expressed by the sentence as a whole is a thought asserting of Venus
whatever the rest of the sentence says.
On Diamond’s account, then, we do arrive at the meaning of a sentence com-
positionally, but crucially also conditionally, and because our hypotheses as to
what the parts of the sentence mean are conditional on what the overall thought
expressed by the sentence actually is, that process is perfectly compatible with
even strong contextualism.
Diamond’s account here, then, suggests one way in which Glock’s objection
might be countered on both an exegetical and a substantial level. I do not want to
endorse Diamond’s account unconditionally; it seems to me that more needs to be
said about exactly what these rules look like, how exactly (if at all) a conditional
application of a rule differs from an application of a conditional rule, and perhaps
too in expressing the process in a way that does not beg any questions.39 That said,
however, something like this account, one which explains our arriving at the mean-
ing of the whole by way of hypotheses about the meanings of the parts, seems to
me at least plausible, and also not to conflict with strong contextualism.
Moreover, the force of Glock’s exegetical conclusion here—that Wittgenstein’s
compositionalism suggests he did not hold strong contextualism—is further under-
mined by Glock’s apparent acceptance elsewhere in the same paper that Wittgenstein
did in the Tractatus hold a strong, restrictive understanding of contextualism; a
stance which Glock explains by way of certain features of Wittgenstein’s picture-
theory of propositions and by Wittgenstein’s extraordinary notion of meaning
(Glock 2004: 227–228). Thus, Glock might be taken to acknowledge that there are,
WITTGENSTEIN, CONTEXTUALISM, AND NONSENSE 117

after all, good reasons to attribute to the early Wittgenstein strong contextualism.
Furthermore, and on the same kind of ad hominem note, it might be thought to be
undermined, too, by Glock’s discussion elsewhere of the Tractatus’s compositional-
ism as forming the “implicit rationale” for Wittgenstein’s early strong contextualism
(even if Glock goes on to say that, as a rationale, it is not strong enough to justify the
strong, restrictive view) (Glock 1996: 87). Whatever the force of those two points,
however, it simply is not the case that strong contextualism and compositionalism
are obviously in conflict, such that one would be forced to abandon one or other
position; again, if Glock’s point here is to work, more argument is needed.

VII.
I have argued that Glock’s criticisms of the austere view of nonsense, and of
the attribution of that view, at least to the early Wittgenstein, by way of the con-
text-principle, do not succeed. Some of those arguments—for instance, against
Glock’s way of presenting the austere view itself, and also the contrast between the
forms of contextualism Glock distinguishes—apply equally to Glock’s arguments
against attributing the austere view to the later Wittgenstein too. There, Glock
claims that Wittgenstein’s contextualism is far weaker than the version present in
the Tractatus:
[I]n the Investigations Wittgenstein quotes Frege’s restrictive principle with
approval. . . . But, with occasional exceptions, Wittgenstein explains the
context-principle in a non-restrictive way, one that is compatible with the
idea that individual words can mean something without actually occurring
in a proposition. . . . What he insists on is that they must be capable of oc-
curring in propositions. (Glock 2004: 229)
There, Glock claims, Wittgenstein’s view is that “the meaning of a word is deter-
mined by how it can be used in sentences,” that a word has a meaning if it has a
use. And, Glock emphasizes, “[t]here is a difference between having a use in the
language and being actually used on a particular occasion” (Glock 2004: 229).
Hence, words even in the context of a nonsense-sentence can, in that sense, have
a meaning.
Glock recognises that “New Wittgensteinians,” as he calls them, would deny that
that is the extent of Wittgenstein’s contextualism in his later work. Thus, the stricture
is applied once more at the level of sentences (which may consist of only one word),
and the range of variables making up the relevant context is extended to include
the “whole language-game,” though, as Lars Hertzberg writes, “there is no way of
determining in advance what contextual considerations will be relevant.”40 But what
Glock does not recognise is that his account of Wittgenstein’s contextualism and
of nonsense is susceptible to a similar objection to that put forward by Diamond
against Hacker in relation to the Tractatus (Diamond 2005). That is, in order for
Glock to make the case for the existence of combinatorial nonsense, consisting of
prohibited combinations of words, he needs the words within a nonsense-sentence
118 EDMUND DAIN

