On the autonomy of linguistic meaning.
Frege taught that an indicative sentence can be put forth with any of a variety of illocutionary forces, and he has been taken thereby to have shown that semantic content does not determine the extralinguistic purpose with which that content may be employed. Frege also held that a fully perspicuous language would contain expressions whose function was to indicate the illocutionary force with which the expressions to which they attach are put forth. Although in ordinary language grammatical mood, intonation and contextual factors usually suffice to make clear to the auditor the force of the speaker's speech act, these features are not an infallible guide thereto, as witnessed by the familiarity of such dialogues as the following:
A. You'll be more punctual in the future.
B. Is that an order or a prediction?
It may be hoped that in a more perspicuous language questions such as B's need never be asked due to the presence somewhere in A's remark of an expression indicating the force with which she is putting forth the content in question. For this language to be an improvement upon our own such putative force indicators would have to indicate with great reliability the force with which the content to which they are attached is being put forth. Indeed, it has often been held that if languages of the sort in question are to be of logical interest the force indicators they contain must indicate force indefeasibly; otherwise these languages would at best differ in degree from our own.
This account of what is required of a force indicator is supported by Frege's own discussion of some expressions that he denies fall into this category. He argues (1979, p. 251) that "true" does not express assertion since we can put forth "that it is true that sea water is salty" without asserting that sea water is salty. In "Negation" (1984, p. 383) Frege remarks that "not" does not express the act of denial, since one can say "it is not the case that A", in a sentence such as "If it is not the case that A, then B" without denying A. Similarly in his later logical works Frege inveighs against his own earlier practice of glossing his assertion sign by analogy with expressions such as "the fact that", or "is a fact".(1) Indeed he holds that there is no word, expression or grammatical component of ordinary language whose function is simply to assert something (1979, p. 185).(2) It seems for the mature Frege sufficient to show that an expression E is not an indicator of illocutionary force f if E can be tokened concatenated with a sentence A without A's being put forth with force f.
At first blush, then, Frege requires the assertion sign to be what I shall call a strong illocutionary force indicating device (hereafter strong ifid):(3) an expression any utterance of which indicates that an associated sentence is being put forth by the speaker with a certain illocutionary force. This definition is meant to include embedded as well as free-standing occurrences of the uttered expression (quotation marks excepted). Let "??__" be, syntactically, a function from sentences into sentences, chosen from a set of connectives each element of which is in the domain of a function IF, whose range comprises illocutionary forces. This allows us to speak of "the force associated with connective `??__'". Then let us say that "??__" is a strong ifid iff for all sentential complements A, the utterance of "??(A)", outside of quotation marks, is the putting forth of A with the illocutionary force associated with "??__". Let us also adopt the policy that utterance of a sentence that contains "??(A)" as a proper part is also an utterance of "??(A)". With this definition we may formulate a necessary condition of a symbolism's being logically correct:
Frege Thesis: A logically correct symbolism would contain strong ifids.
In spite of its many adherents this thesis has come under attack. Wittgenstein (1961, 4.442) dismissed what many have thought of as one central type of strong ifid, the assertion sign, as of mere psychological significance.(4) More recently Davidson (1979, 1982) has argued that no expression in any possible language could be a strong ifid, his reason being that jokers, actors and storytellers would exploit any expressions purporting to be strong ifids to serve their own ends.(5) For instance, any expression purporting to be a strong ifid indicating assertoric force would be prone to be used by the actress on stage to heighten the realism of her performance. Yet the actress does not make assertions on stage. At most she represents the character she is portraying as making assertions and other kinds of speech act.(6) Referring to the expression purporting to indicate assertoric force as the strengthened mood, Davidson writes:
It is easy to see that merely speaking the sentence in the strengthened
mood cannot be counted on to result in an assertion; every
joker, storyteller, and actor will immediately take advantage of
the strengthened mood to simulate assertion. There is no point,
then, in the strengthened mood; the available indicative does as
well as language can do in the service of assertion. (Davidson
1979, p. 311)(7)
If Davidson is right then not only does the use of contentful language underdetermine the extralinguistic purpose to which that language may be put; no piece of language, semantically contentful or not, could be such that its very utterance determines the extralinguistic purpose with which that or any other piece of language is being used. On this account, if only Frege had discerned the appropriate strengthening of his thesis that content does not determine extralinguistic purpose, he would not have sought the chimerical strong ifid.
This might give one pause. Frege was no stranger to the difference between genuine and mock speech acts such as those that occur on stage, and, as we shall see, some of his discussions of fictional discourse occur in close proximity to discussions of his assertion sign. Had there been an objection to the status of his force indicators to be found in the realm of fiction, he would in all likelihood have discerned it. Perhaps the trouble is with our own reconstruction of his notion of a force indicator.
In this essay I shall first look to Frege's writings for a new account of the notion of a force indicator. This new account will justify a modification of Frege's Thesis that will make it immune to Davidson's objection. (I do not propose to defend Frege's Thesis in any form, but aim merely to show that formulated appropriately it cannot be ruled out on such grounds as Davidson's.) I shall then provide examples from English of expressions that are force indicators according to the new criterion. One moral will be the not unfamiliar one that Frege took an unjustifiably jaundiced view of natural language. My main aim, however, is to show that expressions with some kind of strengthened mood are indeed possible. For this will show to be at best a half-truth a conclusion that Davidson draws from his criticism of the conception of force indicators given above. He suggests what he terms a Thesis of the Autonomy of Linguistic Meaning:
Once a feature of language has been given conventional expression,
it can be used to serve many extra-linguistic ends; symbolic
representation necessarily breaks any close tie with extra-linguistic
purpose. (Davidson 1979, p. 113)(8)
Among the extra-linguistic purposes Davidson has in mind are the putting forth of a content with a certain illocutionary force. From this thesis Davidson infers a corollary:
... this means that there cannot be a form of speech which, solely
by dint of its conventional meaning, can be used only for a given
purpose, such as making an assertion or asking a question.
