Content and Community

Author(s): Harry A. Lewis and Andrew Woodfield

Source: Proceedings of the Aristotelian Society, Supplementary Volumes, Vol. 59 (1985), pp.
177-214
Published by: Blackwell Publishing on behalf of The Aristotelian Society
Stable URL: http://www.jstor.org/stable/4106754
Accessed: 22/08/2010 14:42
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at
http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you
may use content in the JSTOR archive only for your personal, non-commercial use.
Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at
http://www.jstor.org/action/showPublisher?publisherCode=black.
Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.

The Aristotelian Society and Blackwell Publishing are collaborating with JSTOR to digitize, preserve and
extend access to Proceedings of the Aristotelian Society, Supplementary Volumes.

http://www.jstor.org

CONTENT

AND COMMUNITY

Harry A. Lewis and Andrew Woodfield

I-Harry A. Lewis
I
Can a being have thoughts and yet not be a member of a
community? Is the content of individuals' thoughts determined
by features of the community of which they are members? These
are distinct questions. A link would be made between the second
question and the first if any fixing of content could be shown to
require a community; for contentless thoughts would be no
thoughts at all. A positive answer to either question may seem
obvious, or paradoxical, depending how it is taken. A hermit
can be allowed thoughts, and if any human being lacks a
community, surely he does. But this example might be held not
to break a necessary tie between thought and community, for
the hermit will have developed his capacity for thought in the
bosom of his family, and if we read his writings, even after his
death, we could be held to enrol him in our community. A thesis
so plastic will not be easy to refute. An empiricaldisproof would
appear not to be open to us: for how could we detect thought if
not by conversing, and thus creating at least a 'community' of
two? But a step or two of argument is needed to show that the
very notion of a solitary thinker is incoherent, and a further step
to show that any thinker may belong to a fully-fledged
community. In this paper I explore some aspects of recent
arguments to community.
I take a being to have thoughts just in case attitudes
('propositional' attitudes) can be truly ascribed to it. I speak of
'content' to distinguish the attitude proper (whether it be belief,
remembering, or wondering whether) from the Fregean thought
or Russellian proposition to which the attitude is directed. I see
no need for a prior commitment to a 'theory' of content, if such is
possible, or to a view of the logical form and ontology of attitudeascriptions in general. In particular, I take no view as to whether
content is 'private' or 'public'-since I regard the question as

178

A. LEWIS

I-HARRY

confused. ('Content' can be ascribed to what one person tells

another just as much as to what a single person believes.) And I
shall henceforward assume that the being to whom attitudes are
ascribed is human. On the other hand, I am not prepared to take
the notion of 'community' as well understood. The arguments
that interest me purport to reveal the necessity of community for
thought, or for language and so for thoughts too. I have
indicated already that the notion of membership of a community
is at risk of being stretched. In addition, the notions of social
group that are invoked are not all the same, nor are the functions
assigned to such groups in the argument.
II
A direct connection between thoughts and community was
claimed by Tyler Burge in his paper 'Individualism and the
Mental'.' Burge purports to show that a person's mental
contents are constituted in part by his social environment: 'No
man's intentional phenomena are insular. Every man is a piece
of the social continent, a part of the social main.' (p. 87). The
central argument that he used has been rehearsed before,2 but it
is of sufficient interest to bear repetition. My purpose will be
served by adapting one of Burge's simpler examples, while
echoing his presentation in other respects. We are to imagine
that Jane has a large number of attitudes commonly attributed
with content clauses containing the word 'sofa' in oblique
occurrence. For example, she thinks (correctly) that sofas are
items of furniture, that they are for sitting on, and that they are
heavily upholstered. In addition to these unsurprising attitudes,
she thinks falsely that John has a new sofa in his living room (for
in fact, it is an armchair). Her false belief results from the
absence from her experience of any corrective to the application
of the word 'sofa' to single-seaters. She talks to John of his new
'sofa', and he points out that it is an armchair: sofas are not
'French, P. A., Uehling, T. E. and Wettstein, H. K. (editors) MidwestStudiesin
Philosophy,Volume 4 (1979) Studiesin Metaphysics,pp. 73-121.
2
By Philip Pettit, in 'Wittgenstein, Individualism and the Mental', at pp. 446-55 of
andthePhilosophy
Weingartner, Paul (editor), Epistemology
of Science:Proceedings of the
7th International Wittgenstein Symposium, Vienna: Halder-Pichler-Tempsky, 1983;
by Steven Stich, in his 'On the Ascription of Content' at pp. 200-203 in Woodfield, A.
(editor) ThoughtandObject,Oxford: Clarendon Press (1982); and by Andrew Woodfield
in 'Thought and the Social Community', Inquiry25 (1982) pp. 435-50.

CONTENT AND CX)MMUNITY

179

single-seaters, as any dictionary could have told her. Jane is

surprised, but gives up her belief. This story corresponds, so far,
to the 'first step' in Burge's 'thought-experiment'. In the second
step, we are invited to imagine the circumstances altered:
whereas nothing has been different in Jane's physical history
and non-intentional mental phenomena, furniture-salesmen,
lexicographers and informed laymen apply 'sofa' not only to
sofas but also to armchairs. The standard use of the term in the
imagined circumstances is to be conceived to encompass Jane's
actual misuse. The third step is an interpretation of this latter
case. In the counterfactual situation, Jane lacks some-probably
all-of the attitudes commonly attributed with content clauses
containing 'sofa' in oblique occurrence. She lacks the occurrent
thoughts or beliefs that John has a new sofa in his living room,
that many of her other friends have sofas, that sofas are for sitting
on, and so on. Thus Jane's counterfactual attitudes differ from
her actual ones, but only in virtue of a difference in social
environment. Burge claims wide scope for this argument: it can
get under way 'in any case where it is intuitively possible to
attribute a mental state or event whose content involves a notion
that the subject incompletely understands'. He regards such
incomplete understanding as common: 'One need only to
thumb through a dictionary for an hour or so to develop a sense
of the extent to which one's beliefs are infected by incomplete
understanding' (p. 79).
What is the 'community' to which this argument appeals, and
what is its function in the argument? The community includes
specialists (the furniture-salesmen of the example), lexicographers and informed laymen. These people agree on their use
of the term in question; but there is no suggestion that they form
part of the community solely by virtue of that agreement. On
the other hand, their function is to establish a 'standard use'
over against Jane's. In addition, they define 'the' notion of

sofa.

Burge moves easily between talk of divergent usage and talk of

divergent belief, as his remark on dictionaries shows. In fact he
goes further: for he embraces the paradox that we (individuals)
'think with' notions of which our mastery is incomplete. I find no
difficulty in accepting that we severally fail to master thejointlyowned notion of sofa, or of harpsichord. (To refer to this

180

I-HARRY

A. LEWIS

phenomenon as the 'linguistic division of labour', after Putnam,3

would only serve to promote the conflation of belief and usage
that I have just noted.) What seems perverse, indeed to beg the
question in favour of Burge, is to claim that we severally'think
with' this jointly-owned notion. Jane would be held to 'think
with' a notion of sofas that excluded armchairs, in her very
judgment that an armchair was a sofa. The alternative would be
to recognize that Jane's notion of sofa was different from that of
others, and that she could only 'think with' her own notion. In
certain cases, it might still be convenient for a standard user of
'sofa' to report Jane's beliefs by using that term; but this would
be a practical decision (as Woodfield suggests).4 Burge offers
normal practice as a reason for rejecting the attribution of a
notion that just catches the misconception. However, practice
has many motives. He seems also to hold that attributing such a
notion to an individual will involve treating her beliefs as true. If
the only way of ascribing unusual notions to an individual is to
allow all her beliefs to be true, then the view that we think with
our own notions seems to be in trouble. (I shall return to this
point later.)
Andrew Woodfield has argued persuasively that Burge's
argument fails,5 through inattention to the variety of languages
that are involved in the imagined cases. Woodfield shows that
the phenomena in question allow of clear characterization
where Jane's attitude is identified by way of a linguistic
disposition, and the language in which the attitude is reported is
thought of as the same as Jane's or alternatively as relevantly
different. One effect of his critique is to split off the social
elements involved from the attitudes. The role of social factors in
Burge's examples is to identify one language that might be used
to reportJane's state of mind. A similar conclusion can be drawn
if we replace the idea of the divergent individual by that of a
small community with its own dialect. Such a dialect-owning
community could diverge from its host-community in just the
way that Jane diverged from hers, whereas any problems
thought to derive from the fact that Jane is a single individual
Papers,Volume2: Mind, Languageand Reality,Cambridge
3Putnam, H., Philosophical
University Press (1975) pp. 227 ff.
4Woodfield, op. cit. p. 446.
5In his paper 'Thought and the Social Community' cited above.