to be capable of having meaning in a sense over-and-above the sense in which

they simply have a use in the language; for it may be the case that the words have
a use in the language, but are not being used in that way here, in some nonsense-
sentence, nor in any other way, and in that case their nonsensicality would be due
not to the meaning they do have, but to their not having any meaning at all in this
occurrence. So, like Hacker, what Glock needs to supplement his view is either
the claim that words can have at most one meaning, or a violation-conception of
the rules for their use, such that any deviation from that would result in nonsense.
Both of these options are incompatible with linguistic creativity, with the idea that
we can give words new uses, without laying down in advance the rules govern-
ing such uses,41 and Glock provides no evidence for thinking either to be true to
Wittgenstein’s later view.

VIII.
I have argued, then, that Glock’s objections to the austere view of nonsense, and
to its ascription to Wittgenstein, early and later, by way of the context-principle, fail
for a number of reasons. First, Glock’s presentation of the austere view distorts what
kind of view it is, and in doing so lays down a requirement for a particular kind of
justification which the austere view need not meet. Further, Glock’s presentation
of the context-principle, and of his different versions of it, impose a version of that
principle upon the austere view that it need not accept. Hence, it seems, Glock’s
objections to the austere view based on his understanding of the context-principle
may not even be relevant to the austere view at all. Finally, I have tried to show
that even were we to accept Glock’s way of presenting the matters at stake, still his
arguments against the austere view do not succeed—for instance, because it is not
clear that resolute readers cannot make some kind of appeal to a technical notion
of meaning in the Tractatus (though such a notion will have to be, in some sense,
“overcome”), or because it is not clear that strong contextualism and composition-
alism are incompatible as Glock suggests. I have also wanted to suggest, in the
previous section, that Wittgenstein’s later contextualism is not at all what Glock
suggests, but that even were it so, that would not suffice to establish the case for
Glock’s understanding of nonsense there. Rather, Glock’s account, like Hacker’s,
must be supplemented with a further idea, and it is implausible to attribute this idea
to later Wittgenstein. These points serve to undermine Glock’s arguments for point
(2), as laid out above (section III), and their substantial counterparts (B) and (C).
Without those points, and without the mispresentation of the austere view assumed
in Glock’s paper, Glock’s claim that the austere view lacks independent plausibility
((3) and (D)) is also undermined. Hence, Glock’s case against austerity fails.
In a debate that has too often been characterized, on both sides, by a lack of
genuine engagement between proponents of differing readings, this attempt to
show how Glock’s criticisms of the austere view of nonsense fail not only at a
very general level, but also as it were on their own terms, seems to me especially
important. More positively, however, I have also tried, among other things, in
WITTGENSTEIN, CONTEXTUALISM, AND NONSENSE 119

arguing against Glock, to shed some light on the kind of view that the austere
view is, as well as on the relations that might obtain between that view and two
other central aspects of Wittgenstein’s thought and of the philosophy of language:
contextualism and compositionalism.42