(Davidson 1979, pp. 113-4)
The possibility, to be maintained below, of expressions with strengthened mood will undercut this corollary and thereby the Autonomy Thesis. Symbolic representation can be, as one might say, internally related to extralinguistic purpose.(9)
2. Speech acts, fictional discourse and the Begriffsschrift
Some might reply to Davidson's objection to Frege's Thesis by suggesting that Davidson's examples involve non-serious or, to borrow Austin's (1962, p. 22) term, etiolated, uses of language. Perhaps the utterances quantified over in our definition of a strong ifid should be restricted to the serious ones, this restriction involving a consequent weakening in Frege's Thesis. Davidson (1979, p. 111) rejects such a response on the ground that we have no more than an intuitive grasp of the notion of a serious utterance. I would suggest instead that the problem is that our intuitive handle on the notion of seriousness leaves doubt as to whether it is the correct one for present purposes. The actress may be utterly serious in delivering her lines, and the storyteller may relate a somber and sobering tale. At best invocation of the notion of seriousness rules out the joker. (The notion of sincerity is equally unhelpful here. Actors, if they play their roles well, are often said to give sincere performances. Indeed an actor may believe true the indicative sentences that he utters, but even in this case it is not clear that he is making assertions.)
Hare (1989) instead argues in reply to Davidson that it is in principle possible to have a law holding that anyone using an expression E in uttering sentence S is to be taken as having put forth S with the illocutionary force associated with E. Just as in our institution for check writing, if a person signs a check made out correctly the payee is entitled to draw funds from her account, so too, Hare argues, we could have a law holding that anyone attaching an agreed-upon sign of assertion to an indicative sentence is to be taken to be committed to that sentence assertorically (as opposed to not being committed to it at all, or merely as a supposition for the sake of argument). Since assertion of S does not require belief in the truth of S, the possibility that the user of the assertion sign in such a scenario might not believe S to be true would be consistent with his having asserted it. Hare concedes that if we had such an "assertion law" then play-acting would be impossible,(10) but then again even in certain actual societies an actor cannot use blasphemous language on stage without suffering the consequences from the religious authorities.
The debate between Hare and Davidson threatens to reach an impasse at this point, since if Davidson (1979, p. 114) is right to hold that to assert S one must intend to be taken as doing so, there will be a crucial disanalogy between check writing and assertion. There could not be a convention that, say, affirming the consequent is a logically valid inference rule. At most there could be a convention that people are to believe and say that this rule is logically valid. (The existence of such a convention does not of course imply that they always do believe and say this; at most it implies that they should do so.) Likewise if a speaker asserts S only if she intends to do so, then the existence of a convention such as Hare imagines would imply merely that a speaker uttering a sentence containing a sign of assertion is to be treated as having made an assertion. It could only imply that she has made an assertion if it implied that she has the requisite intention, but this it cannot do.
Rather than attempt to settle this debate between Hare and Davidson I propose to make a fresh start on the question of the possibility of force indicators. As a first step, we may observe that the counterexamples to Frege's Thesis we have been considering share in common the feature that speakers involved in them are not--or at least need not be--performing speech acts at all. The actress on stage does not in general make assertions, ask questions, or issue commands. Rather the character that she portrays performs these speech acts, and the actress portrays the character as doing so by making as if to (among other things) perform these acts. In the terminology of Walton (1990), the actress uses herself as a prop in a game of make-believe, and appreciation of her activity involves imagining her to be performing such acts as greeting, requesting and imprecating. Similarly, the storyteller differs from the journalist or the historian in that she makes as if to portray a character who relates a sequence of events, and unlike the journalist or the historian the storyteller does not typically perform speech acts. A practitioner of the New Journalism may intentionally write fiction that is at least in large measure also veracious, but then the sentences she utters will not provide examples fuelling Davidson's skepticism.(11)
This feature uniting the counterexamples that Davidson has offered suggests that we do well to look for a conception of force indicators that applies only to speakers performing speech acts. Frege would have characterized the participants in Davidson's examples as performing only mock or sham speech acts:
As stage thunder is only sham thunder and a stage fight only a
sham fight, so stage assertion is only sham assertion. It is only
acting, only fiction. When playing his part the actor is not asserting
anything: nor is he lying, even if he says something of whose
falsehood he is convinced. (Frege 1984, p. 356)
Indeed Frege seems to have held that a mock assertion must also have a mock or fictional content:
Assertions in fiction are not to be taken seriously, they are only
mock assertions. Even the thoughts are not to be taken seriously
as in the sciences: they are only mock thoughts. (Frege 1979, p.
There is no good inference from the premise that the actor is only pretending to make assertions to the conclusion that the contents of those mock assertions are only mock thoughts, but Frege appears to reason this way. The first of the above two passages occurs in the context of a discussion of whether "true" is an indicator of assertion. When an indicative sentence is put forth without assertoric force, he argues, prefixing it with "It is true that" cannot be counted upon to give it assertoric force. His example is precisely one of utterances made on stage. See also Frege (1979, p. 251). Had Frege meant his assertion sign to be a strong ifid, it is unlikely that he would have failed to consider the possibility of the use of such a sign on stage as a counterexample to his view that that sign imbues a sentence with assertoric force. Likewise, in "A brief survey of my logical doctrines", he writes:
In the Begriffsschrift I use a special sign to convey assertoric
force: the judgement-stroke. The languages known to me lack
such a sign, and assertoric force is closely bound up with the
indicative mood of the sentence that forms the main clause. Of
course in fiction even such sentences are uttered without assertoric
force; but logic has nothing to do with fiction. (Frege 1979,
It is not plausible that Frege would have omitted to consider the question of what the possibility of fictional discourse shows for the tenableness of his assertion sign. What is more, he was in a position to answer that it shows nothing at all. Lumping poetry, novels and drama under the rubric of fiction, Frege stipulated that such uses of fictional language are irrelevant for purposes of logic. He writes:
The logician does not have to bother with mock thoughts, just as
a physicist, who sets out to investigate thunder, will not pay any
attention to stage-thunder. When we speak of thoughts in what
follows we mean thoughts proper, thoughts that are either true or
false. (Frege 1979, p. 130) (See also 1979, pp. 329, 362, 373, 394,
and 1984, p. 152.)
Stage thunder is not a kind of thunder, and so the physicist is entitled to ignore it. Similarly, mock speech acts are not a kind of speech act, and so the logician is entitled to take no notice of them if her aim is merely to introduce a symbol that will determine which force an utterance has (but not whether that utterance is a speech act at all). I shall argue that Frege aimed for this and no more.