CONTENT AND COMMUNITY

181

could not exist for a community. But its usage would by Burge's
argument fail to determine the contents of the attitudes
expressed by it, until reference was made to the ways of the hostcommunity. The only contrast needed in order to characterize
the phenomena presented by Burge is that between languages.
He makes no use of the contrast between individual and
community except to that end.
The claimed direct link between content and community is
thus revealed as mediated by the notion of language identity.
This suggests that the argument needs to be re-cast. The link
taken for granted by Burge between (on the one hand) the social
factors that fix the language and the notions it expresses and (on
the other) the contents of individuals' thoughts, needs to be
made good. Even if it established its ostensible conclusion,
Burge's case would fall short of showing that there could not be a
solitary thinker; for it is an unargued assumption that his
thinkers are members of communities. He does not suggest that
there would be no attitudes without community, but only that
where a person already belongs to a community, some facts
about the community enter into the determination of the
content of her thoughts.
III
We shall consider a less direct argument to show the impossibility
of a solitary thinker, to wit: any thinker must have a language;
any language-user must have a community; so any thinker must
have a community.
We should not assimilate thinking too readily to a use of
language. There are many attitudes whose content is not
literally expressed by any sentence of their owner's language. A
variety of examples may be used to argue for this observation,
if indeed argument is required. If there are de re attitudes
properly expressed using indexical pronouns, their content will
not typically be expressed by any complete sentence, where
what the sentence expresses is taken to be what it brings to its
contexts. Any figurative or metaphorical use of language may
lend itself to the expression of attitudes whose contents cannot be
'translated' into another sentence taken literally. Nonetheless, if
I say 'It's raining and it's not raining', what I think, and what I
convey, may be true, even though the sentence I utter, taken

182

I-HARRY

A. LEWIS

literally, is false. Does anyone wish to insist that in such a case

the speaker and the hearer must have in mind some other
sentence, whose sense 'It's raining and it's not raining' used
figuratively, precisely captures? For another class of examples,
consider what might be called 'indexical predicates' as in
'Peter's shirt was this colour', 'Mary jumped up like this'. Does
anyone wish to insist that understanding such sentences, as
uttered with appropriate gestures, requires that speaker and
hearer apply or invent a word to express what the whole
predicate, complete with context, contrives to convey? A
person's beliefs are sometimes identified with that subset of the
sentences of her language that for her express truths. But many
things held true cannot be found a sentence of the believer's
language that (taken literally) expresses them. The relation
between content and language is looser than that. Thus a strong
form of the argument from thought to language is invalid: it is
not the case that all contents must be expressible by literal or
portable uses of sentences. But the evidence just adduced has no
tendency to show that thought is possible for a creature without
language, since it draws on examples where whatever is
expressed or conveyed exploits a use of language.
For a defence of the first step of our indirect argument, we
may turn to the view defended by Donald Davidson in his
'Thought and Talk',6 namely, that any thinker must be an
interpreter of another's speech. I do not need to consider all of
Davidson's argument, but only to review the social elements
involved.
The alleged relation between thought and interpretation is
indirect: an 'interpreter' is not said to be interpreting whenever
he thinks, just as our hermit is not forever learning from his
parents. Davidson is not concerned with the development of the
capacity for thought in the individual's lifetime, but with the
general conditions for thought.' This enables us to suggest that
the capacities in question might be exercised in thought while
remaining unexercised in interpretation. Thus insofar as a social
6Davidson, D., Inquiriesinto Truthand
Oxford: Clarendon Press (1984)
Interpretation,
pp. 155-70.
SI recognizethat this account of Davidson'spurposemay be at odds with his talk (on p.
170) of the 'emergence' of the contrast between truth and error. see below,
p. 193.

CONTENT AND COMMUNITY

183

element in thought is held to derive from the thinker's

interpreting activities, it could remain a theoretical possibility
rather than a necessity. A similar conclusion might be reached
by another route, if it could be accepted that there are human
beings who both think and use language but have never
practised radical interpretation. Someone who has learned
only one language has arguably never indulged in radical
interpretation, since the latter activity (as described by Davidson)
requires that the interpreter already possess a language into
which the alien sentences are translated. The capacity for
radical interpretation in such a person would remain a capacity
only. It is only half an answer to this to propose that normal
communication between users of 'the same' natural language
involves translation between idiolects; for the learning of one's
first language cannot be construed as translation. To be sure,
Davidson says that 'all understanding of the speech of another
involves radical interpretation'.8 It follows that he holds the
notion of idiolect intelligible; for he recognises that the
assumption that a fellow-member of the same speech-community
shares one's language requires an empirical grounding, and that
must be found in facts about that individual. I shall be returning
shortly to the topic of idiolects.
Even if we allowed that any thinker must at some time have
practised interpretation, would that show that she must have
been a member of a community? In what ways is interpretation
a community activity? It requires two players; and two
languages. If we stipulate that it is the interpretation of another
person's speech that is in question, then of course the two players
must be different people. On the other hand, without that
stipulation it is not obvious that one person could not play both
parts. An example might be that of a bilingual translating her
own work. The fact that the one human being understands both
languages does not guarantee that the translation is effortless.
Such translation, however, could hardly be termed 'radical', at
least not in Quine's terms ('the translation of a language of a
hitherto untouched people' from which 'all help of interpreters
is excluded'9). Nonetheless we should recognize that inter8'Radical

Interpretation',

at p. 125 in Inquiries into Truth and Intelpretation, op. cil.

9Quine, W. V. 0., f'ord and Object,M.I.T. Press (1960) p. 28.

184

I-HARRY

A. LEWIS

pretation requires in the first place two roles rather than two
people, even if it is only when we credit the bilingual with two
personalities that we can make sense of the hypothesis that both
roles could be played by one person.
If interpretation is taking place, must interpreter and
informant be members of a community? If we hark back to the
forerunner of radical interpretation, Quine's radical translation,
we are apt to assume that they are severally members of
communities, but not that they are fellow-members of any one
community. For the moment we may hold in suspense the
assumption that there are two communities in the background,
responsible for the languages, and focus on the activity of
interpretation. This is usually talked of as a collaborative
activity; for more than the passive observation of speechbehaviour is needed to establish experimentally the patterns of
assents, dissents or neither (Quine), or the pattern of holdings*true (Davidson), that forms the empirical basis of interpretation.
There is, however, a lack of mutuality in this supposedly 'social'
activity, for it is motivated only by the interpreter's need to
associate sentences held true with observable situations or with
the interpreter's occurrent beliefs. There seems no reason to
honour a pair related by no more than the minimal requirements
for interpretation with the title 'community'.
Interpretation requires also two languages. These are usually
represented as the languages of communities; but that they are
communal
languages is a background assumption, not an
Moreover it is not clear that we have a
conclusion.
argued
ready-made argument for the communal nature of language
from the premiss (even if granted) that any language must at
some time have been interpreted. For all this could yield would
be further pairs of interpreters and informants. Although an
actual human community could provide the personnel for such
activities, they could be carried on without it; after all, the facts
of human generation by pairwise relations of mothers and fathers
do not make siblings of us all. Furthermore, both Quine and
Davidson find ultimately that pairs of speakers of 'the same'
natural languages are indulging in radical translation or
interpretation. As I have noted already, this view has to find
intelligible the restriction of a language to a single speaker.
Thus the first step provides no easy transition from thought to

CONTENT AND COMMUNITY

185

community. At best it brings us to the second: the thesis that to

have a language is to be a member of a community. Perhaps, by
another path, we can show that anyone who possesses a
language must also have a community.
Before proceeding we should draw one lesson from the discussion of radical interpretation. It might seem that an activity
requiring a minimum cast of two fails to qualify as communal
on arithmetical groundsalone. If we could find a need for a third
player, or a twenty-second, perhaps it would then be communal. It is not numbers that count, however; the activities in
which the numbered personsare involved, and the concepts we
must invoke to describe those activities adequately, determine
whether they are to be deemed a 'community'. Thus I am
prepared to allow that there could be a community of two; but
also, there could be activities involving a cast of thousands (e.g.
treading this path) that fail to constitute them a community.
Many activities involving language are social (conversing,
debating, commanding) and othersare not ( writing one's diary,
doing a crossword, soliloquy). Thus someone who has a
language need not use it exclusively in social activities.
Moreover it is conceivable that our hermit, having once learned
his language, never again uses it in a social activity. Is there a
sense in which having a language, as distinct from using it,
requires a community? A positive answer is immediately
forthcoming if the only possible languages are languages of
communities. For example, if one holds that any language will
contain observation sentences in the sense of QuineA sentence is observationalinsofaras its truth-value,on any
occasion, would be agreed to by just about any member of
the speech community witnessing the occasion.10
-then indeed it will be communal. However, Quine's approach
appears to take it as an axiom that a necessary condition of
objectivity in perception-reportsis communal agreement. I am
not prepared to accept that without argument, and shall
confront it anon. An alternative reason that might be offeredfor
accepting that there will be social elements in any language, is
that the proper purpose of language is communication (or some
"

Quine, W. V. O., The Rootsof Reference,

Open Court (1974) p. 39.

186

I-HARRY

A. LEWIS

other supposedly social activity). It will be clear by now that I do

not regard such claims as arguments. Language is used for many
things, not all of them social. If, percontra,a proper purpose is one
for which something might never actually be used-as can
certainly happen in the case of artefacts-then the required
conclusion does not follow.
Care over the distinction between having a language and
using it is necessary because of an associated distinction, that
between the community responsible for a language and the
community within which it is used. (The common tendency to
talk of 'the' community blurs this distinction, among others.) An
example would be a bilateral international negotiation-a
transient community of diplomats-in which each side spoke its
native tongue. If the participants were appropriately trained,
communication could be achieved and the negotiation would
proceed; but the communities to which the languages were
answerable would be different from the community within
which the languages were used. (As Davidson observes,"
communication between two parties does not require identity of
language. I note .too that the claim that anyone who has a
language is a member of a community is not the same as the
claim that any language is a community's language. In
particular, a person may use a community's language but be
excluded from it; or fail to understand its language but be
included in it. To date we have found no reason to deem all and
only the speakers of some language to be thereby the members of
a genuine community, rather than a mere set.)