ENDNOTES
1. That view, as I see it, has both a trivial and a non-trivial aspect. The trivial aspect is that
any string of signs could, by appropriate assignment, be given a meaning, and hence that, if
such a string is nonsense, that will be because we have failed to make just such an assign-
ment. The non-trivial aspect is this: there is no further, non-trivial story to be told, and so
nonsense is only ever a matter of our failure to give signs a meaning. Hence, on this view,
logically speaking, all nonsense is on a par. The view is originally outlined in Diamond
1995: 95–114. See also, e.g., Conant 2000. Glock also finds two strands to the austere view
(see below).
2. Glock (2007: 56–57) goes so far as to suggest that if Wittgenstein held the austere view
of nonsense, that would provide reason for “abandoning . . . the philosophical study of his
writings.”
3. Resolute readers are those who, like Diamond and Conant, reject both the idea that
Wittgenstein is concerned in the Tractatus to communicate ineffable insights of any kind via
his nonsensical “propositions” and the idea that some theory of sense is required in order to
recognize those nonsensical “propositions” as nonsense. See especially Conant and Diamond
2004: 47–48.
4. See Glock 2004: 222, where he writes: “[B]oth early and late [Wittgenstein] allowed
that nonsense can result not just from failure to assign a meaning, but also from combining
meaningful expressions in a way that is prohibited by the rules for the use of these expres-
sions.” See also Hacker 2000 and 2003
5. Glock 2004: 222. That Glock means logically distinct kinds here is to be inferred from
the material preceding this remark.
6. Cf. Morris and Dodd 2007: §4. Similarly to Glock, Morris and Dodd distinguish two
aspects to the austere view: (i) that nonsense arises only from the failure to give words a
meaning, and (ii) that nonsense is only ever sheer lack of sense. Frege, Morris and Dodd
claim (apparently on the basis of TLP 5.4733 and their understanding of the twin demands
of compositionality and the context-principle), held (ii) but not (i), endorsing the idea of
combinatorial nonsense, whereas Wittgenstein in the Tractatus accepts both (i) and (ii).
But, again similarly to Glock, Morris and Dodd argue that Wittgenstein could only have
held (i) on the basis of a strict form of contextualism, itself grounded in or motivated by the
general requirement that a proposition must share a logical form with the reality it depicts.
This seems to me to involve no less a misunderstanding of the austere view than Glock’s
(see section IV below).
7. Frege’s context-principle is nowhere explicitly adapted by him to his later distinction
between sense and meaning (Sinn and Bedeutung). Though the principle does appear in
Frege’s later work (i.e., post the Sinn/Bedeutung distinction), there is a question-mark over
how it might be so adapted and, indeed, over whether it can be or should be so adapted. (Much
120 EDMUND DAIN

of the evidence of Frege’s later contextualism is compiled in Janssen 2001, see especially
125ff.) Wittgenstein not only reformulates the sense and meaning distinction in the Tractatus,
but also adapts the context-principle to it. Although Tractatus 3.3 talks explicitly of names,
I do not think it absurd to hold that Wittgenstein already in this work is attacking the view
that all words function as names for entities, just as he does in the opening sections of the
Investigations. Wittgenstein’s reference to Frege’s context-principle in the Investigations
comes at §49.
8. See, for instance, Diamond 1995: especially 97–100 (although this paper of Diamond’s
is concerned largely with the discussion of Frege, it is also intended to suggest parallels
with Wittgenstein’s Tractatus (see 112 and n7)); and Conant 2000: especially section 5,
189–195.
9. Glock 2004: 227–228 and 229. Since Glock seems to note the point that “it is individual
words rather than whole sentences that have a meaning” (229; similarly, 226) as though it
counts against strong contextualism, it might be thought that he associates the latter view
with the absurd position (no longer, I think, a form of contextualism at all) that the individual
words of a sentence no more have a meaning than do the individual letters of a word. See
also here Dummett 1973: 3.
10. The most charitable interpretation of Glock is that he sees the latter view as one way
of holding a weak contextualism. However, I shall argue that his argument depends on as-
sociating the two in a much stronger sense than that.
11. So, for instance, Wittgenstein writes in that passage that a proposition “is understood
by anyone who understands its constituents” (4.024); or that translation between languages
proceeds not by translating whole propositions “but merely by translating the constituents
of propositions” (4.025); or that a proposition “must use old expressions to communicate a
new sense” (4.03). See below, section VI.
12. Glock 2004: 228. Glock also notes the incompatibility of strong contextualism and
compositionalism as a substantial (not simply exegetical) objection to strong contextualism
on 226–227.
13. Glock 2004: 228 writes as follows: “This [the compositionalism of Tractatus 4.024–
4.026] suggests that the early Wittgenstein did not take TLP 3.3 [Wittgenstein’s Tractarian
context-principle] as literally as the proponents of the austere conception suppose, and as
they themselves need to do in order to construct the aforementioned argument against com-
binatorial nonsense.” (The argument Glock refers to here is given at the bottom of 225 and
is cited here below, section IV.) Nevertheless, Glock also seems to concede that the early
Wittgenstein did hold the strong view (227 and 228), but argues that that view involves a
notion of meaning incompatible with the aims of a resolute reading. See below, n17.
14. But also to those, like Adrian Moore, whose allegiance is less clear, but who nevertheless
ascribe to Wittgenstein an austere view. See Moore 2003.
15. Similar arguments to some of Glock’s here based around Wittgenstein’s interpretation of
the context-principle are rehearsed by Genia Schönbaumsfeld in her paper, ‘Is Wittgenstein’s
Ladder Real?’ (2005), specifically in relation to Conant’s work on the Tractatus. Like
Glock, Schönbaumsfeld claims that if we take Wittgenstein’s Tractarian context-principle
“absolutely literally”—in the way presupposed by Conant’s reading, as implying that
“words only have meaning in the context of a proposition”—then (1) “it just has to be
false”; (2) it “contradict[s] the later Wittgenstein’s dictum that the meaning of a word is its
WITTGENSTEIN, CONTEXTUALISM, AND NONSENSE 121