He aimed for at least this much. He criticized Peano on the ground that the latter's conceptual notation leaves it up to the reader to determine when an expression is put forth as an assertion or rather as a supposition (Frege 1984, p. 247). There and elsewhere (1984, p. 149; 1968, p. 92) he points out that his own notation makes this distinction manifest, and the marking of this distinction does add to the perspicuity of a notation. For instance, one who wishes to propound the logical law of modus ponens might do so by offering an instance such as: "Snow is white, and if snow is white, then grass is green; ergo Grass is green." However, the audience may be confused as between what is being put forth assertorically and what is not. It is for this reason useful to make clear that in the second premise of this line of reasoning, "Grass is green" is not being put forth assertorically, while the conditional in which that sentence occurs is. Or to use an example from Geach (1976, p. 63), it is valuable to be able to distinguish between a valid reductio ad absurdum argument, such as
P, Q; suppose that R; then not R; ergo not R
and an argument in which the arguer reduces himself to absurdity, as in
P, Q, R; ergo not R.(12)
Correct placement of a sign will distinguish what is being put forth assertorically from what is not.(13)
On the other hand we have evidence that Frege did not intend his assertion sign to determine whether a particular tokening was a component of a speech act. That job was to have been performed by the very construction of the Begriffsschrift.(14) Use of an expression lacking a reference would for Frege have resulted in expression of a thought lacking in reference, and thus, according to him, in that thought's being mock or fictional. Frege (1893, [sections] 6) ruled that a case of this sort cannot arise in his Begriffsschrift by stipulating that all expressions within that formalism must be endowed with a reference. He did not explicitly consider the case of an actor on stage using a Begriffsschrift formula prefixed with the assertion sign, but because he took it that mock assertions must also have mock thoughts as their contents, he would have held that in the very act of simulating assertion the actor has brought it about that, there tokened, those words have no reference. By Frege's lights it would follow that in such a case the actor was merely using expressions homonymous with those of the Begriffsschrift.
Frege's view seems to be, then, that if one is using a formula of the Begriffsschrift and not, say, some orthographically identical formula lacking a reference, then prefixing it with an assertion sign guarantees that it is put forth with assertoric force. The assertion sign can usefully be employed in the marking of such a distinction without being enlisted in the further task of determining whether a speaker is performing speech acts in the first place. This latter determination will be effected by the speaker's use of the Begriffsschrift notation.
I have argued that for Frege cases of the sort Davidson considers provide no counterexample to the possibility of certain signs being stipulated to be strong ifids. This, however, is only because Frege holds both of:
(1) All expressions of the Begriffsschrift have a reference.
(2) If an expression is used by an actor or storyteller then that
expression lacks reference.
One sympathetic to the spirit of Frege's Thesis may look for a modification thereof that does not depend upon these assumptions, for the latter of these is not plausible.
For such a modification one may again begin with Frege. Merely because Frege denies on the ground of embedding considerations that any expression in natural language is a sign of assertion, it does not follow that according to him for an expression to be an assertion sign it must indicate assertion indefeasibly. A weaker and, for all we have said so far, no less adequate principle as far as concerns his logical enterprise is that for an expression to be an assertion sign it must indicate assertion whenever it is used in the performance of a speech act. The hard work comes in formulating such a principle satisfactorily.
Such a formulation should meet at least four desiderata. First of all, if it is to be plausible independently of our ascription of it to Frege the tenableness of this principle should not depend upon thesis (2).(15) Belief in such a thesis may justify Frege in holding that one cannot use a language such as the Begriffsschrift without performing a speech act, but we should not follow him on this point. Second, Frege erred in taking it that the emptiness of a singular term makes for the fictionality of any speech act employing that term.(16) We should strive for a conception of force indicators that does not presuppose a too protean conception of fiction. Third, although Frege held that cases of fiction aside, the utterance of a freestanding indicative sentence of ordinary language is the putting forth of that sentence with assertoric force,(17) this is incorrect, as shown by the fact that one can say:
Let us suppose the next sentence for the sake of argument.
Snow is white.
"Snow is white" is not embedded within any larger sentential context, is in the indicative mood, is put forth in a genuine as opposed to a "mock" speech act, and yet is not put forth with assertoric force. A satisfactory account of the notion of force indicator should accommodate the possibility of such a case. Finally, Frege and many who have followed him have held that no force indicator can have semantic content.(18) For instance in footnote 7 of "Function and concept" he writes:
The assertion sign cannot be used to construct a functional
expression; for it does not serve, in conjunction with other signs, to
designate an object. "??2+3 = 5" does not designate anything; it
asserts something. (Frege 1984, p. 149)
There is no obvious conflict between an expression's designating something and its effecting an assertion in that same usage. On the other hand Frege's highly prized distinction between predication and assertion seems to require that an expression that has semantic content cannot also be a force indicator. A useful clarification of the notion of a force indicator will shed light on this issue.
3. A modification of Frege's Thesis
Assistance is available in speech-act theory, which provides an organon for assessing arguments in terms more inclusive than that of truth-preservation. For instance the theory of speech acts allows us to assess arguments containing non-declarative. One who poses a question is committed to whatever presuppositions that question might have, but, particularly for the case of win-questions, it is not without violence to the phenomena that the interrogator's commitment can be understood in terms of the truthpreservation model. Likewise, those attempting to explain inferences that contain imperatives have often found reason to depart from truth-preservation as their criterion.(19) Speech-act theory provides a notion of validity that applies to declarative sentences (or their contents) and to non-declaratives (or their contents) as well. Further, one might be committed to content A assertorically, but be committed to the content B merely as a supposition for the sake of argument, and if A and B together imply C we must take care to discern the manner in which this pair of commitments engenders commitment to C. A notion of validity that takes account of these differences in the mode of commitment widens our account of how to tell good arguments from bad. One such is an illocutionary notion of validity. LetS be an arbitrary speaker, <[??.sub.l] [A.sub.l] [??.sub.n] [A.sub.n] [??.sub.B]> a sequence of force/content pairs; then:
<[[??.sub.l] [A.sub.l] ..., [??.sub.n] [A.sub.n], ??B> is illocutionarily
valid if if speaker S is committed to each [A.sub.i] under mode [??.sub.i]
then S is committed to B
under mode ??.(20)
Because it concerns what force/content pairs commit an agent to what others, illocutionary validity is an essentially deontic notion. It also requires a codification of the conditions for the successful performance of various kinds of speech act. I will not attempt to provide such a codification here,(21) but it should be observed that one can be committed to a content C under mode m without performing a speech act the content of which is C and having the force associated with m. For instance one can be committed assertorically to a propositional content without having asserted that proposition but rather by having asserted propositions that imply it. Similarly, being committed to a content C under mode m should not be confused with being committed to performing a speech act that has force m and content C. One is committed assertorically to whatever follows logically from what one asserts, but this does not mean that one is committed to asserting whatever follows logically from what one asserts. Instead, in the central case one committed to C assertorically is committed to giving reasons for C if challenged, and is entitled to draw further inferences from C to other contents, to which she will again be committed assertorically. In addition, one committed to C assertorically gives others the prerogative to accept or assert C as well, deferring to the original assertor for justification of C.(22)
We may use the notion of illocutionary validity in providing a new characterization of the notion of force indicator, and on that basis a new formulation of Frege's Thesis. As before let "??__" be, syntactically, a function from sentences into sentences, chosen from a set of connectives each element of which is in the domain of a function IF, whose range comprises illocutionary forces. This allows us to speak of "the force associated with connective `??__'". left [f.sub.??] denote that force. Then we may say that "[??]--" is a weak illocutionary force indicating device (hereafter weak ifid) just in case for all illocutionary forces f' and sentences A, the inference
f'...??(A)... / [f.sub.??] A
is illocutionarily valid. According to this inference, which for reasons that should be clear I shall call a weak-ifid elimination inference, if a speaker is committed under any mode at all to a sentence in which ??(A) occurs, then she is also undertaking or expressing commitment to A with the force [f.sub.??]. (Her doing so is compatible with her putting forth A with other forces as well. It is also compatible with her intending any of various perlocutionary effects.) If there were any weak ifids then although their presence in an utterance would not purport to answer the question whether the utterer is performing a speech act at all, it would provide at least a partial answer to the question what commitments the utterer is undertaking and under what mode.