IV
We can now refine the second step, and look for argument to
either of two conclusions: that any language is a community's
language (not just one individual's); or that anyone who has a
language must belong to a community (whether or not her
language is also the community's language). If either conclusion
is true, there will be some latent contradiction in the notion of a
lifelong social isolate who possesses an idiolect: a language all his
own, spoken by no one else.
If social isolation is conceptually, if not psychologically,
" In 'Thought and Talk', op. cit. p. 157.

CONTENT AND COMMUNITY

187

acceptable, possession of an idiolect is not as favourably placed.

It is not enough to define 'idiolect' as 'the language of one
person'. For a helpful discussion, the description must be filled
out to give a hand-hold to arguments about the intelligibility of
the notion. I propose to be generous: let us allow S's idiolect to
have a syntax, to have perceptible sentence-tokens, and to
include sentences that fail to express its owner's beliefs, and
sentences that express falsehoods. It is thus not a 'private'
language in any of the following senses: it is not restricted to
reporting 'private' phenomena such as sensations; it does not
have only private tokens (such as inner speech-acts); and it is not
(at least not by stipulation) unlearnable by others.
My generosity is not motivated. A 'language' so definedas to
express its speaker's beliefs unerringly, would be no language.
Whatever seemed right would indeed be right in that case. The
language could not be learned except by someone with
independent access to those beliefs; and so it could not be
learned. For parallel reasons, a 'language' defined as unerringly
expressing truths could convey nothing to anyone; it would, for
example, be useless in testimony. Thus if the idiolect is to have
the autonomy required of a genuine language, neither of these
approaches to definition is open to us. It might seem inconsistent
to reject them while also rejecting Burge's claim that we 'think
with' communal notions, as I did above. For Burge sees the only
alternative to be to attribute (to Jane) notions that exactly
capture her misconceptions; and this seems tantamount to
attributing to her a language that exactly captures her beliefs, or
alternatively one that cannot fail to express truths. (These
apparent alternatives collapse into each other when what 'seems
right'-is believed-'is right'-is true.) I accept that to attribute
notions exactly capturing misconceptions leads to absurdity,
but reject the thesis that there is no other alternative to
attributing communal notions. The third possibility involves
showing how an idiolect-speaker could detect error-how he
could come to recognize a mistaken classification, for example.
I believe that it can be shown, and I shall attempt to do so
shortly.
Does possession of an idiolect, as defined, reveal its owner as a
member of a community? Empirically, for human speakers, I
have no doubt that the answer is 'yes'. But is this due to more

188

I-HARRY

A. LEWIS

thanthelimitation
ofhumanintelligence?
Letusaskhowsucha
languagemightbe acquired.
IfS, ouridiolectophone,
is to bea
lifelongsocialisolate,he cannothavelearned
it
at
his
mother's
knee.(ButI notethatsomeonecouldlearn
an idiolectsocially.
His parentsmightinventit, and
teachit to him,withouthis
meetinganyotherpersonwhospokeit.) Thus
wemustassume
heinventedit. Thishe dideitherby
replicating
withinhimself
thoseelementsof languagelearninggiven
socially
to therestof
us;orbyfindinganalternative
wayintolanguage.Itwouldbeg
thequestionto saythatthe latteris
impossible;
forthatisjust
whatthehumanrace,takencollectively,
achieved.
Wehaveyet
touncovertheessential
socialelementinlanguagecreation,that
wouldshowhow the racecouldachieve
whatno individual,
however
well-endowed
intellectually,
could.
Anassumption
of thistreatment
ofidiolectshasbeenthatan
idiolect
will be differentfromanyother
language.Nothingin
thedefinitions
proposed
required
this,
however,
andit neednot
beso.Twopeoplecouldsharethesame
idiolect
just
sharethe sameheight;exceptthat 'share' astheycould
is a dangerously
ambiguous
termto use there.If we allowthat the
notionof
idiolect
is intelligible,I cannotsee that we
can
rule
out the
possibility
of the independent
invention
by
two
people
of the
same
language,howeverimprobable
it
may
be.
A versionof Wittgenstein's
commentson his
reporting
languagemight still be thought to sensationa
conclusive
refutation
of thethesisthatanidiolect,asprovide
described,
is
possible.In spiteofmygenerous
description,
it mightbeheld
thata solitaryuser of the 'language'
could
not mustera
distinction
betweencorrectandincorrect
use.Inthecaseofthe
impoverished
'language'of Philosophical
Investigations
258,
Wittgenstein
hadanargument
toshowwhy'whatever
is
going
to
seem
righttomeisright'.Soa defenceofthe
idiolect
should
take
the
formofshowinghowitsusercouldcometo
detecterror.(Itis
not
enoughto claimthatthelanguagecaninfact
bemisused;
a
sense
mustbe givento discovery
of themisuse.)Thiscallsforthe
possibility
of checkingon the correctness
of pastusage.
I am makingthe simplifying
assumption
that the usagein
question
is 'fact-stating
discourse'.
This
abstraction
awayfrom
the
complexity
of therealworldis defensible
ad
hominem;
andI
know
ofnoonewhoallowsthata 'fact-stating'
idiolectmightbe

CONTENT AND COMMUNITY

189

possible, but that (e.g.) irony demands a communal language.'2

The fundamental idea in Wittgenstein's argument is that of
comparison: crudely stated, it is a comparison of the way it
seems to S with the way it is. That statement masks a vital
distinction, however. We should not confuse two quite separate
claims: the one, that S has nothing with which to compare the
way it seems to him; the other, that the detection of error
requires access to the facts. The former claim is one that (I
accept) was made by Wittgenstein of the user of the private
sensation-language.'3 Some have also interpreted his discussion
of rule-following as implying that any idiolect would be as
deficient as the sensation-language."'4 The counter to that claim
requires only the demonstration that S can check his usage
against something else and come to a reasonable decision. To
characterize this process as checking for errors,the acquisition of
the knowledgethat he was wrong, requires the further judgment
that S (eventually) gets it right. Now, as far as I can see, what
Wittgenstein has to say reduces to this: the only possible point of
comparison available to the user of the sensation-language
derives from the judgments of his community. It is a separate
matter whether the community is right, or whether we have to
say: 'Whatever seems right to the community is right, and
there's an end on 't.' (This last is the position ascribed to
Wittgenstein by Crispin Wright.'") The interpretation that
finds in Wittgenstein the view that any rule-following requires
the community is driven also to the conclusion that for
Wittgenstein the sole possible source of objectivity lies in the
community. If one succumbs to the temptation of saying,
resignedly, that to avoid scepticism we must hold that what the
community accepts is, by definition, true, then the conflation of
the two claims is complete. I find just such a conflation not only
'2Cf. Davidson, pp. 164-5.
(translated by G. E. M. Anscombe),
13Wittgenstein, L., PhilosophicalInvestigations
Oxford: Basil Blackwell (1953) sections 258 to 323.
14For example, Peacocke, C., 'Reply [to Baker]: Rule-following: The nature of
Wittgenstein's arguments'; Wright, C., 'Rule-following, Objectivity and the Theory of
Meaning'--both papers to be found in Holtzman, H. and Leich, C. M. (editors)
tofollow a rule,London: Routledge and Kegan Paul, 1981, respectively at
Wittgenstein:
onRulesandPrivateLanguage:an
pp. 72-95 and pp. 99-117; and Kripke, S., Wittgenstein
elementary
exposition,Oxford: Basil Blackwell, 1982.
5 Wright, op. cit. p. 106.

190

I-HARRY

A. LEWIS

in the cited discussions of Wittgenstein, but also in 'Thought and

talk', where Davidson moves from a notion of the 'social truth' of
sentences to 'the idea of an objective, public truth'.'6
So, in defence of my idiolectophone S, I need primarily to
show that he can do as well as a member of a community in
comparing his usage with some standard. I do not need to show
that he is bound to get it right. For not every community is
bound to get it right either. (This embarrassing fact is often
suppressed by talking of the source of comparisons as 'the'
community, presumed unique and unanimous.)
Wittgenstein himself was particularly concerned to deny that
meaning could be grasped 'in a flash':
it mean that one
But what about this consensus-doesn't
human being by himself could not calculate? Well, one
human being could at any rate not calculate just oncein his

life.'7
Taking our cue from this, let us compare the individual's
instantaneous and unchecked judgment both with the same
judgment in the context of communal agreement and with the
same judgment as one of a series by that individual. For I submit
that any reinforcement provided for an individual's momentary
judgment by his community can be paralleled by reinforcement
available from the historyof his own judgments. Schematically:
many minds on one occasion of judgment can be paralleled by
one mind on several occasions. If 'agreement in judgment' is
possible at one moment for many minds, it is also possible for one
mind at many times.18 Appropriate but entirely analogous
assumptions are needed in both cases. Let the judgment in
question be 'This is an oak tree'. If several people assent to that
sentence at the time, they must (in order to agree) refer to the
"

Davidson, op. cit. pp. 165, 170.

of Mathematics(edited by G. H. von
17Wittgenstein, L., Remarkson the Foundations

Wright, R. Rhees and G. E. M. Anscombe, translated by G. E. M. Anscombe), Oxford:

Basil Blackwell, 1956: II 68. (On Wittgenstein'sspecial concern with the instant, see also
Baker, G. P. and Hacker, P. M. S., Scepticism,Rules and Language,Oxford: Basil
Blackwell, 1984; and cf. Lear,J. 'The Disappearing "We"', Proceedings
of theAristotelian
Society,Supplementary Volume 58 (1984), p. 235.)
8 The same proposal is made by Simon Blackburn, 'The Individual Strikes Back',
58 (1984) pp. 281-302, at p. 294. Blackburn'spaper is congenial to the position
Synthese
argued here; it came to my attention too late to be taken properly into account in the
preparation of the present paper.