use in the language”; (3) it “renders superfluous the employment of dictionaries”; and (4)
it “leaves it altogether mysterious how a sentence gets to mean anything in the first place”
(Schönbaumsfeld 2005: 17–20). Much of my argument against Glock is also applicable to
Schönbaumsfeld’s objections: thus, Glock makes points similar to (1) and (2), and point (4),
though distinct from any Glock makes, effectively insists on the need for a strong understand-
ing of the consequences of compositionalism and is susceptible to similar objections to those
outlined below. Schönbaumsfeld’s third point relies on equating (as Glock also seems to at
one point) strong contextualism with the absurd view that the individual words of a sentence
no more have a meaning than do the individual letters of a word (see above, n9), and would
only be true if strong contextualism were indeed incompatible with compositionalism. But
of course, the strong principle actually insists on compositionalism.
16. The latter view, Glock associates with the position he attributes to the later Wittgenstein,
and which Glock himself endorses, that a word has a meaning if it is capable of occurring
in a proposition—if, that is, it has a (rule-governed) use in the language.
17. As noted above (n13), it is not entirely clear which version of contextualism Glock
thinks is to be found in the Tractatus. On the one hand, Glock 2004 writes of “the early
Wittgenstein’s [strong] restrictive principle” (227), and explains its restrictiveness in terms
of Wittgenstein’s early notion of meaning and in terms of the picture-theory (228). On the
other hand, Glock asserts that the compositionalism of Tractatus 4.024–4.03 contradicts
the strong, restrictive principle, suggesting “that the early Wittgenstein did not take TLP
3.3 [Wittgenstein’s context-principle] as literally as proponents of the austere conception
suppose” (228). (The purpose of the latter point may be to suggest that Wittgenstein already
in the Tractatus had reasons to be moving towards the weak version of contextualism Glock
claims is to be found in the Investigations. A version of the point also appears as part of
Glock’s substantial case against austerity (226–227).) Glock’s argument, as I reconstruct it
above, covers both eventualities.
18. See above, section I.
19. Presumably Glock means that they will be nonsense in some sense over-and-above that in-
volved in our simply having failed to give the words a meaning in that context and to date.
20. Again, the words are meaningful here in the sense that they have a use in the language,
even if they are on such occasions being misused, and so cannot be said to have, or “stand
for,” a meaning.
21. See above, section I.
22. Glock 2004: 222 and 237. As noted in my 2006: 100, the latter objection, that the austere
view of nonsense is incompatible with Wittgenstein’s later use of reductio ad absurdum
arguments, ignores the fact that the problem in employing a nonsense-sentence within the
context of a reductio is presumably not what kind of nonsense it is, but that it is nonsense
at all and so lacks a truth-value. Thus, Glock’s objection would apply equally to his own
account of nonsense. Indeed, Glock acknowledges that this remains an unresolved issue on
his view (2004: 222).
23. Nevertheless, there is some confusion about the sense of conflict between those two views
in Glock. Thus, in the same paper (2004: 227) Glock also describes strong contextualism
as the “flipside” of compositionalism: “The flipside of this compositionalism is that the
role of names is to contribute to the determination of the sense of an elementary proposi-
tion. Outside that context, Wittgenstein seems to have held, they cannot have such a role.”
122 EDMUND DAIN