We are now in a position to consider a
Modified Frege Thesis: A logically correct symbolism would
contain weak ifids.
This is a necessary but not sufficient condition for the correctness of a symbolism. Can it be shown that there could not be a language conforming to Frege's Thesis thus modified? It is at least difficult to see how Davidson's argument could do so, for actors, dissimulators, jokers and storytellers are in general not performing speech acts and so do not satisfy the premise of the above illocutionary inference. Suppose for instance that in a play about Frege the actor inscribes the sentence
?? Arithmetic derives from logic
where "??" is a weak ifid indicating assertoric force. Then the actor is presumably not asserting the sentence thereby tokened but rather is portraying a character as doing so. He therefore provides no counterexample to the inference according to which, for all illocutionary forces f', and sentences A,
f'...??(A)... / [f.??] A.
If on the other hand the actor happens to be Frege portraying himself then we might well say that the actor is asserting the sentence he has tokened, but then the case serves as no counterexample to the claim that "??" is a weak ifid. Similarly, suppose that an agent tokens
?? That is a gorgeous tie you have on
with the clear intention of denigrating the auditor's necktie. If his intention is manifest he will in all likelihood achieve his effect, but he is not undertaking or expressing assertoric commitment to the proposition that the auditor's tie is gorgeous. Indeed he is not putting forth "That is a gorgeous tie you have on" with any illocutionary force whatever, but is instead only making as if to put forth the displayed sentence with assertoric force.(23)
Empty singular terms may be divided into those that are de jure empty ("Zeus", "Sherlock Holmes") and those that are de facto empty ("Vulcan"). Perhaps one can only make as if to perform a speech act with atomic predications containing terms of the former sort; if so then such cases will not provide an objection to the possibility of weak ifids. For negations of such sentences, and for sentences containing de facto empty singular terms, make-believe is a less plausible rubric. On the other hand if one can perform a speech act in the utterance of "Vulcan is beyond Pluto" or "It is not the case that Zeus exists", this is no grounds for doubt concerning the Modified Frege Thesis. Prefixing either such sentence with a weak ifid will simply make it clear which speech act one is performing.
We also do well to consider cases in which a person utters
?? The window is open
with the clear intention of requesting that the auditor close the window. It might be thought that in spite of using a conventional sign for assertion the speaker has not made an assertion but has rather performed a speech act of requesting that the auditor open the window. But from the premise that he makes a request in speaking as he does it does not follow that he is not also making an assertion. Indeed the making of an assertion in that context is what allows the auditor to infer that the speaker wants him to close the window.(24)
It is natural still to worry that if asserting that A requires that one intend to be taken as so doing (a question on which I wish to remain neutral), then we would have a counterexample to a putative illocutionary validity concerning the assertion sign in anyone lacking the requisite intention. That is, if it were held that "??__" is a weak ifid indicating assertoric force, the concern would arise that a speaker might put forth A with some force or other without intending to be taken as putting forth A assertorically. However, according to our definition of illocutionary validity the above inference only implies that the speaker is committed to A assertorically. It does not imply that she asserts A. Even if asserting A requires the intention to be taken as so doing, commitment to A assertorically does not. Otherwise we should be unable to hold that a person is committed assertorically to whatever follows logically from what she asserts.(25)
In propounding the Modified Frege Thesis we are eschewing the ideal of a language in which the very utterance of an expression is an infallible guide to whether the speaker is performing a speech act, and if so, which one. We are instead characterizing a more modest ideal, namely a language in which, if it is known that the speaker is performing a speech act, then the utterance of one of a designated set of expressions allows us to determine what commitments she is undertaking (or expressing) and the mode under which she is doing so. Epistemically, this is to give up nothing as compared with Frege's Thesis. For we should certainly agree that there will be many cases in which it is difficult to tell whether a speaker is performing a speech act or just making as if to do so, and hence whether the presence of a weak ifid can be used to determine the force of her speech act. On the other hand the proponent of Frege's Thesis, even armed with postulates (1) (that all expressions of the Begriffsschrift have a reference) and (2) (if an expression is used by an actor or storyteller then that expression lacks reference), will be no more able to determine in such cases whether the speaker is using the language of the Begriffsschrift (or whatever logically correct language she chooses) or some other homonymous language. But then given the implausibility of (1), it seems that if we must choose between these approaches we cannot do better than to choose that of the Modified Frege Thesis over that of its rival. If, however, this is correct then the necessary condition laid down by the Modified Frege Thesis has already been met, for, as I shall now argue, English contains weak ifids.
4. Weak ifids in English
Urmson (1952) discusses parenthetical expressions in English such as the following:
Mary is, I claim, extremely creative.
Mary is extremely creative, I claim.
As Urmson points out the parenthetical clause serves to indicate that the content, that Mary is extremely creative, is being put forth with assertoric force. While an actor or a student of elocution might utter either of the above sentences without undertaking or expressing assertoric commitment to this content, anyone putting forth either of the above-displayed sentences with some illocutionary force is undertaking or expressing assertoric commitment to this content. This is quite a weak claim, however, for it is not grammatical to embed either of these sentences, and such discourses as the following seem barely intelligible:
Let us ask whether the following is true:
Mary is, I claim, extremely creative.