CONTENT AND COMMUNITY

191

same object, and mean the same by 'oak tree'. If one person
assents to it on several occasions, in order for hisjudgments to be
in agreement, his reference and his language must likewise
concur. The many-person case is not yet revealed as privileged.
We are by now, however, alerted to the ambiguity of 'agree'.
Two people may assent to the same thought, without coming to
an agreement that it is true. The former is a necessary
condition of the latter, but it is not sufficient. If the latter is
conceived as a social act involving two people, there will be no
obvious analogue for a single person. However, a step from the
one to the other is the recognitionof sameness of assent. And the
persisting individual may follow the momentary colleague at
least to this point. As this is the crucial step for comparison and
so for checking, let us look closely at it.
First I must note that the term 'judgment' is misleading. The
kind of agreement that (according to the 'communitarian'
interpretation) underpins rule-following is agreement (or better,
coincidence) in pre-linguistic responses, not in judgment
proper. And that is how it should be, if the claim is that language
could not have developed without coincidence of responses.
Of course, a possible position would be that some social
communality of pre-linguistic responses was required for any
individual to have a language. But that would be an unsupported
claim, and only if some content could be given to the alleged
communality could it be evaluated. Once such content was given
(e.g. along the lines of mutual perception of responses) I could
seek to model it in the persisting individual. There is a further
problem in the notion of genuinely social sharing that is
supposed to take place prior to language: the mutuality or
communality involved would have to be possible for creatures
without language.
So the claim that needs scrutiny is this: the only intelligible
kind of checking of responses is inter-personal; even given
time, I cannot compare my own responses one with another.
And this claim hardly needs to be refuted. To compare any two
responses, the least that is needed is the awareness of the
responses, the awareness of that to which they are responses, and
the awareness that they are both responses to that stimulus.
(This is intended to be a minimal characterization. Clearly no
less will do. I do not find it plausible myself that all these

192

I-HARRY

A. LEWIS

attitudes could inhabit a mind as yet incapable of language.)

The responses in question being construed as behavioural, I see
no difficulty in allowing the responder himself to be aware of
them. He can be simultaneously aware of them, either by
memory of one or both, or by their persistence (if they are traces,
such as marks on the bark of a tree). Simultaneous perception of
responses in two people is not obviously easier in practice, and
surely not in principle. Awareness of the cause is entirely
analogous. The crucial step is the next. How could an individual
come to know, of his own responses to some event, that they are
caused by it? As long as we recognize in the 'response' an item of
observable behaviour, I cannot see that such causal knowledge
is any more difficult for the individual of himself than for any
person of anyone, whether himself or another. Such knowledge
is no more difficult to acquire than the knowledge that contact
with a particular kind of plant causes a rash.
Thus, if comparison amounted to no more than this, the
idiolectophone could indulge in it. On the other hand, the
account does not yet appear to give us any notion of
decision-for example, of choice between two different responses
to the same stimulus. The short answer to this objection is that
any description adequate to allow for such choice will ascribe a
language to the chooser, and so rule out the process of
comparison from counting as a precondition of language. But a
better reaction is to point out that even this inadequate account
already undermines the claim that all objectivity requires
agreement in response as a precondition. For any perception of
agreement in responses requires in the perceiver a capacity to
recognize the same responses again. The correctness of that
recognition is required for the process to start at all, and so
cannot be explained by it.
Can we help the argument to community by lifting the two
artificial constraints that we have imposed, namely, that the
judgment of the member of a community was to be taken at an
instant, and that the relevant conception of agreement in
responses was to be pre-linguistic? Let us lift them in turn and in
the orderjust given. We can see that inter-personal comparison
(as distinct from mere coincidence) of judgment needs time
anyway. But if we hold firm to the contrast between comparison
and correction, we can see that nothing can now undermine the

CONTENT AND COMMUNITY

193

ability of S (our solitary) to make comparisons, once it has been

granted. On the other hand, if comparison requires a fullyfledged language, we are still withholding that. Moving from
consideration of comparison to that of correction, what we find is
that the cost of conceding to the social being the extra reasons for
preferring one response over another that become available
once (e.g.) memory is allowed its due role, is the creation of new
opportunities for scepticism. If the agreement in judgment has
to be discovered, it can only be discovered by one who already
knows what we must then allow the solitary to know as
well.
If we lift the second constraint, we shall allow ourselves to
ascribe a genuine language to S, and seek to find in this a covert
reference to a community, to be revealed by scrutinizing the
process of checking. Users of natural languages do 'correct' their
own usages without social prompting or comparison in each
case; we are all (in that respect) hermits for much of the time. We
may detect inconsistencies in our own writings or in our
remembered thoughts; we may come to see as false recorded or
remembered reports of perceived objects or events. In both cases
a retention of meaning through time is required: but it is as
much an assumption of social as of individual comparison that
such retention occurs in individuals. (How else could the radical
translator or interpreter build up an account of his informant's
language?) Once more we do well to keep apart the notions of
comparison and of coming to know one was wrong. 'Social
truth' can be seen to provide no more than a point of comparison
not available to the solitary, unless it is held apriori to guarantee
objective truth. I ought to note here a claim of Davidson's that
would, if proven, settle the question in favour of the community:
'Someone cannot have a belief unless he understands the
possibility of being mistaken, and this requires the contrast
between truth and error-true belief and false belief. But this
contrast, I have argued, can emerge only in the context of
interpretation which alone forces on us the idea of an objective,
public truth."' I do not pretend to understand completely how
this very compressed argument works. But for present purposes
the crucial sentence is the last, with its claim that the sole
9Davidson, 'Thought and Talk', op. cit. p. 170.

194

I-HARRY

A. LEWIS

possible source of a notion of incorrectness is 'the context of

interpretation'. We may accept that the individual who
compares and modifies his responses has the concept of
incorrectness (even if his decisions are sometimes wrong). We
may further accept that someone who indulges in radical
interpretation needs that concept. However I cannot see that
any argument is offered by Davidson to show that the concept
could arise in an individual in no other way. Indeed, it would
seem to be an intellectual rather than an apriori difficulty: could
not a very clever and inventive isolate come up with the
concept?
V
The role of 'community' in the arguments considered gives little
support to the conclusion that the oft-repeated claim, 'language
is a social art', is anything more than a general empirical truth.
The most that the arguments establish is a series of contrasts: in
Burge's case, a contrast between the possessor of an attitude with
her language and the language in which an attribution of that
attitude is made; in the discussion of radical interpretation, a
contrast of two roles, interpreter and informant, and two
languages; in Wittgenstein's discussion of rules, a contrast
between individual and communal judgement. As we have seen,
analogies exist between the relation of individual to community
and the relation of one community to a larger of which it is part;
and between the relation of two non-intersecting communities
and the relation of two individuals. At the limits, we can place
the hypotheses of the isolated user of an idiolect; and of an error
by all mankind. It needs no new argument to defend the
intelligibility of the latter. The former seems to me equally to
survive renewed attacks. In both cases, it is possible to
understand why we cannot produce examples. No one can
produce an example of a current belief of his own that is false.
Likewise I cannot, unfortunately, introduce you to a lifelong
social isolate, or invite you to interview an idiolectophone.
In between these extremes we should recognize, not a sharp
division between individual and community, but roles and
relationships, vertical and horizontal, that may hold now
between individuals, now between groups, now even within a
single person. Wittgenstein, when he was, according to some, in

CONTENT AND COMMUNITY

195

the midst of arguing that any rule-use required a community,

remarked:
A human being can encourage himself, give himself orders,
obey, blame and punish himself; he can ask himself a
question and answer it. We could even imagine human
beings who spoke only in monologue; who accompanied
their activities by talking to themselves.-An explorer who
watched them and listened to their talk might succeed in
translating their language into ours.20
The multiple possibilities of language creation and languageuse argue the need for care over terminology. A good example is
the use of the word 'public', as in 'public language'. Does this
mean, 'language with perceptible sentence-tokens', 'language
learnable by third parties', or 'communal language'? Part of my
own position is that a language with perceptible (and enduring)
sentence tokens need not be communal; careless talk of 'publicity'
could render this view unintelligible. (Perceptible 'clothing' of
thoughts is a necessary, but not a sufficient, condition for
community of language.) A related ambiguity in the expression
'grasping the meaning' has been remarked upon by Gilbert.2'
Our critical discussion suggests an opposite conclusion to that
being argued for by the supporters of community; far from its
being a prerequisite of language, language is a prerequisite of
community, and no mass conspiracy to set it up, no wordless
social contract, needs to be postulated.
A key notion to emerge is that of the reliability of the
individual language-user even when a member of a community.
It is clear that agreement alone could not guarantee objectivity;
at best, it could be only one among many factors. If the idea of
the individual as an autonomous 'responder' or 'measuringinstrument', in Mark Wilson's phrase,22 is introduced, a
complementary view to the one attributed to Wittgenstein
becomes attractive. Agreement in judgment shows us the truth
because only the assumption that the world is as it is thus
represented can explain the fact that we react in a similar way.
There are three contributors to this process: a world, open to
Investigations, op. cit. section 243.
Philosophical
M., 'Has Language a Social Nature?', Synthese56 (1983) pp. 301-18.
22 Wilson, M., 'Predicate Meets Property', Philosophical
Review91 (1982) p. 255.
20

Wittgenstein,

21 Gilbert,

196

I-HARRY

A. LEWIS

perception; a reliable human perceiver; and a language. In my

view the role of the community is secondary. It can, but need
not, provide a language. It does provide alternative human
perceivers, so that what we collectively agree on, the intersection of our several belief-sets, is likelier to represent the world
than any one individual's total belief-set. It is because people are
likelier to be reliable about what they perceive than about what
they merely imagine that inter-personal agreement is a test of
objectivity. We consign that about which we disagree to the
zone of the 'subjective' not primarily because it is uncomfortable,
but because disagreement is evidence that it is not 'out there' at
all. There is an analogy here between two quite different areas of
divergence of belief: morals, and advanced science. For what is
least disputedly 'objective' is that which all of us can agree about
with minimum training. Such agreement, however, is the
evidence, not the essence.23

2" In preparing this paper I have benefited from discussions with Mark Johnston,
Philip Pettit, Pamela Tate and Roger White.

