212 Semantics and Metaphysics
analysisyntetie dtncion, We cannot quote such cases, but cis
inability may well be the reule of ignorance ofthe sciences. In any
case, the existence if they do exist of statements about which iis
pointless o pres the question wheter they ae analytic or synthetic,
‘hoes not enti the nonexistence of statements which are clearly cls:
Siable in one or othe ofthese ways and of satements ur hestation
lover which ha diferent sources, such asthe possiblity of atemaive
imerpreatons ofthe linguist forms in which they ar expressed.
"This concludes our examination of Quin’ arie- ew be ev
dene that our purpose has been wholly negative. We have aimed t0
show merely that Quine case agains the existence of the analytic
"thetic dstincon isnot made ot, His article has wo par. In one
Of them, the notions of the analytcty group are erie on the
round tht they have nt been adequately explained. In the other, a
Positive theory of tath is outlined purporting to be incompatible
‘with views 10 which believers inthe analtiesyarhetic disunction
{ther must be, or at likely to be, commie. In fact, we have con-
tended, no singe point is established which those who accep the no-
‘dons of the analytic group would fel any stain in accommodating
in their own system of ef, This snot to deny that many ofthe
pins raised ae ofthe fist importance in connection with the probs
lem of giving 2 satisfactory general acount of analytic and related
concept, We are here onl eizig the contention that these pois
justly the rejection, a illusory, ofthe analyc-symthetc distinction
and th notions which belong to the same family.
Consider the following sentences
“Those spots mean (meant) measles.”
Those spy did’ mean anything me, but wo the doctor they
“The recent budget means that we shall have a hard year”
(1 | canoo sa, “Tose pots meat meas, bt he hada ot
teas” and I camo ey, The reset get ea that we Sal
ves hard yea but we shan hae? Tat io say nce the
above, meat that par meas tha peal
(2) camo age om “Thx spt ean fea) meas" to
ny concasion about "what fs) mea By hse mos fe
ela ot ened oy,“ was meu hy hse pos as
iar him,” guano hee tot
seen badgt the conan Whats meant the ec bet
is that we shall have a hard year.”
(3) exon argue om “Toe sos meant mess” ro any cone
chson othe ete tha someon ober mean by tose sp
Sedo, Mats mated he same rote sentence sot he
‘ame buee
(For noneof he bore examples cana restatement be od in
which te ver “mean flowed ys eect sc gua
on ma Ts hoe 0 etl abe em
ated a “Tine sp mean neato Those po ean he
bas measles!” oc ‘
(5) On the other han, for lhe examples an approximate reae Semantics and Metaphysics
statement can be found begining with the phrase “The face that
INP for example, “The fact that he had dose spots mean that he
ihad measles” and “The fact thatthe recent buaget wat a6 i was
‘meats that we shal havea hard year”
"Now contrast he specimen renences with the following
“Those thre rings the bell (ofthe bus) mean that the bus is
‘al
“That sar, ‘Smith could't get on without his wouble and
stif meane tht Smith found hs woe indispensable”
(1) cam ase the int ofthese ad go on to 5a, “Buc it isn't in face
full—the conductor has made a mistake; and I can use the second
ane go, "But in fact Smith deserted he seven yeas ago." That is
to say here x means tha p and mean hat p do not ental
(2) Tan argue from the St to some statement about "what i
(vas) meant” by the rings on the bell and from the second to some
statement about “what is (as) meant” by the quoted remark
(3) can argue from the fst sentence othe conclusion that some
body namely the conductor) meant, of at any rate should have
sean, by the rings char the bus fll, and Lean argue analogously
forthe second sentence
() The fiat ventence can be restated in 2 form in which the verb
“mean” followed by «phrase in quotation mark, that is, "Those
three rings on the bell mean ‘he busi ful So also can the second
(5) Such 4 sentence as *The fact that the ell ha been rang three
‘imes means thatthe busi fll” snot 2 restatement ofthe meaning
ofthe fst sentence, Boh may be true, but they do not have, even
approximately, the same meaning.