On the other hand addition of the proform "as" before the pronoun allows both of these uses. Thus consider the expression "--, (as) I claim" taking VPs as inputs and yielding VP's as outputs, and taking Ss as inputs and yielding Ss as outputs.(26) This expression embeds arbitrarily deeply and when not embedded can occur in indicative sentences put forth with other than assertoric force.(27) Yet no matter how deeply embedded, if a speaker puts forth the embedding sentence with some force or other she undertakes or expresses assertoric commitment to the complement of "--, (as) I claim". Similarly, if a speaker puts forth with other than assertoric force a free-standing indicative sentence in which this expression occurs, then so long as she puts forth that sentence with some force or other then she undertakes assertoric commitment to the complement of this parenthetical expression. These facts may readily be explained by the hypothesis that the expression, "--, (as) I claim" is a weak ifid. That is to say that this expression participates in the following illocutionary validity: for all illocutionary forcesf' and sentences A,
f'... A, (as) I claim ... / [f.sub.claim] A.
Since claiming is a species of assertion,(28) anyone who puts forth with some force or other a sentence in which "A, (as) I claim" occurs undertakes or expresses assertoric commitment to A. Analogous remarks may be made concerning other parenthetical speech act verbs in the first person present indicative active. For instance it is also the case that for all illocutionary forces f',
f'... A, (as) I presume ... / [f.sub.presumption] A.
For other indicators of assertoric commitment English also has, "--, (as) I hold", "--, (as) I contend", "--, (as) I aver", and of course also "--, (as) I assert.(29)
In contrast, it is not the case that connectives such as "I assert that --" are weak ifids. For a speaker may put forth "I assert that snow is black" as a supposition for the sake of argument without undertaking or expressing any commitment to the complement, "snow is black". It is thus not the case that for all illocutionary forces f'
f'...I assert that A ... / [f.sub.??] A.
Our proposal then is that Frege erred in claiming that there is no word, expression or construction of natural language whose function is to assert something; "--, (as) I assert" does just this if used in a speech act. Since German contains entirely analogous constructions, it might be wondered why Frege failed to see this. Given his remark in footnote 7 of"Function and Concept", my conjecture is that had he considered parentheticals such as those we have been here scrutinizing he would have inferred, from the fact that they have semantic content, that the only role they can play is to contribute that content to the contents of larger sentences in which they occur. The argument for this thesis appears to be that if an expression has semantic content then it may be embedded within a larger sentential context (such as the antecedent of a conditional or within a temporal operator). But that will then leave undetermined the force with which that larger expression is being put forth.
Developing such a line of reasoning, Stenius (1967) invokes Wittgenstein's example of the picture of a boxer: the showing of the picture may have an indicative function (of indicating how a certain man stood in a certain situation), an imperative function (of indicating how a personought to stand), or an interrogative function (of presenting the person to whom it is shown the task of deciding whether it obtains). Next Stenius argues that if we were to imbue a force indicator (what he calls a "modal element") with meaning, the force indicator would make a contribution to the semantic content of the sentence in which it occurs, "... forming an element of it as a picture. But then there ought to be an additional modal element to show the function of this picture" (Stenius 1967, pp. 258--9).
Stenius apparently holds that nothing can both be an element of a picture and indicate the function of that picture. If this means that no element of a picture can guarantee that there is exactly one purpose such that in every use the picture can only be used for that purpose, then the claim is correct simply because nothing can guarantee such a thing. If on the other hand Stenius intends the stronger claim that there cannot be a purpose such that in every use the picture is used at least for that purpose, then we do well to resist it. The view that a picture can be used for any of a variety of ends does not establish this claim. Further, suppose that the pugilist is Teddy Roosevelt, caricatured in the fashion of Thomas Nast. Then although the gross exaggeration of the moustache, the pince-nez, etc., are parts of the picture, these features also help to indicate that Roosevelt is being characterized as truculent. One cannot appreciate this picture without grasping this characterization.
So it is with expressions having semantic content. There is no bar in principle to our speaking English(*), as similar to English as is compatible with the fact that English(*) contains a verb "to swave" that makes the same semantic contribution to sentences in which it occurs as does "to wave" but differing from the latter in that if a speaker puts forth with some force or other a sentence containing it then she undertakes or expresses assertoric commitment to the smallest indicative sentence in which it occurs (if there is one).(30) In English* one who asserts, for instance, "If John swayed his hand, then Mary will stop her car", is performing a complex speech act (a speech act that comprises more than one force/content pair) in which she
(i) asserts that if John waved his hand, then Mary will stop her car, and
(ii) undertakes or expresses assertoric commitment to the proposition that John waved his hand.
In English*, an assertion of "It is not the case that John swayed his hand" would commit the speaker to an inconsistency; but this, I take it, is a change with which we could learn to live.
Our hypotheses about English* are compatible with a case in which the smallest indicative sentence containing "swave" is put forth with some force in addition to the assertoric. It may seem mysterious how this can be if other than by a perlocutionary effect ("I am swaying to cool off" as a way of asking the auditor to open the window). That mystery is lessened once we see that "swave" performs much the same role as does "-- (as) I claim" when attached to a complement containing "wave". For, as we have seen, although a speaker using the parenthetical will also undertake or express her assertoric commitment to the complement thereof, that complement may be put forth with, say, interrogative force or the force of supposition.(31)
The analogy suggests a question. How could assertion of "It is not the case that, as I claim, John waved his hand" commit the speaker to an inconsistency if not by entailing a contradiction? Yet if this sentence did entail a contradiction we would be ascribing an incoherence to Mary in saying "Mary thinks it is not the case that, as I claim, John waved his hand". We are, however, imputing no incoherence to Mary with this ascription. The simplest account of these phenomena begins with the observation that a sentence containing a parenthetical has the same truth conditions as a sentence just like it but stripped of the parenthetical. That is, I suggest that we take the truth conditions of a sentence that has a parenthetical first person speech act verb in the present indicative active as its major operator to be:
V || A, as I [Phi] || = T iff V || A || = T
where "[Phi]" is a speech act verb in the present indicative active. These truth conditions bear out the intuitions of writers such as Huddleston (1984, p. 399) and Nunberg (1990, p. 39), according to whom parentheticals provide background or secondary information (Huddleston), or supply information that is communicatively supererogatory to the propositional content of the clause in which they appear (Nunberg). Such truth conditions nevertheless allow for the view that parentheticals are literally parts of the sentences in which they occur, and that such sentences, if indicative, have truth conditions. We thus need not follow the drastic approach of McCawley (1981) and Bach and Harnish (1979), according to whom sentences containing non-restrictive relative clauses (a species of parenthetical) lack truth conditions.(32)
As a matter of its contribution to the truth conditions of the sentence in which it occurs, then, the parenthetical clause works in the way that "It is true that" works for a complement such as "snow is white". In light of this account, the two sentences
Snow is white
Snow is white, I claim
have the same truth conditions, namely snow's being white. "I" and "claim", in spite of having a semantic content, do not contribute that content to the truth conditions of sentences in which they occur as parts of parentheticals. That content rather discharges its function via the mechanism of a weak-ifid elimination inference.