CONTENT

AND COMMUNITY

Harry A. Lewis and Andrew Woodfield

II-Andrew

Woodfield
I

Two Kindsof Meaning

I agree with the individualist thrust of Harry Lewis's paper as
far as thoughts and concepts are concerned. The conclusion of
my paper will echo Lewis's view, but I hope to display the issues
from a slightly different angle. Before discussing thought, the
argument will touch on the notion of linguistic meaning. Here I
think I do disagree with some of the things that Lewis says.
When he talks about content in connection with linguistic items,
I find that he is too individualist. With any luck, there will be
enough disagreement over this to provoke a pleasant debate.
In section II, I maintain, despite my individualism, that there
are also such things as communal thoughts, which are emergent
from, or supervenient upon, community activities. I make use of
Grice's distinction between sentence or word meaning and
speaker's meaning.' Words and sentences possess standard
meanings which remain invariant over the many uses to which
particular utterances may be put,2 whereas what a speaker
means on a particular occasion is up to the speaker.
I rely upon this intuitive contrast, because I want to hold up
standard linguistic meanings as my models when I introduce the
notions of communalthoughtand communalconcept.The senses in
which a thought or a concept might be said to be essentially
communal become clearer if we consider the matter in the light
of this contrast.
In fact, it is possible to construct several different notions of
communal thought. There is nothing ontologically wrong with
thoughts that are communal. Controversy comes in over the
questions whether they are the only kind of thoughts, and
whether they are more basic than individualist thoughts. I do
'Grice, H. P. (1968) 'Utterer's Meaning, Sentence Meaning and Word Meaning',
Foundations of Language 4, 1-18.
2 Davidson, D. (1975) 'Thought and Talk', p. 164, Inquiriesinto Truth and Interpretation,

Clarendon, Oxford 1984.

198

II-ANDREW WOODFIELD

not think they are the only kind, nor do I think they are basic, as
section III will reveal.
Let us start by considering actual human languages, such as
Arabic, English, Mandarin or any of the several thousand or so
others that are in daily use. Some, like Cornish, are dying out. It
is said that only two persons are left who can speak fluent
Cornish. I find it helpful to say that the basic purpose of a
language is to facilitate communication. The truth of this
functional claim depends, I think, upon facts about how the
language originated. A genuine language has to have evolved
because it has successfully mediated communication in the past,
or if it is an artificial language, it has to have been designed for
that purpose. The concept of a natural language is partly
explicated by the platitude that natural languages have evolved
for sending messages. The various components of a language are
there for the same purpose, but derivatively, as the organs of a
body are there to keep the whole animal alive, or as the parts of a
machine are tailored to its overall purpose. The thesis holds, as
Aristotle says, 'generally and for the most part'. It is not falsified
by putative counter-examples of the 'Robinson Crusoe' kind in
which a language is not used for communication. All sorts of
exceptional situations are imaginable. Platitudes about functions
can admit of exceptions and yet reveal essential truths.
Lewis is right that individuals who talk to one another do not
ipsofacto bind themselves into a community. It happens that
people who converse a lot tend to live close together, consequently they tend to enter into social relations and adopt shared
norms. Using the same tool is not sharing a norm.3
Yet, if one considers any living natural language, with all its
richness and tradition, one knows that there must have been a
3A related point is made by Bach and Harnish, when they note that successful
communication in L requireswhat they call the 'Linguistic Presumption':'The mutual
belief in the linguistic community C that
(i) the members of C share L, and
(ii) that whenever any member S utters any e (expression)in L to any other member
H, H can identify what S is saying, given that H knows the meaning(s) of e in L and is
aware of the appropriate background information' (Kent Bach and Robert M. Harnish
and SpeechActs, p. 7. MIT Press, Cambridge, Mass.
(1979) LinguisticCommunication
Without such a presumption, neither S nor H would be entitled to assume that e
means to the other what it means to himself. So the mutual belief is something else that is
shared by people who converse in L. But of course this is not sufficient to forge a social
community.

CONTENT AND COMMUNITY

199

society that sustainedits development.I am inclinedto think

that a word'shavinga meaningin L is an emergentproperty,
not completely analysablein terms of what the totality of
speakersof L have meant when they have used the word.
So insteadof tryingto define'socialcommunity',let uslookat
how it couldpossiblyhavecomeaboutthatwordsandsentences
have standardmeanings.
SupposethatwordW hasa meaningM in L now.Itsmeaning
is whatit is now becausetheactivitiesofmanypeoplein thepast
establishedit as such,and also becauseW is (orwouldbe) used
with that meaningby a sufElciency
of the presentpopulation.It
is not a coincidencethat the presentusersaccordit the same
meaningas the earlierpeople.The currentlot do this because
they know that M is the establishedmeaning.Generally,they
speakin accordancewitha ruleto theeffect:'Respectpreviously
establishedmeanings;don't be a Humpty-Dumpty'.They do
this becauseit facilitatesthe efElcientexchangingof messages.
Roughly,then, W has a standardmeaningM in L because,
and in so far as, people treat it as having that meaning.This
statementseemscircular,like saying that a coin is a 1 coin
because,and in so far as, it is treatedby peopleas a 1 coin. In
bothcases,the shortstatementtriesto sumup a longstory.The
story is not viciouslycircular,it describesa historicalprocess
that contains cyclical subprocesses.It is a kind of historical
explanation rather than a meaning-analysis.Similarly, to
explainwhya certaintypeof coin is worthonepoundSterlingis
not to ofTera semanticanalysisof 'Thiscoin is worthone pound
Sterling'.The conceptof the poundSterlingbelongstoa higher
levelof discourseand is linkedin withotherconceptsto do with
finance and economics.The conceptsexercisedin the backgroundstorydo not belong to this finishedsurfacescheme.
Thereare severalhallmarkswhich serveto fix the notionof
standardmeaningmoreclearly.I shallmentionthree.First,the
meaningthat a given wordhas in L is a matterthattranscends
what any individualmay say or believe,and the answerto it is
usuallyobjectivelydecidable,once the word has enteredthe
lexicon.A newcomerto the languagehasto discovertheanswer
by empiricalinquiry.Evena native-speaker
can be wrongabout
the meaning of a word in L. It is even possibleto imagine
circumstancesin which a majorityof the speakersof L have

200

II-ANDREW

WOODFIELD

mistaken beliefs about the meaning of W. But this would be an

unusual situation. Generally, if most people think that W means
M, this is a good sign that it does; there is a continuous gentle
pressure maintaining things thus. When a new word ('wally',
'frisbee', 'software') is coined, it does not have a standard
meaning; the process of standardisation goes hand in hand with
the assimilation of the word into the lexicon. One person can be
responsible for starting this process, but the word has to catch on
among a significant subset of the population before it can be said
to be a part of the language. The initiator is not in control of the
word once it has entered the public domain. Its eventual
standard meaning could be different from the meaning that the
inventor originally assigned.
The second hallmark is stability. Once W has acquired a
meaning, it will keep it, no matter how often or how rarely the
word is uttered, and regardless of the purposes for which it is
uttered (cf. Davidson's 'autonomy of meaning', op. cit. note 2).
Once in the communal pool, a meaning stays there indefinitely,
and the fact will be registered by all competent dictionaries.
Some words lose currency and are perceived as obsolete. But
they can sometimes be revived. Given the current craze for oldfashioned cuisine, rare words like 'posset' and 'medlar' could
become popular again. They continue to bear the original
meanings that they had in Shakespeare's day.
Certain words ('nice' is an example) have an old meaning
which is hardly known, plus a new meaning. Such words are
ambiguous. Both meanings are standard, in my sense of the
term, even if one is better known the other.4
I do not say it is impossible for a word to change its standard
meaning(s), or to lose all meaning. Clearly, keeping a meaning
depends in some loose way upon a continuity of practice among
temporally spread-out L-speakers. But it is difficult to define a
precise requirement. If a word were forgotten by everybody
throughout the twentieth century, then remembered in the
twenty-first, it could be said, at the later time, that the word
meant so-and-so during the previous century but no one living
then was aware of the fact. Languages, like libraries and other
4 Dictionaries sometimes list a 'standard' (i.e. a dominant, received) meaning plus a
'non-standard' (subsidiary, slang, regional) meaning. On my use of the term, both
dictionary meanings would be standard, that is, communal and stable.