‘When the expresions "mean," “means something” "means that”
are wed inthe Kind of way in which they are usd nthe Sst st of|
Seovences, shall peak ofthe Senseo Sense, in which they are sed,
15 the natal sense, oF senses, ofthe expressions in question. When
the expressions are used inthe kind of way in which dey reused in
the second set of sentences, I shal speak of the sens, oF ses, in|
which they are used a the monn Sense, or senses, ofthe expres
‘Sons in queson, I shall ure the abbreviation "mesphy.” 10 istn-
‘ish the nonnatual sense o senses
T propose, for convenience lio to include under the head of nau
ral seme of “mean” sch senses of "mean" as may be exemplified in
Meaning as
seences ofthe pattern “A means (meant) todo so-and-so (by 2)"
‘where A is a human agent. By contast, a the previous examples
show, linclade under the head of nonnaturl senses of mean” any
senses of “mean” found in sentences of the parterns "A means
{emcan) something by x” of "Aenean (reat) by x that. (Ths
‘overigids bu it wil sre as an indication)
do not want to maintain that ll our uss of “mean” fal easily,
‘obviously and ly ato one ofthe two group have distinguished;
bot I think that in most ease we should be at leas finy strongly
Inlined to assimilate a wee of “mean” to one group rather than to
the other The question which now arises is this. “What more can be
said about che distinction benween the cases where we should sy
that the word is applied ina natural sense and the cass whete we
should say that the word i aplid in a nonnatural sense?” Asking
this question wil not of course prob us fom eying to give an
soanatin af een. ems fone aoe ar ee
‘This question about the distinction between natural and non-
‘narural meaning iy I think, what people are geting at when they
splay an interest ina dstincson between "natural" and “conven
tional” signs. But I think my formulation is better. For some things
‘hich can mean, something are not signs (words ae ns) and
Some are ot conventional im any ordinary sense fig. cesta ge
tures; while somethings which mean naturally are not signs of what
they mean (cf the recent budget example.
want fist to consider briefly, and reject, what I might tem a
causal type of answer to the question, “What ie meaning?” We
might ty to sy, for instance, more or les with C. Stevens, that
forxto mean, someting, x mist have (roughly tendency to pro
duce in an audience some atitude (cognitive or otherwise) and a en
‘dency inthe ease ofa speaker, o be prodced by tha atid, these
tendencies being dependent on "an elaborate proces of conditioning
attending the ste of the sgn in communication This lary wl
or do,
(1) Lees considera case where an uterance fit qualifies a all as
‘meanings, something, willbe ofa descriptive or informative kind and
the relevaneateue, therefore, will be cogiive one, for example,
1 fe nd ona SH,26 Semantics and Metaphysis
4 belie (ae “utterance” a a neutal word to apply to any cand-
‘ate for meaning thas 3 convenient ac-objec ambiguity ei 00
“oubt the ease that many peopl have a tendency to pat on a taleoat
when they think shy are about co go to a dance, andi no doubt
iso the case that many people, on scing someone put ona tlcoat,
would conclude thatthe person in question was about tO go t0 8
‘ance, Does tis satisfy us that putting on atalloat means that one
is about to go toa dance (or indeed means, anything a all? Ob-
ously not. eis wo help o eefer othe qualifying phrase “dependent
fon an elaborate process of conditioning,” Foe ll his means i that
the response to the sight ofa ealcoat being pt oni in some way
learned or acquire, it will not exclude the preset case fom being
‘ne of teagan But we hae to take seriously the second part of
the qualifying phrase ("attending the use ofthe sign in communica
‘on”) then the account of meaning obviously circa. We might
just a6 well sy, "X has meanings fi wsed in communication,”
‘which, though tr, not help
(2) if this snot enough, cher isa dfculry—relly the same dt
culty, think wich Stevenson recognizes: how we aero avoid say
ing for example, that “Jones is tal” is pare of whats meant by
“Tomes isan athlete,” since to tell someone that Jones isan athlete
would tend to make him believe that Jones i all Stevenson bere
‘sorts to invoking linguiric rules, namely, a peemisive rae of lan
tage tht “athletes may be nonal.” This amount to saying that we
Ste not prohibited by rue from speaking of "noatll athletes.” But
‘why are we not probed? Not because eis nt bad grammar, ors
‘ot impolite, and so 0, but presumably because its not meaningless
(oe if this i too strong, doesnot in any way woate the eues of
meaning for the expresions concemed). But this seems to iavo¥e ws
in another ice. Moreover one wants to ask why, if i latte
to appeal here to rules to distinguish what i mean fom what is
sugested, this appeal was not made eal inthe ease of groan, for
‘example, to deal with which Stevenson originally introduced the
‘walifying phrase about dependence on conditioning.