According to the present view, assertion of "It is not the case that, as I claim, John waved his hand" commits the speaker (assertorically) by means of semantic entailment to the content that it is not the case that John waved his hand. On the other hand assertion of this sentence commits the speaker by means of illocutionary entailment to the content that John waved his hand. Since assertoric commitment is closed under logical consequence, the speaker is committed to a contradiction (see Harrah (1980) for further discussion). On the other hand if the speaker were merely to put forth "It is not the case that, as I claim, John waved his hand" with the force of a supposition, she would only undertake assertoric commitment to "John waved his hand". Further, in saying "Mary thinks it is not the case that, as I claim, John waved his hand", the speaker ascribes to Mary acceptance of a content that holds just in case it is not the case that John raised his hand. This content is coherent. In speaking as he does about Mary the speaker himself undertakes or expresses assertoric commitment to the content that John raised his hand, and so to his disagreement with Mary.
In contrast to the view espoused here, Urmson (1952) held that the two sentences
1. Your house is, I presume, very old
2. Your house is very old, I presume have the same content as
3. I presume that your house is very old.
He infers from this, together with the premise that in (1) and (2) the parenthetical makes no contribution to the truth conditions of what is said, that in (3) "I presume that" makes no contribution to the content of the sentence in which it occurs. Holding parenthetical expressions to be weak ifids does not require such a view, however. We may all agree that assertion of (3) has the, or an, upshot of putting forth the complement clause with assertoric force, but this does not imply Urmson's conclusion. (Otherwise we would have to infer that "John rightly thinks that A" is just a way of asserting A; it is a way of so asserting but is also more than this.) Further, taking the attitudinative to make no contribution to the truth conditions of (3) makes a mystery of the validity of such inferences as
If I suppose that A, then someone supposes that A
I suppose that A
Someone supposes that A
(I assume that the premise has different truth conditions from "If A, then someone supposes that A".) Similarly for the putting forth of (3) as a supposition for the sake of argument or as an hypothesis to be entertained. (See Recanati (1987) for further discussion.)
Hare infers that in his scenario concerning a sign of assertion, play-acting would be impossible. As already mentioned in note 10, however, all that actually follows is that actors using the assertion sign would be unable only to engage in make-believe assertions in the use of that sign; if they engage in such pretense then by Hare's hypothesis they also undertake assertoric commitment to the sentences prefixed by the assertion sign. It is thus consistent with what Hare's argument establishes that a sentence to which is attached a sign of assertion can be used with other than assertoric force. All that is required of such a case is that the sentence be used with assertoric force as well. On the other hand Davidson's argument against the possibility of a strengthened mood would, if sound, show that a signature on a personal check would not entitle the payee to draw funds on the signatory's account (for a joker or actor on stage might exploit such a convention in the service of his own dramatic ends ...). Davidson establishes only that a sign of assertoric force could not be incapable of being used for anything else than the making of assertions. Similarly for other indicators of force.
From the premise, then, that any conventional expression can be used in pretense it does not follow that no conventional expression is an indicator of force. All that follows is that no conventional expression can be used for nothing other than the indication of force. Nothing Hare has established contradicts this. One aim of the previous section was to dramatize this fact by illustrating with reference to weak ifids in natural language how a sentence can be put forth with more than one force even when that sentence is the complement of a weak ifid. Even if I am incorrect in my claims about English, however, the speciousness of this part of the debate between those who defend, and those who deny, the possibility of force indicators may now be clear.
We may return now to the corollary that Davidson draws from his Thesis of the Autonomy of Linguistic Meaning: there cannot be a form of speech which, solely by dint of its conventional meaning, can be used only for a given purpose, such as making an assertion or asking a question. We have not defended Hare's "strict liability" conception of a force indicator (corresponding to our strong ifid) against Davidson's criticism, and so do not claim that the mere use of any expression is sufficient for the performance of a speech act. But Davidson's corollary, transposed to apply to speech acts rather than uses, admits of two readings:
I. There cannot be a form of speech F and purpose P such that, solely
by dint of the conventional meaning of F, for all speech acts S,
if F is used in S then F is used only to bring about P.
II. There cannot be a form of speech F and purpose P such that, solely
by dint of the conventional meaning of F, for all speech acts S
if F is used in S then F is used to bring about P.
Although we have not denied (I) we have argued that (II) is false: parenthetical speech act verbs in the first person present indicative active are forms of speech belying that claim. But (I) and (II) are each justified, respectively, by the material before and after the semicolon in Davidson's Autonomy Thesis:
Once a feature of language has been given conventional expression,
it can be used to serve many extra-linguistic ends; symbolic
representation necessarily breaks any close tie with extra-linguistic
This is why I said at the beginning of this essay that the Autonomy Thesis is only a half-truth: the claim following the semicolon is incorrect. English shows that there are contentful expressions the use of which in any speech act is sufficient, inter alia, for the achievement of an extra-linguistic purpose.
If our metaphor for the relation of propositional content and force is to be the chemically inspired one of sentence radical and modal element, it must be so revised that it is understood that some radicals, if they are combined with modal elements at all are such that there are certain modal elements with which they must combine. Propositional content and force do not swing as freely from one another as chemical the analogy suggests. The possibility of free radicals, nicely mirrored on the linguistic side by the possibility of putting forth propositions with no force whatever, should not be allowed to obscure this point.(33) (1) See Dudman (1970) for a conjecture as to what might have led Frege into this error.