Likewise, elsewhere Glock refers to compositionalism as forming the “implicit rationale”

behind strong contextualism, even if, as Glock goes on to say, as such a rationale, it proves
insufficient (1996: 87).
24. It should be noted that this set of arguments does not exhaust the content of Glock’s paper.
There is a great deal more there worth discussing, and a great deal more too that I would
wish to dispute, but doing so would take me too far from my purposes in this paper.
25. The suggestion that Glock treats the austere view as itself amounting to a substantial
philosophical position is also made by William H. Brenner in his review of the collection in
which Glock’s paper appears. Brenner (2005: 380) writes: “I believe that Glock’s criticisms
of Diamond depend on ascribing to her some special, draconian notion of nonsensicality,
over and above the ordinary notion of having failed to say anything.”
26. See also here Glock 2007: 56 and 64n9.
27. Thus, far from involving some theory of sense articulating the conditions a sentence must
meet in order to be considered senseful, the austere view involves nothing more than, or
merely articulates one aspect of, “our ordinary capacity to think and speak” (Conant 2005:
51–52). It is, as Conant argues, part (a very basic part) of that capacity that we are able to
tell when a sign with which we are familiar is employed in a sentence in a way with which
we are not; if we do come to understand the sentence, that will be because we have been
able to grasp the new use; but if we do not come to understand it, if it resists all our attempts
to do so, we may come to suspect that the sign has not been given a meaning at all, and so
that the sentence is in fact nonsense.
28. That, however, is not to say that that fact is not worth emphasizing, nor that the view
itself is not worth defending, precisely because and in spite of its apparent triviality, it is
nevertheless very often the case that it is assumed that nonsense can be more substantial
than that, and for a variety of purposes. Moreover, that acknowledgement brings the view
into line with what the later Wittgenstein has to say about what philosophy does result in,
or consist of (PI §§126–128).
29. Thus, the rhetorical role played by Glock’s argument (2004: 225; quoted above, sec-
tion IV) is crucial; it is the construction of that argument on behalf of the austere view that
makes it seem as if strong contextualism is what is required when in fact nothing nearly so
strong is.
30. That might seem like little consolation, since Frege’s later view here is widely regarded
as false, as a retrograde step. My point here, however, is just that Diamond clearly does
not think that the austere view of nonsense must deny that words can be used meaningfully
outside propositions, and so need not take the strong view. Rather, austerity only rejects
certain ways—such as Glock’s—of cashing out the idea that words can have meanings in
contexts other than their immediate use in propositions.
31. Diamond 1995: 109. Of course, were Diamond talking of the early Wittgenstein here that
use of the word ‘senseless’ might be misleading, given his distinction between ‘senseless’
and ‘nonsense.’
32. Diamond 1995: 109. See also Diamond 1995: 79, 81.
33. A very different kind of response to Glock here, one which could accommodate such
apparently meaningful uses of words within the Tractatus’s account, is suggested by Michael
Kremer 2002.
WITTGENSTEIN, CONTEXTUALISM, AND NONSENSE 123