'A farther decency ina causal theory of the type jst expounded
seems to be that, even if we acepe it a8 it stands, we are furnished
‘with an analysis only of statement about the stndard meaning, oF
‘the meaning in general ofa “sgn.” No provision is made for deal
with statements about what 2 paricular speaker oe weiter means By
4 sign on » particular occasion (which may well diverge from the
Moning 27
standard meaning ofthe ig): nist obvios how the theory could
be adapted to make such provision. One might even go futher in
iim and mainiain thatthe causal theory ignores the Fact ha the
‘meaning in general ofa sign needs to be explained interme of what
ers ofthe sign do (or should) mean by it on particule ocessons
tnd so the later notion, which i unexplained bythe causal theory,
in face che fundamental one. Iam sympathetic to this more radial
‘iticsm, though Tam aware thatthe point is controversial
ido not propose to consider any father theories of the “casa:
tendency” type. Isuspet no such theory could avoid dickies anal
‘ogour 9 those Ihave outlined without uly losing its lim to rank
‘8a theory ofthis ype
Twill sow ty 2 dierent and, | hope more promising line we
can elucidate the meaning of
‘x meant something (on a paticular occasion)” and
‘x mean that so-and-so (ona partculatoccaion)®
and of
A meant, something by «(ona particular oecssion)” and
“A meant, by x that So-andso (om 3 particular oceason),”
‘his might reasonably be expected wo help us with
‘x means timeless something (hat so-and-so)”
“A means (imeles) by something (hat so-and so”
and with the explication of “means che same as" “undertands”“en-
tals and soon. Let us forthe moment pretend that we have to desl
only with uterances which might be informative or desripaive.
‘Aft shot would bet suggest that "x meant, something” would
be true if = was intended by it tere to induce a bli in some
audience" and that say what the belie was wold be to say what
2xmeantge This will not do. I might lave BY handkerchief near the
Scene of murder in orde to induce the detective to believe that B
‘wae the murderer; but we should not want to say thatthe handler
het or my leaving it there) meant anything ofthat had means.
by leaving ie that B was the madre, Cleary we musta east 24d
thas, for to have mean. anything, aot merly must ie have been
“tered withthe intention of inducing 3 certain ble bu also the
err must have intended an *2udince™ to recognize the intention
behind the uerance.218 Semantics and Metaphysis
“This, though pethaps beter i not god enough. Consider the fol
lowing eases:
(1) Herod presents Salome with the head of St.John the Bapist on
charge.
(2) Feeling fang, child lets its mother se how pale its thoping
‘that she may draw her own conclusions and bel)
(3) Ileave the china my daughter has broken Ing around fr my
wife ose.
Here we seem so have cases which satisfy the conditions so far gven
for meanings For example, Hered inended to make Salome believe
thar St. Jon the Baptit was dead and no doubt ls intended Salome
to recognize chat he intended her to Believe that St.John the Baptist
wae dead, Similarly for the other eases. Yee etainly do not think
that we should want to say that we have hee cases of meaning.
‘What we want to inde the diferenceberween, for example, “e-
liberately and openly leting someone know” and “telling” and be
‘seen “getting someone to think” and “telling”
“The way out is perhaps a follows. Compare the following 80
(0) [show Mr-X a photograph of Me. Y displaying undue falar
ity wo Me X,
(2) I draw » picure of Me, ¥ behaving in this manner and show it
fo Ms. X
‘find that 1 wane o deny that in (1) the phorograph (or my showing
ieeo Me X) meant anything all while wane oasert that in (2)
the picture (or my drawing and showing if) mean, something (hat
‘Mr. Yhad been unduly familiar) ora las that Ha mean by it
that Me. Vhad been unduly familie. Wha is he ference between
the wo cases? Surely that in ca (1) Me. Xs recopition of my ite
tion to mae him believe tha there is something between Me ¥ and
Mrs. Xie (more oles) irrelevant to the produsion ofthis eect by
the photograph. Me X would be ld bythe photograph at leas to
Suspect Mrs. X even if instead of showing to im, had lef
his room by accident; and I (the photograph shower) would not be
tanaware of thi. Buti will make a dierence tothe effec of my pic
ture on Mi. X whether or not he takes me te intending to inform
him (make him believe something) about Mrs. X, and no tobe ust
doodling or teyng to produce a work of art.