(2) Frege also holds that the grammatical mood of a sentence's verb is not an indicator of force on the ground that a sentence in the indicative mood can occur as a subordinate clause (1979, p. 198). As we shall see, he also denies that the grammatical mood of a sentence's main verb is an indicator of force due to the possibility of using that sentence in fiction.
(3) I borrow the term "illocutionary force indicating device" from Searle and Vanderveken (1985), but so far as I can discern I do not attach the same meaning to the term as do these authors.
(4) "... in the works of Frege (and Russell) [the assertion sign] simply indicates that these authors hold the proposition marked with this sign to be true" (1961, 4.442). Apparently Wittgenstein would have denied Frege's Thesis on the ground that the presence of strong ifids does not attest to the logical perspicuity of a symbolism.
(5) Davidson would evidently hold that if Frege's Thesis is true there could not be a logically correct symbolism. This does not, of course, commit him to the denial of Frege's Thesis.
(6) I shall elaborate on and qualify this point in the next section.
(7) We should distinguish Davidson's argument against the possibility of a strengthened mood from one argument he gives for the claim that mood is a fallible guide to force. He cites as examples in support of this latter claim such cases as "There is a draft coming through that window", uttered with the intent of requesting the auditor to close the window referred to. As Dummett (1993, p. 209) remarks, however, so long as we do not conflate the point with which a speech act is performed (what Dummett characterizes as a strategic notion) and the force (what he terms a tactical notion) with which it is performed, we may maintain that this utterance has the force of an assertion and the point of a request. Dummett could as well have invoked Austin's (1962) distinction between illocutions and perlocutions.
(8) For other endorsements of this thesis and its implications see also Davidson (1975) and (1982).
(9) The present paper complements the arguments of Clarke (1981), Pendlebury (1986), Belnap (1990), and Segal (1990), according to which grammatical mood is relevant to semantic content. We here argue that semantic content is illocutionarily relevant. Together these works militate against the dominant view of the relation of force and mood, on the one hand, and content on the other, according to which force and mood are to content as functional unity is to radical. (See Stenius (1981, Ch. IX) for elaboration of this Tractarian chemical metaphor.)
(10) Hare need not in fact make this concession. All that follows concerning his scenario is that it would be there impossible merely to play-act in the use of the assertion sign. Similarly our institution of check writing does not make it impossible for an actor (who is perhaps portraying himself) to sign his name to one of his own checks in the course of playing his part. His doing so will cause him to owe money to the payee, but this is compatible with his making as if to do so. I shall return to this point in [sections] 5 below.
(11) Dummett (1973, p. 311) contends that stage assertions are a species of assertion. For a reply see Recanati (1987, p. 266). Also, I need not take issue with the view that the author of a fictional work makes assertions about, and sometimes direct quotations of the words of, fictional entities. If that view is correct then Davidson's examples will not undercut the possibility of strong ifids. On the other hand, the present account of Davidson's examples is not committed to the view that if a sentence is put forth as part of a fictional work or performance, then that sentence is neither true nor false. We shall see that Frege seems to reason this way. Similarly, Urmson writes: "I assert nothing when I make up a story as fiction, so a fortiori I do not assert something that is true or false, even by coincidence" (1976, p. 155). This is a fallacy that depends on equivocation over the "ing/ed" ambiguity with "assert". A writer of fiction might make as if to assert a sentence that is as a matter of fact true. Urmson remarks that it is inappropriate to ask whether the sentence the writer utters is true, but even if we concede this it does not follow that the question cannot be raised.
(12) Bell (1979, p. 85) suggests that the force indicator is not just a valuable device for communication, but is required for an account of validity. This is incorrect. If an argument's validity consists in there being no way for the premises to be true and the conclusion false, then any arbitrary attachment of illocutionary forces to the premises or the conclusion of that argument will make no difference to the validity thereof. The most that may be said is that indication of force is important to the notion of inference, and in particular to the use of inference rules. It is difficult to see how one could understand the inference rule of reductio, for instance, unless one were able to grasp the distinction between assertion and supposition for the sake of argument. As we shall see below ([sections] 3), force indicators are usefully employed in what I shall call an illocutionary account of validity, but that is not an account that Frege foresaw.
(13) Anscombe (1963, p. 113) gives as one of Frege's reasons for use of the assertion sign that it solves a problem raised by the apparent possibility of equivocation in such inferences as modus ponens:
If P, then Q; P; ergo Q.
The two occurrences of P in this inference differ in that one is put forth assertorically while the other is not. Yet we need to be able to say that it is one and the same content that occurs in the antecedent of the conditional as occurs free-standing in the second premise of that inference. A distinction between force and content, marked by the presence of an assertion sign before the first occurrence of "P" but not before the second occurrence, allows us to see how modus ponens does not rest upon a fallacy of equivocation. This may be a good ground for introducing a sign of assertion, but I know of no passages in which Frege explicitly cites this problem as a justification for the use of his assertion sign.
(14) In what follows I shall use "Begriffsschrift" to refer to Frege's concept script (both that of the 1879 work and that of the 1893 work, the ambiguity will be harmless here), and use "Frege 1879" or some variant thereof to refer to the book entitled "Begriffsschrift".
(15) In addition to the implausibility of that conditional there is also the difficulty of making sense of the idea of fictional thought. See Evans (1982, Ch. 1) for fuller discussion. (16) Frege writes:
Is that lime tree my idea? By using the expression "that lime tree" in
this question I am really already anticipating the answer, for I mean
to use this expression to designate what I see and other people too
can look at and touch. There are now two possibilities. If my
intention is realized, if I do designate something with the expression
"that lime tree", then the thought expressed in the sentence "That
lime tree is my idea" must obviously be denied. But if my intention
is not realized, if I only think I see without really seeing, if on th at
account the designation "that lime tree" is empty, then I have
wandered into the realm of fiction without knowing it or meaning to.
In that case neither the content of the sentence "That lime tree is
my idea" nor the content of the sentence "That lime tree is not my
idea" is true, for in both cases I have a predication which lacks an
object. So then I can refuse to answer the question, on the ground
that the content of the sentence "That lime tree is my idea" is
fictional. (Frege 1984, pp. 361-2)
(17) "Fiction apart, it seems that it is only in subordinate clauses that we can express thoughts without asserting them" (Frege 1979, p. 198).
(18) Others who have shared this view of Frege's are Reichenbach (1947), Geach (1965, 1975), Stenius (1967), and Hare (1970, 1989).
(19) See Belnap (1990), Pendlebury (1986), and Harnish (1994) for arguments in support of taking non-declaratives at face value.