CONTENT AND COMMUNITY

201

repositoriesof traditions, grow bigger all the time, or get wiped

out, but they rarely shrink.In George Orwell's1984, 'Newspeak'
is deliberately impoverished English; but the case is fictional.
The third characteristic is the presence of social procedures
for determining correct meanings. One would expect to find
such procedures wherever a community prizes its native
language and is concerned that it should function well.
Occasionally, an official body is set up, like the Royal
Commission on Abortion, to adjudicate the meanings of legally
important key terms ('foetus', 'viable', 'alive'). When the
problematic term is technical or scientific, or is a natural kind
term,5scientific experts will be accorded a big say. If the word is
not technical, savants like etymologists, dialect-specialists, and
philosophers are asked for their views. Eventually the public
'decides' by adopting a certain meaning as standard. It is hard
for an authority to compel the public to use a word in a certain
way. The Acad6mie Frangaise failed in its efforts to ban 'le
weekend' and 'le parking' from French.
In cases of urgency which attract official attention, one can
imagine that a single individual might have the power to
legislate a meaning. But this could occur only if the personwere
acting in an official capacity, on behalf of the whole community.
If a nonentity tried to issue an edict, it would carry no weight.
No doubt there are other criteria of standardisation.We shall
let these three suffice for our purposes.
Lewis spoke of idiolects. By 'S's idiolect', I mean the version of
a natural language which is spoken by S. One is free to christen
the version by its own name, as if it were a different language
altogether. For instance, it might be said that Ronald Reagan
speaks Reaganish. But this is clearly not a private language (not
for nothing is Reagan known among his staff as 'The Great
Communicator'), and there could be other people who spoke in
exactly the same idiosyncraticway as Reagan. Reaganish would
then be their dialect, which would not stop it being Reagan's
idiolect. Although it is nominally essential that S speaks his
own idiolect, it is never an essential property of S's idiolect that it be an idiolect. The words in it will have standard
meanings; S had to learn it, he can have false beliefs about
5Putnam, H. (1975) speaks of a 'division of linguistic labour' in 'The Meaning of
Papers,vol.II: Mind,LanguageandReality,C.U.P., Cambridge.
"Meaning" ', Philosophical

202

II-ANDREW

WOODFIELD

meanings in it, and he is not responsible for those meanings.6

We now turn quickly to speaker'smeaning. I do not wish to
examine Grice's notion in depth.' I simply wish to bring out the
fact that the notion is individualistic.
S ironically says 'You cooked a fine stew', meaning that the
person he is addressingcooked an unpleasant hotpot. The fact
that S meant this is independent of whether anyone understood
him; it is also indifferent to whether S is a member of a
community; also the sentence he uttered does not mean what he
meant. What S meant depends solely on S's state of mind
at the time. S, who believes there is someone within
earshot, intends to make that person go through an inference
about S's reason for speaking. The hoped-for inference is to
include a number of 'mutual contextual beliefs'. Ironical
remarks demand more decoding than straightforwardsincere
utterances do. John Searle' has applied the same idea to explain
how people understand metaphors. The interpreter has to see
that the words uttered are not to be taken literally, but can be
used as clues for puzzling out the speaker'smeaning. The same
idea works for many other figures of speech, such as hyperbole,
euphemism, and malapropism.
While the example well illustrates the contrast between two
levels of meaning, it is noteworthy that the hearer has to know
the sentence-meaning first in order to retrieve the speaker's
meaning. Indeed, in sophisticatedspeech-acts,S can be counted
on to exploit H's previous knowledge of L.
It might be urged, therefore, that the example is no good for
proving that speaker'smeaning is the more basic kind, and also
that it casts doubt on the basis that speaker'smeaning is purely
individualistic.The scenariopresupposesthat there are sentencemeanings available, which, so I have argued, are communal.
Well, I am not in fact arguing that speaker'smeaning is basic.
But even so, the objection fails to establish the contrarythesis, on
two counts. The notion of speaker'smeaning, pure and simple,
does not entail prior knowledge of any language. Some of
Grice's (1957) examples are of 'primitive' utterances lacking
SThis point is made by Michael Dummett in 'The Social Character of Meaning'
(1974) p. 425, in Truthand OtherEnigmas,Duckworth, London 1978.
7 See Bach and Harnish (op. cit.) pp. 149-154 for some criticisms of it.
8 Searle, J. R. (1979) 'Metaphor', in A. Ortony (ed) Metaphorand Thought,C.U.P.,
Cambridge.

CONTENT AND COMMUNITY

203

prior linguistic significance.' Secondly, even in real life where S

and H share an environment, a language and a community,
these contextual facts are not truth-conditions of 'S meant that
someone had cooked an unpleasant hotpot'. What is necessaryis
that S should suppose that these conditions obtain.
It is clear that we are dealing with a duality of kinds of
meaning, not just a distinction between types and tokens. S can
on one occasion utter a token of sentence-type z and mean that
P, then on another occasion S can utter a second token of
sentence z and mean that not-P. But the two sentence-tokens
have the same standard meaning on both occasions. They
inherit it from the sentence-type of which they are tokens. This
reveals an important link between the two contrasts. The
fundamental possessorsof standard meanings are types (words
or sentences). With speaker'smeaning the polarity is reversed;
particular utterances are their vehicles. To determine which
message an utterance carries you must look to the psychological
circumstances of production. And this is precisely the reason
why the notion is individualist.
II
SomeNon-Individualist
Conceptions
of Thoughtsand Concepts
Taking the concept of standard sentence meaning as our model,
we may now introduce the concept of a standard thought in
community C. We shall henceforth speak, for example, of 'the
standard thought that P', and also of 'the standard K concept'.
For brevity's sake we may omit the reference to community C.
These notions are hardly new. The standardthought that P is
just what many philosophers have always meant by 'the
proposition that P'. The link with standard meaning brings this
out very clearly, since many writers define a proposition as
something expressed by a declarative sentence (in virtue of its
standard meaning).
Questions need to be raised about the ontological status of
standard thoughts. Are they creations arising out of communal
activity? Are they supervenient upon social activities and
relations? If the former, are they artefacts, or are they natural
products? If the latter, are they epiphenomena or are they
causally efficacious?
'Grice, H. P. (1957) 'Meaning', PhilosophicalReview66, 377-88.

204

II-ANDREW

WOODFIELD

Not all philosopherswho espouse standardthoughts have tied

them closely to sentences. Frege, despite saying 'thoughts are
senses of sentences',' weakens the connection by allowing
thoughts that are never expressed. I have been unable to
discover whether Frege ever said that thoughts can be
independent of language altogether. He does state, however,
that a thought exists prior to anyone's grasping it."
Frege's precursor Bolzano, whom a German commentator
recently described as a 'logical Plato','2 held similar views. He
wrote, '.

. by proposition in itself I mean any assertion that

something is or is not the case, regardlesswhether or not somebody has put it into words, and regardlesseven whether or not it
has been thought.''"He also affirmsthat 'There is indeedno doubt
that thinkabilityis a propertyof any proposition,but it is also obvious that it does not form part of the concept of a proposition'."4
Again, the connection with language is not completely
severed by this formulation, since a propositionmight exist only
in so far as it couldbe put into words. Moreover, Bolzano is not
committed to propositions as Platonic entities, for he says 'One
must not ascribe being, existence or reality to propositions in
themselves'.' The Fregean tradition flourishes today. Popper
advocates the objectivity of 'World 3', whose denizens are
autonomous of any mind.'" Christopher Peacocke consciously
adopts Fregean terminology in his recent book.'7
Frege's logicist motivation is well-known. Both he and
Bolzano deplored the post-Kantian tendency to 'psychologise'
logic, which in their view propagated doubts about its objective
validity. If thoughts can be shown to be objective, then
presumablylogical relationsbetween them can also be objective.
0oFrege,G. (1918 originally) 'Thoughts', p. 4, in LogicalInvestigations,
(ed) P. T.
Geach, Blackwell, Oxford 1977.
"
Frege (op. cit.), p. 18 footnote: 'A person sees a thing, has an idea, graspsor thinksa
thought. When he graspsof thinksa thought he does not create it but only comes to stand
in a relation to what already existed-a differentrelation from seeing a thing or having
an idea.'
"2See the editor's introduction (p. xxix) to BernardBolzano (originally 1837) Theory
of
Science,edited and translated by Rolf George, Blackwell, Oxford 1972.
'"Bolzano, (op. cit.) pp. 20-1.
'4ibid. p. 26.
'5ibid. p. 21.
16 Popper, K. (1972) Objective
Knowledge,esp. chapters 3 & 4. Clarendon, Oxford.
17Peacocke, C. (1983) Senseand ContentClarendon, Oxford (see esp. p. 106).