Meaning 219
‘But now we seem tobe landed in a further dificult if we accept
this account. For consider now, sa, owning. IT fown sponta
‘neously the ordinary course of evens, someone looking a me may
‘wel teat the frown asa natural sgh of dspeasre. But | fown
elberstly (to convey my displeasure), an onlooker may be ex
pected, provided he recognizes my intention, silo conclude that
fam diplesed, Ought we not then to say ence could no be ex
pected to make any difference tothe onlookers reaction whether he
Tegatds my frown ae spontaneous or as intended to be informative,
thar my frown (deliberate) docs mo! meamn anything? I tink this
dliiculty can be met; fr though in general a deliberate frown may
have the same effec (wih espct to inducing belie in ny displeasure
25a spontaneous frown, i an be expected to have the same effect
‘only provided the audience takes it ak ntended vo convey diplesure
“Thats, f we take aoray the recognition of intention, leaving the ote
circumstances including the ecogition ofthe frown as deliberate,
the blit-producng tendency of the frown must be regarded 35 being
impaired or destroyed.
echape we may sum up whats necessary for Ato mean something
by. as follows. A mus imend o induce by x a ble in an audience,
tnd he mus alo intend his urerance tobe recognized 3 0 intended.
But ehese intentions are not independent; the ecogiton i intended
by A to play its par in inducing he bell, andi dos not do 0
someching will have gone wrong withthe flllmen of A’ intetions.
“Moreover A's intending thatthe ceconition should play this pare
lips, ink, char he assumes that here is ome chance that wil
in fac play this pare, that he docs not regard it ea foregone cont
son that he bli wl be induced in the audience whether or nt the
fneention behind the utterance is recognized. Short, pethaps, we
tay say that “A meant, something by x" is roughly equivalent
"uttered » with the intention of inducing a belie by means of the
sccogiion ofthis intention.” (This Seems to involve a elexve par
sido, bu it doesnot realy do 30)
"Now pethaps i is time to drop the pretense that we have to deal
only with “informative” cases. Let us start with some examples of|
imperatives or quastimperatives. have 2 very avaricious main my
‘oom, and T wane him to go: so [throw a pound note out of the
window: Is thee hee any uterance with 4 meaning? No, becase
in behaving aI did Id not intend his recognition of my prpose
to bein anyway elective in geing him to go. This parallel wo the220 Semantics and Metaphysis
photograph cas. If, on the other hand, I had pointed ro the door of
iver him Ile pur, then my behavior might well be held to con
state a meaningfl, utterance, ast because the recognition of my
intention would be intended by me to be ellecive in speeding his de-
parnure, Another pair of cases would be 1) a policeman who stops 2
‘ar by standing in is way and (2) a policeman who stops a car by
‘waving.
‘Or to tur briey co another ype of cas if, 8am examine, ail
mda, | may wel cause him distress or indignation or hamilaion;
ndil Lam vindictive, may intend this elec apd even intend him 0
fecognize my intention, But T should not be inlined to say that my
failing him meant anything. On the other hand, if cut someone
im the sret, 1 do eel inclined to assimilate this to the eases of
‘eating and this inclination seems o me dependent on the fet
that I could noe easonaby expect hi tobe dressed (indignant,
hunted) unless he recognized my intention to affect hin in this
sway If my college stopped my salary altogether, should acase them
‘of ruining me if they cut it by one pound, I might accuse them of
tnsking me; with some lager eutsT might no know quite what
Tethap then we may make the following generalization.
(() “A meant, something by 3” is roughly) equialen to “A in-
tended the witerance of to produce some eet in an audience by
means ofthe recognition ofthis intention"; and we may add that to
tse what A meant i to ask fora specication of the intended efece
though, of cours, ¢ may nor always be posible to get a straight
answer involving a “that” clause, for example, °abebet that).
(2) "x meant something” is (oughly) equivalent to “Somebody
‘mean, something by x.” Here agai there wil be eases where thi
will ot quite work. el inclined to say that (a regards traf ight)
the change to red meant, that che tae was fStop; but it would
be very unnatural to say, “Somebody (eg the Corporation) mean.
by the relight change thatthe teafic was ro stop.” Nevertheless,
there sem tobe some sort of reference somebody intentions.
(2) "x means, (mcs) that so-and-0” might asa Bs shot be
‘uated with ome statement or dnjunction of statement about what
“people” (vague) intend with qualfcations about “recognition”
cffct by | shall have a wotd to say about this.