34. Glock 2004: 226. Again, Glock goes on to note that a further reason for thinking
strong contextualism wrong is that it clashes with compositionalism; I discuss this below
in section VI.
35. Diamond 1995: 98. (Note, by the way, the appeal Diamond makes here to compositional-
ism in her explanation of (some form of) contextualism.)
36. One reason for thinking that might be this: that there would then be no such thing as
recognising the occurrence of a sign—such as ‘meaning’—as a logical element at all, as one
symbol rather than another. That is correct, but it ignores the idea—central to Wittgenstein’s
work—that we very often do, in practicing philosophy, imagine ourselves to make sense,
imagine our words to make sense, where in fact we, and they, make none: our use of words
can have all the appearance to us of making sense when all along we are failing to say any-
thing. What Glock ignores is just the idea that Wittgenstein’s words might be written with the
intention that they induce in the reader the illusion that they are being put to some technical
use, but that they fall apart when we try to make sense of them from the inside as it were.
An apparently technical notion of meaning might be utilised in this way, and may bring us
to an understanding of nonsense as not consisting of words with meaning. But abandoning
the idea that there was any sense behind those remarks would leave us not abandoning the
austere view too, since that view follows not from a theory of sense or meaning, but from
the rejection of such a theory.
37. Denis McManus, for instance, is one reader who does consider the context-principle to
be elucidatory nonsense (in conversation). Glock’s point here may just be that the context-
principle must eventually be discarded by resolute readers.
38. Of course, one does not wish to leave this an utter mystery, but it might be objected that
Glock places a demand for an altogether different, and far stronger, kind of explanation
here.
39. Thus, there is a difference between the thought that “the object ‘Venus’ stands for, what-
ever that is, stands in whatever relation it is the relational expression stands for to whatever
object it is ‘Mercury’ stands for” and the thought that “Venus is more massive than Mercury.”
The thought typically expressed by Diamond’s example is the latter, but Diamond cannot
simply help herself to a direct statement of that thought in describing how we arrive at it.
Thanks to Alessandra Tanesini for bringing these points to my attention.
40. Hertzberg 2005: 6. (NB: This quotation is taken from a passage omitted from the 2001
version.)
41. I argue this point at greater length in chapter two of my Nonsense and the New Wittgen-
stein (manuscript in preparation).
42. This paper is a much revised and extended treatment of the issues discussed in my
“Contextualism and Nonsense in Wittgenstein’s Tractatus,” originally presented at the
Philosophical Society of Southern Africa (PSSA) Annual Conference 2006 and subse-
quently published (with selected proceedings) in the South African Journal of Philosophy,
25(2): 91–101. I am grateful to the audience on that occasion (especially D. H. Mellor and
David Spurrett) for their questions, and to the editor of SAJP for permission to reproduce
material from that paper here. This paper was also presented at the University of Chicago’s
Wittgenstein Workshops and at Hamilton College, and I am grateful to audiences on both
occasions, as well as to Silver Bronzo, James Conant, Cora Diamond, Michael Kremer, Marie
124 EDMUND DAIN

McGinn, Denis McManus, Genia Schönbaumsfeld, and Alessandra Tanesini for comments,
correspondence, or conversations on the issues here discussed.