CONTENT AND COMMUNITY

205

Still, it is tempting to wonder if Platonism is really necessary,

if one's main concern is to ensure the shareability of thoughts.
All one really needs (so it might be urged) is the distinction
between thought-particularsand thought-universals.The latter
are by their very nature shareable, that is, possessableby many
subjects. The same goes for ideas, pains, indeed any mental
entity for which the particular/universaldistinctionmakes sense.
But Frege did not say this, because he would not have been
satisfied that it ensured the supra-psychologicalstatus of logic.
The Aristotelian conception of a 'thought-kind'entails (a) that a
kind exists only if at least one particular of that kind exists, and
(b) that relations obtain only among the kinds that exist. This
limits the scope of logic to the domain of what has actually been
thought.
It is very important for Frege, then, that the thought-type
comes first.This guaranteesthe anti-individualismand the antipsychologism. But there is no need for him to view thoughts as if
they were Platonic Forms, disjoined altogether from human
activities. For there remains the other possible position, that
thoughts are cultural products, emergent out of the intellectual
and linguistic life of human collectives.
At this point we need to say more about what a constitutively
communal thought might be. It seems to me that there are
several ways to go here, and it is up to each theorist to construct
the concept he wishes to use. But certain general specifications
can be laid down. On any view, a standard thought is going to
be something like a cognitive tradition or institution. We
already employ such notions as, for example, 'the Spanish
concept of masculine pride', 'the Britishnotion of fair-play', etc.
However, we do not wish to limit ourselves to ideas that are
cherished by the culture in which they arise. We want to include
concepts that have never yet been explicitly recognised or
formulated, but which are in some senseavailable in the cultural
repertoire. Every participant in the culture, including infants
who are born into it, has all the standard concepts of C
available, waiting to be grasped. This legacy is to be counted
among the advantages of civilised living in C.
The model of standard meaning serves us well, since we
already know what it means to say that a possible sentence of L,
never yet uttered, already has a meaning in L. This is because

206

II-ANDREW

WOODFIELD

sentencesare composed of words which already have meanings;

a similarcompositionalitycarriesover to the domain of thoughts
and concepts.
Standard thoughts are composed of standardconcepts, then,
but it is left as an open possibilitythat some standardconcepts in
C are not easily expressible in L. There may be room for such
things as implicit standard concepts, which explain features of
the life that is led in C, styles, formsof pictorial representation,
and so on. One can imagine that the Spanish language might
never have possessedthe word 'machismo', but that Spanishlife
might still have included the same range of characteristicmale
responses to certain life-situations. There could be a point in
saying that Spanish life contained dramatic representationsof
male pride which were not explicitly recognisedas such. There
is no doubt that literary critics, historians of ideas, anthropologists, and sociologists do find it useful to operate with such
theoretical notions, which are analogous to notions used in
individual psychology. Indeed a persuasivecase could be made
that the English-speakingworld implicitly possessedthe concept
of machismo, even before the time, not long ago, when it
borrowed the word from Spanish.
Do standard thoughts lodge in the group mind? There is no
harm in using the metaphor, but no need to commit oneself to
any group mind. The notion of standard thoughts can be
explicated reasonably well without it, on analogy with the
concept of standardmeaning.18Analogues of the three marksof
standardness are easily discernible. The fact that stable
cognitive traditions exist is empirically verifiable. If it were not
so, there would be no such disciplines as sociology or history of
ideas. Also the participants in C can be collectively responsible
for determining the identity of a standard concept or thought.
Questions like the following can be raised: which thought does
sentence z express? Which standard thought was speaker S
expressing?There will be cases where what has to be decided is
the rightness or wrongness of certain ascriptions. It will be an
unfailing characteristicof such cases that the individual thinker
is not authoritative about the content of the thought he was
entertaining. Other people, interlocutors or outsiders, can
8 Cf. M. Dummett, (op. cit.) pp. 427-8, where he speaksof 'knowledge possessedby
the community as a whole'.

CONTENT AND COMMUNITY

207

correct S about its identity. Tyler Burge, in his ingenious antiindividualist argument, sets up clearly-defined situations in
which precisely this kind of phenomenon occurs. The Cartesian
intimacy between subject and thought-content gets sundered
when the thought is communalised. S can aim at a certain
thought and miss it, yet be under the impression he has hit it.
Thus we find Burge saying that his argument gets under way 'in
any case where it is intuitively possible to attribute a mental
state or event whose content involves a notion that the subject
incompletely understands."'9 Burge treats standard concepts as
symbolic instruments for people to 'think with'. Operating with
thoughts is seen as a cognitive skill, at which some people are
more adept than others. Burge sees the community as 'defining'
the notionof sofa; my third criterion is clearly applicable to his
conception of a notion.
I have said there are many alternative methods of constructing
conceptions of communal thought. To see some of these, it is
helpful to focus on the different ways one might try to explicate
the 'grasping' relation. I shall sketch just four possible methods,
two of which seem reasonably plausible.
(1) The 'Dasein' Model. For a (suitably receptive) S to think a
standard thought it is sufficient for S to be appropriately
ensconced in the relevant cultural environment. Nothing hinges
on S's inner condition. Two individuals could be internally
identical, one of whom thinks that P, the other does not, solely
because the former stands in the external relation, the latter does
not. This Hegelian or Heideggerian position entails the
wholesale rejection of the Cartesian concept of mind. The model
not only conflates 'having a thought' with 'having a thought
available', it is also wholly at odds with what is known about the
dependence of cognition upon brain processes.
(2) The 'Internalised Token' Model. S thinks standard
thought T iff there is a token of T, (call it t), inside S, and S
stands in the 'thinks*' relation to t.20The crucial feature of this
model is that the inner token t is itself a standardised entity. The
9 Burge T. (1979) 'Individualism and the Mental' p. 79, MidwestStudiesin Philosophy
vol. IV (eds) P. A. French, T. E Uehling and H. K. Wettstein, University of Minnesota,
Minneapolis.
"
For more on 'thinks*'see Hartry Field (1978) 'Mental Representation',Erkenntnis
13,
9-61.

208

II-ANDREW

WOODFIELD

token is not identical with the neural state, if any, which

'realizes' it. Token-thoughts are like pound-notes; they
supervene on the physical stuff, but their identity-conditions
distinguish them from anything physical. Each token is of a
certain type essentially; the type, in its turn, is essentially
standard; so the token is too.
The thought-token is analogous to a speech-act token in this
respect. S says 'Je demande que la porte soit fermne', thereby
performing the act of asking in French that the door be shut.
What S does is constitutively social under that description. The
sounds produced would not warrant that description unless the
French language existed, which in turn allows one to infer, given
that French is a natural language, that French society sustained
its evolution.
Many writers espouse this position, including Hegel, Marx,
L. S. Vygotsky,"2 JulianJaynes.22 Many in this camp like to say
that the individual thinker 'internalises' or 'appropriates' the cultural objects. When individuals make contact with a standard
thought, they create a copy of it inside their heads, rather as a
magnetic tape records a message. A great deal of social and technological stage-setting is required before this can happen.
(3) The 'Cultural Osmosis' Model. When S thinks a standard
thought, S enters into a complex relation with something public,
but S can do this only by satisfying a number of internal and
external conditions. They would probably include at least the
following: S's brain is currently in a certain type of physical
state; S has gone through certain kinds of appropriate training
(has learned a language, learned the norms of a group, acquired
certain skills); S is in social contact with others and counts as a
member of a community. Also certain linguistic and social rules
must be in force in C; standard thoughts supervene on the
shared practices. To gain the right kind of access to them, S must
be brought into the group and must agree with others in
judgments and behaviour. S must be disposed to react in certain
ways spontaneously, without reflecting. This position is inspired
by Wittgenstein.
Certainly, where hybrid psychological states are concerned,
21Vygotsky, L. S. (1962) Thoughtand LanguageMIT Press, Cambridge, Mass.
in the Breakdown
J. (1976) The Originof Consciousness
of theBicameralMind,
Boston, Houghton Mifflin.
22Jaynes,

CONTENT AND COMMUNITY

209

contextual conditions, as well as purely psychologicalconditions,

affect the truth of ascriptions. But the osmotic model takes
externalism further than the mere recognition of hybrids; it
denies the validity of the inner/outer distinction which is
employed by those who talk of 'purely psychological states'.
Individual minds themselves supervene on group life, rather as
pieces of metal are brought to life by being used as money.
(4) The 'Measuring Stick' Model. This view, which I hold,
says that when S grasps a standard thought, S has an individual
thought which is complete in itself and has its own content. The
individual thought stands in a certain relation to the standard
thought. It has a similar role in S to the role that the standard
thought has in the systemof standard thoughts. Other people, in
ascribing thoughts to S, identify S's individual thought by
specifying the standard thought which most closely resembles
S's individual thought, availing themselves of the imperfect
resourcesof their own language. The background to this view is
a theoryof content-ascriptionswhich I have outlined elsewhere.23
It is an individualist view, yet it does not deny that standard
thoughts exist. It treats them on a par with standard meanings.
Nor does it deny that individuals can enter into cognitive contact with them. But the contact is indirect, being mediated by
individualthoughts.It is possiblefor S to be awarethat he or she is
thinking the content ofa standard thought that P. This would be
the case if S thought that P, and knew that this content was
specifiable through the sentence 'P'. In other words, S has an individualthought plusa backgroundof semanticknowledge.I shall
say more in defence of the fourth model in the next section.
III
IndividualistThoughtsAre MoreBasic
Are there such things as irreducibly individualist thoughts, in
addition to standard thoughts? In this final section I defend an
affirmative answer.
No sane believer in standard thoughts will wish to deny
outright that people perform individual acts of thinking. But he
may well try to offer a reductive analysis, saying that an
individual thought is nothing but a person entering into a
23Woodfield, A. (1982a) 'On Specifying the Contents of Thoughts' in Thoughtand
Object(ed) A. Woodfield, Clarendon, Oxford.