‘Will any kind of intended eect do or may there be cates where an
eects intended (with the required Gualifcations) and yet we should
Meaning aa
ot want to talk of meaning? Suppose I discovered some person
constituted that, when I old him that whenever I granted in 2 Sec
‘way I wanted him to blush or to incur some physical mad, there
alter whenever he recognized the grunt (and with fey intention), be
‘id blush or incur the malady: Should we then want fo say tha the
‘grunt meant, something? T donor think so. This pont 0 the fact
that for x to have meanings the intended elect mast be something
hich in some sense is within the contol ofthe audience, otha in
Some sense of “reason” the recognition ofthe intention behind i
forthe audience a reason and not merely 4 case. Ie might lok a if
there isa sort of pun here ("eeason for beeving™ and “reason for
sing”), but Ido no think this is Sevios. For though no doube fro
‘one pine of view questions about reasons fr believing ar questions
about evidence and so quite different from uetions about reasons
for doing, nevertheless to recognize an uterer intention in uteing
2 (descripve uuerance), to have 2 reason for belering that so-and
So, isa eat quite like “having a motive foe” accepting so-and-so,
Decisions “that” seem to involve decisions “to™ (and thsi why we
fan “refuse 10 believe” and aso be “eompelled to believe”). (The
“euting” case needs slightly diferent ueatment, for one ean in
any straightforward sense “decide” to he fended; bu one can etuse
tobe ofende,) Ie looks, then, 25 the intended effect must be some
thing within the control ofthe audience, or atleast the srt of hig
which is within ts contol
(One point before passing to an objection of two. think follows
that from what Ihave said about the connection between mesniniva
and recognition of intention that (insofar as Tam right) only what
may call the primary intention of an utterer is eelevant to the
meaning. ofan unterance. Forfar intending (wit thea of
the ecogition of chi intention to induce an eft E, and intend this
cfc Eto lead wo a farther elect F, then insoae a the ocarrence of
Fis thought tobe dependent solely on F,Leannot reperd F asin the
least dependent on recognition of my intention to induce E. Thats,
f(y) Timend to get man to do something by giving bit some
Jnformation, it cannot be regarded at rlevant 0 the meatig. of
my uterance to desrbe what Iiend him odo
Now some question maybe raised about my se fay fee of uch
words as intention” and "recognition." must disclaim any inten
on of peopling all our talking life with armies of complicated py-
chological cccurtences.I donot hope to salve any philosophical pus22 Semantics and Metaphysics
es about intending, but I do wane briefly to argue that no special
{icles ae erred by my ase of the word “intention” i connection
‘vith meaning Firs, there wl be eases where an uteranc is 2c
panied or preceded by a conscious “plan,” or explicit formulation of|
Intevion (eg. I delare how Lam going to wee or ask myself how
to "get something across”). The presence of such an expii “plan”
obviously counts fay heavily In favor of the utere’s intention
(meaning) being as “planned”; though iis not, think, conclusive:
forexample, speaker who has declared an intention psc familar
cxpresion in an unfamiliar way may slip nto the familia se Sin
layin onlingusc eases: we are asking about an agen iten-
‘on previous expresion counts heavily; nevertheless aman might
plan to throw alee inthe dustbin and yet take iw the post; when
lifting his hand he mighe "come to" and say either “dn intend
to do ths a all” oF “I suppose I must have been intending £0 pt
Explicitly formulated linguistic (or quasinguisie)incentons ae
ro doubt compatatively eae. In thir absence we would seem rely
fon very much the sme kinds of extern a8 we dom he case of son-
linguist imentions where there i a general sage. An teers held
to imtend to convey whats normally conveyed (or normally intended
to be conveyed), and we fequte a good reason for accepting that 2
particular use diverges fom the general usage eg he never knew or
had forgotten the general usage Similarly in nonlingusi ass: we
are presimed to intend the normal consequences of ou actions.
‘Again, incase where there is doube, say, about which of wo oF
‘more things am utere intends to convey, we tend 0 rele othe co
‘ext (linguistic or etberwse) of the uterance and ak which ofthe
alternatives would be elevant o other things he is saying or doing,
for which intention ina partcular situation would Bein with some
Durpose e obviously hs (eg # man who cls for a “pump” ata fre
‘would not wane a bieye pump. Nonlingustic parallels ae obvious:
‘context is 2 eiterion in sting the question of why a tan who has
just put a cgarene in his mouth has put his hand in bis pocker; ele
‘vance to an obvious end criterion in sting why a man running
say from a bal
certain linguistic eases we ask the uterer afterward abou hi
Jncenion, and in fe ofthese case (che very dificult one, sch 38
4 philosopher being asked to explain the meaning ofan unclear pas
‘26 n oe of is works) the answer is nr based on what he femem-
Meaning 23
bers bucis mor like a decision, a decision about how wha he sid is
tobe taken. canoe ind 2 noelingusic parallel here; but the case
S0 special as not to seem to contribute a vital diflerence.
{All this i very obvious; but surly 0 show thatthe criteria for
jing linguistic intentions ae very ike the ete for edging non
linguistic intentions ito show that Linguistic intentions are very like
onlnguiscitetions.