BIBLIOGRAPHY
Brenner, William H. 2005. “Review of Wittgenstein and Scepticism and Wittgenstein at
Work: Method in the Philosophical Investigations.” Philosophical Investigations 28(4):
375–380.
Conant, James. 2000. “Elucidation and Nonsense in Wittgenstein’s Tractatus.” In The
New Wittgenstein, ed. Alice Crary and Rupert Read. London & New York: Routledge,
174–217.
. 2005. “What ‘Ethics’ in the Tractatus Is Not.” In Religion and Wittgenstein’s
Legacy, ed. D. Z. Phillips. Aldershot: Ashgate, 39–88.
Conant, James, and Cora Diamond. 2004. “On Reading the Tractatus Resolutely: Reply to
Meredith Williams and Peter Sullivan.” In Wittgenstein’s Lasting Significance, ed. Max
Kölbel and Bernard Weiss. London and New York: Routledge, 46–99.
Dain, Edmund. 2006. “Contextualism and Nonsense in Wittgenstein’s Tractatus.” South
African Journal of Philosophy 25(2): 91–101.
. Manuscript in preparation. Nonsense and the New Wittgenstein.
Diamond, Cora. 1995. The Realistic Spirit: Wittgenstein, Philosophy, and the Mind. Cam-
bridge, Mass.: MIT Press.
. 2005. “Logical Syntax in Wittgenstein’s Tractatus.” The Philosophical Quarterly
55(218): 78–89.
Dummett, Michael. 1973. Frege: Philosophy of Language. London: Duckworth.
Frege, Gottlob. 1980. The Foundations of Arithmetic: A Logico-mathematical Enquiry into
the Concept of Number, trans. J. L. Austin. Oxford: Basil Blackwell.
Glock, Hans-Johann. 1996. A Wittgenstein Dictionary. Oxford: Blackwell.
. 2004. “All Kinds of Nonsense.” In Wittgenstein at Work: Method in the Philo-
sophical Investigations, ed. Erich Ammereller and Eugen Fischer. London & New York:
Routledge, 221–245.
. 2007. “Perspectives on Wittgenstein: An Intermittently Opinionated Survey.”
In Wittgenstein and his Interpreters, ed. Guy Kahane, Edward Kanterian, and Oskari
Kuusela. Oxford: Blackwell, 37–65.
Hacker, P. M. S. 2000. “Was He Trying to Whistle It?” In The New Wittgenstein, ed. Alice
Crary and Rupert Read. London & New York: Routledge, 353–388.
. 2003. “Wittgenstein, Carnap and the New American Wittgensteinians.” The
Philosophical Quarterly 53(210): 1–23.
Hertzberg, Lars. 2005. “The Sense is Where You Find It.” Available at: http://www.abo
.fi/fak/hf/filosofi/Staff/lhertzbe/The_Sense_Is_Where_You_Find_It.doc [Also published
in Wittgenstein in America, ed., Timothy McCarthy and Sean C. Stidd. 2001. Oxford:
Clarendon Press, 90–103.]
Janssen, Theo M.V. 2001. “Frege, Contextuality and Compositionality.” Journal of Logic,
Language and Information 10: 115–136.
WITTGENSTEIN, CONTEXTUALISM, AND NONSENSE 125

Kremer, Michael. 2002. “Mathematics and Meaning in Wittgenstein’s Tractatus.” Philo-

sophical Investigations 25(3): 272–303.
Moore, Adrian. 2003. “Ineffability and Nonsense.” Proceedings of the Aristotelian Society:
Supplementary Volume LXXVII: 169–193.
Morris, Michael and Julian Dodd. 2007. “Mysticism and Nonsense in the Tractatus.” Eu-
ropean Journal of Philosophy: 1–30.
Read, Rupert. 2000. “What ‘There Can Be No Such Thing As Meaning Anything By Any
Word’ Could Possibly Mean.” In The New Wittgenstein, ed. Alice Crary and Rupert
Read. London and New York: Routledge, 74–82.
Schönbaumsfeld, Genia. 2005. “Is Wittgenstein’s Ladder Real?” Unpublished manuscript.
Wittgenstein, Ludwig. 1958. Philosophical Investigations, trans. G. E. M. Anscombe.
Oxford: Blackwell.
. 1969. The Blue and Brown Books: Preliminary Studies for the “Philosophical
Investigations.” Oxford: Blackwell.
. 1974. Tractatus Logico-Philosophicus, trans. D. F. Pears and B. F. McGuinness.
London: Routledge.
. 1975. Philosophical Remarks, ed. Rush Rhees, trans. Raymond Hargreaves and
Roger White. Oxford: Basil Blackwell.
. 1978. Philosophical Grammar, ed. Rush Rhees, trans. Anthony Kenny. Berkeley:
University of California Press.
. 1993. “Wittgenstein’s Lectures in 1930–1933.” Notes by G. E. Moore. In
Philosophical Occasions 1912–1951, ed. James C. Klagge and Alfred Nordmann.
Indianapolis: Hackett, 46–114.