210

II-ANDREWWOODFIELD

relation with a standard thought. He then uses one of the models

just sketched of the nature of this relation. I call this position
'reductive', but it goes in the opposite direction from a
microreduction. It analyses the mental partly in terms of the
sociological.
At first sight, the position seems paradoxical. To claim that
individual acts of thinking are always acts of grasping communal
contents undermines the initial contrast, which my exposition
relied on, between speaker's meaning and standard meaning.
For speaker's meaning is a matter of speaker's thoughts and
intentions. The proposal is, in effect, to relocate speaker's
meaning on the communal side of the divide.
The claim must be that individualist thoughts (and speaker's
meanings) are social in a different and possibly deeper sense
than that which was invoked when the original contrast was
drawn. The deeper sense of 'social' must turn on a theoretical
account of what it is for a person or a sentence to mean, with the
theory showing that all meaning involves something social in the
background. And it might be argued that the 'paradox' is more
apparent than real. It seems like a paradox only because of my
mode of exposition of the contrast.
Nevertheless, even if speaker's meaning did ultimately
involve a social relation, I should still maintain that there are
thoughts and (non-semantic) intentions which are irreducibly
individualist. Ultimately the case for individualism does not rest
essentially on the theory of meaning, but upon positive
independent considerations, some of which I shall shortly
provide.
To be individualist, a thought has to meet the following
conditions. These are obverses of the conditions for communality, not only because they say opposite things, but also
because they apply to particulars first, kinds second. Clearly,
individualist thought-particulars, if they exist, are not going to
be ontologically derived from Platonic thought-types. Individualist thought-kinds are Aristotelian: they exist through
having members.
First, each individualist thought depends upon the cognitive
activity ofjust one person, and belongs to that person. But two or
more people can have individualist thoughts that are similar.
These may be grouped together, and it may be natural for a

CONTENT AND COMMUNITY

211

third-personto characterise the various thoughts using the same

'that'-clause each time.
Second, the subject of an individualist thought is responsible
for keeping it alive or letting it disappear. With long-lasting
thought-contents, such as the contents of (core) memories and
(core) beliefs, S keeps them in existence by storing them and not
forgetting them. S's stock of retained thoughts is a distinctive
possession;the more thoughts there are in stock, the less chance
there is of another person having an exactly similar set.
Third, the subject is the ultimate authority concerning
questions about the identity of his or her own individualist
thoughts (though S's verbal expressions of those thoughts may
be liable to linguistic correction).
There can be no doubt that we do operate with such a notion.
That we have it is proved by the fact that readers of Harry
Lewis's paper, and of this paper, will find our individualist
locutions perfectly natural. They will tend to agree with Lewis
when he says that it is perverse to claim that people severally
'think with' the jointly-ownednotion of sofa, for 'Jane would be
held to "think with" a notion of sofas that excluded armchairs,
in her very judgment that an armchair was a sofa' (Lewis, p.
180). In the earlier paper of mine to which Lewis refers,24I
described the typical 'correction-reaction' of a man who
discovers he has been under a standing misapprehension as to
the meaning of a certain word. He realizes he may have some
beliefs that were picked up through reading or hearing a
sentence containing that word. His interpretation of those
sentences was faulty at the time. In the interests of truth and
consistency, his beliefs must be reviewed, and some beliefs may
have to be modified or jettisoned. The most natural way to
describe this checking-routine,which is familiar to everybody, is
to make full use of the individualist notion of belief. One can
then convey the needed contrast between the standard belief,
expressed by the sentence that S misunderstood, and the belief
which S acquired upon being stimulated by that sentence
(though the latter belief may be difficult or impossibleto capture
in a sentence of L).
The individualistic, 'Cartesian' conception of thought that
24Woodfield, A. (1982b) 'Thought and the Social Community', Inquiry25, 435-50.

212

II-ANDREWWOODFIELD

we use so readily does not countenance thought-contents which

their subjects 'misunderstand', or 'incompletely grasp'. No
wedge can be driven between subjects and their respective
thoughts. Thinking is an act whose content is created by the
subject; contents are not at arm's length, nor are they objects for
S. S neither grasps nor fails to grasp them, nor stands in any
'cognitive relation' whatever to them. To say that S 'thinks with'
concepts can be misleading, unless you constantly bear in mind
that concepts are not objects for S.
Not only do we use these notions, we need to use them.
Individualism is deeply embedded in the common-sense view of
what a mind is. When we try to explain what is going on in
language-learning, and in cases of conceptual innovation, there
really is no alternative to the individualistic account.
Consider the case of a little girl learning to speak. Her
stumbling speech-acts are intelligently controlled, despite her
frequent mistakes. Her parents can usually guess what she
means (speaker's meaning), and they can often furnish intentional explanations of why she made the mistakes that she made.
They see that her responses were rational in the light of the
limited verbal data to which she had been exposed. Such
explanations attribute thoughts, beliefs and intentions to the
child. In the attributions, English words will be used in their
standard senses, but it will not be assumed that the girl has
mastered those senses. On the contrary, the parents will
explicitly say things like 'Susie believes that "cat" means any
small black animal with a tail'.
Now the point I wish to make is simple. The parents are using
a type of explanation which presupposes that Susie thinks
intelligently when she speaks, but she 'thinks with' notions that
are not the standard notions. She 'uses' her own primitive
prototype notions of other things (like black, animal, tail) in the
course of 'building' a concept resembling the standard cat
concept. Her success will depend largely upon her learning the
standard meaning of 'cat'. But the parents need to ascribe
thoughts to her in order to explain how she learns this. So she
must have her own individual thoughts prior to the time when
she can avail herself of the standard thoughts.
What would Burge say about language-learning? Well, he
warns against a typical 'philosopher's response' to his thought

CONTENT AND COMMUNITY

213

experiments, that of reinterpreting the subject's false belief

about the world, such as Jane's belief that an armchair is a sofa,
as a false belief about the meaning of the word 'sofa'.25It is wrong,
he says, to equate having the concept of sofa with knowing the
meaning of the word 'sofa'. A person could have one without the
other. A foreigner might have the concept without knowing the
English word. I agree entirely with this point, of course. And I
wish to emphasise that standard meanings are not the same as
standard concepts. Some standard concepts may be tacit in C;
So people could, in principle, possess the sofa notion in the
absence of any synonym in their language for 'sofa'. But I think a
major motivation for separating 'having a concept' from
'understanding a meaning' is that children obviously possess
concepts before they acquire word-meanings. Their concepts
cannot always be the standard concepts in C, since these require
that socialisation should have already occurred.
I agree with Burge that Susie's thoughts must not be
reinterpreted as having metalinguistic contents. Still, it remains
true that his examples are of people whose grasp is imperfect,
mostly becausethey have not fully mastered certain words. The
obvious rejoinder is: take a person like Susie who is only just
starting to learn language. She certainly has an incomplete
grasp of standard concepts, in fact she has no grasp at all of most
of them. But she thinks. Therefore, her thoughts must be
individualist.
Secondly, we need individualist thoughts to explain how it is
possible for a person to think a new thought that is not
standardly available in C. Individuals are capable of creating
new thoughts and new concepts, just as artists can create works
of art and geniuses can invent new gadgets. Our culture moulds
and trains us, of course, up to a certain level, but much of the
responsibility for innovations lies with individual agents. If it
were not so, the origin of our communal concepts would be
mysterious. There is only one remotely sensible explanation of
how they arose. It relies on an assumption that individuals of
past generations, who stood in social relations, contributed their
ideas to a communal pool. Some ideas caught on and became
standardised, as words do. There was a continuous stream of
25Burge (op. cit.) p. 89 et seq.

214

II-ANDREWWOODFIELD

individual inputs, and the community as a whole appropriated

some of them, thus enabling them to survive across many
generations.
Lastly, I offer an a priori reason for individualism. The thesis
that there are only standard thoughts leads to a regress. It can be
stopped, but only by reintroducing individualist thoughts.
Consider the sentence 'S subscribes to the thought that P', where
this is read as saying that S enters into a cognitive relation with a
standard thought. Is 'subscribing to' an intelligent act, as well as
being an act of entering into a relation? If it is, as it ought to be,
then the story of what goes on in S when S subscribes is bound to
require that S should have some thoughts or views concerning
the thought that P. The communal thought, lying outside S, is
something that not only can be an object for S, but must be, in
so far as S takes up a stance towards it. And then, of course, S's
cogitations about it must be said to have individualist contents.
To avoid any further regress, it must be said that these contents
are not (further) objects of S's consideration, but that they
directly characterise S's mental life.
Tyler Burge says 'The thought-experiment does appear to
depend on the possibility of someone's having a propositional
attitude despite an incomplete mastery of some notion in its
content'." He could also have said it depends on someone's
having incomplete mastery of the whole content. But we could
go further in this direction, loosening the bond between the
subject and the propositional attitude. Why not say that a
person may have an incomplete grasp of the whole propositional
attitude together with its content? Thus, for instance, instead of
saying 'S stands in the "hoping" relation to the content that the
weather is fine', we could say 'S stands in the "having"
("clutching"?) relation to the hope that the weather is fine'.
There is just as much reason to standardise hopes as there is to
standardise the contents of hopes. And I have no objection to
this, provided we don't kid ourselves that we have replaced
individualist mental states by these standard forms. We still
need the individualist ones. The communal entities emerge out
of them, and newcomers to the community gain access to the
communal ones through thinking their own thoughts.
26ibid. p. 83.