(20) An argument is illocutionarily sound just in case it is both illocutionarily valid and all its premises are such that their conditions of satisfaction are met. A fuller account of illocutionary validity would employ further distinctions. For example, two assertoric commitments may be alike save for differing in strength; these differences are usually described as differences in personal probability, and we could if we wished distinguish among differences in the strength of commitment. Again, we could distinguish among the different possible objects of commitment, since there is nothing to rule out being committed to a question or to an imperative. These distinctions within the dimensions of mode, strength and object of commitment are taken into account in Green (1993). See also Harrah (1980) for a discussion of assertoric, erotetic, and projective commitment.
The truth-preservation notion of validity may be seen as a special case of the commitment-preservation notion as follows: treat each of the sentences in the argument counting as valid in the former sense as being put forth in assertoric mode and treat each such sentence as declarative. Illocutionary validity is thus not a rival to the truth-preservation notion, but is instead a generalization thereof.
(21) Some of this work has been carried out by Searle end Vanderveken (1985, Ch. 4) under the rubric of weak illocutionary commitment, and in Vanderveken (1990). See also McArthur and Welker (1974).
(22) See Brandon (1983, 1994), Green (1993, 1995), and Belnap and Green (1993) for a fuller account of the deontic status that characterizes assertion.
(23) The approach of the Modified Frege Thesis need not follow Frege's postulate (2) (see p. 226 above) in inferring the fictionality of the content of an act from the fictionality of that act itself. Also, the present view of irony sits comfortably with Grice's analysis of irony in terms of implicature. (See Grice 1975, p. 34, and pp. 53-4.) For such an analysis will start with the datum that the speaker has made as 0if to say that the tie in question is a gorgeous one, and Grice's definition of conversational implicature explicitly provides for the case in which a speaker has only made as if to perform a speech act. (See Grice, 1975, pp. 30-1.) The implicature is generated in part by the fact that if the speaker were actually asserting the above-displayed sentence he would be violating the Maxim of Quality, which enjoins one not to say what is believed to be false.
(24) Here again Dummett's (1993) distinction between strategic and tactical aspects of language is helpful. For a defense of the approach to indirection suggested in the text see Bertolet (1994).
(25) I am presupposing that unlike assertoric commitment, assertion is not closed under logical consequence. one does not assert whatever follows from what one asserts.
(26) I shall only discuss the latter, assuming that the two structures are to be identified at LF by means of a transformation rule such as that postulated by McCawley (1982).
(27) To see the latter point, imagine a speaker to say: "Suppose the following for the sake of argument: Mary is, as I claim, extremely creative". She might wish to make clear her assertoric commitment to the content before an audience that does not accept it, while entering this content as a supposition in the hopes at least of convincing her audience of the truth of a conditional whose antecedent this is.
(28) See Vanderveken (1990, vol. I, p. 169) for an extensive list of assertive verbs.
(29) Parenthetical speech act verbs in the first person present indicative active are a special case of the larger category of parenthetical speech act verbs. Thus English also allows
If, as John holds, snow is white, then grass is green.
Either grass is green, or, as Mary used to contend, roses are red.
It is not the case that violets are, as everyone in the room presumes, blue. Putting forth the second of these sentences, for instance, with some force or other results in the speaker's undertaking or expressing assertoric commitment not to the content that roses are red, but to the content that Mary used to contend that roses are red. Similarly, strictly speaking the putting forth with some force or other of "A, as I claim" illocutionarily implies the undertaking or expressing of assertoric commitment to "I claim that A". However, the following inference is il- locutionarily valid:
?? I claim that A / [f.sub.claim] A.
As a result we have as a derived illocutionary inference rule that for all f'
f'... A, (as) I claim ... / [f.sub.claim] A.
Parenthetical speech act verbs in the first person present indicative active thus participate in weak-ifid elimination inferences as displayed in the text, but I do not make this claim of parenthetical speech act verbs generally.
(30) In fact so-called factive verbs approximate to this ideal. For instance, many embedded occurrences of the verb "regret" are ones in which the speaker undertakes commitment to the complement. An assertion of "If John regrets having eaten all the brownies, then he should replace them", typically commits the speaker to John's having eaten all the brownies, as does, "John does not regret having eaten all the brownies". Yet an assertion of "John doesn't regret having eaten all the brownies, because he isn't the one who ate them", does not seem to commit the speaker to the complement of this occurrence of "regret". (See Hooper (1975) for fuller discussion.)
(31) Clearly nothing in our account rules out the possibility of an unembedded indicative sentence lacking a weak ifid being put forth with other than assertoric force. Thus one of the desiderata mentioned at the end of [subsection] 3 has been satisfied in our notion of a weak ifid. However, the present approach also allows one to see how an unembedded indicative sentence can be put forth both (say) as a supposition and as an assertion without either such speech act being indirect.
These remarks also show that while the general strategy of explaining judging in terms of predication is not plausible (since there are so many predicates that can be predicated of a singular term without any judgement being effected) it does not follow that that strategy is useless in all cases. "A judges true `B swaves to C' just in case A predicates swaying to T of `B'" cannot be ruled out merely on the basis of embedding considerations, since even in accepting "If Bswaves to C, then C will veer to the left" A is judging true the antecedent. (32) An analogy with the truth conditions of sentences containing parenthetical speech act verbs in the first person present indicative active may be found in the truth conditions of sentences conjoined with "but" as opposed to "and". "A but B" has the same truth conditions as "A and B", while, unlike the latter, use of the former in a speech act commits the speaker to there being some kind of conflict or contrast between A and B. Notice finally that a clause such as
V || A, (as) I [Phi] || = T iff
V || A || = T and V || IPhi] that A || = T
will not do because of embedding considerations.
(33) My thanks to the following people for helpful discussions: Nuel D. Belnap James Cargile, Joseph Camp, Louis De Rossett, Ray Jennings, and John Marshall.
Anscombe, G.E.M. 1963: An Introduction to Wittgenstein's Tractatus, 2nd edition. London: Hutchinson University Library.
Austin, J.L. 1962: How to Do Things with Words. Cambridge, Mass.: Harvard University Press.
Bach, Kent and Harnish, Robert 1979: Linguistic Communication and Speech Acts. Cambridge, Mass.: MIT Press.
Bell, David 1979: Frege's Theory of Judgement. Oxford: Oxford University Press.
Belnap, Nuel 1990: "Declaratives Are Not Enough". Philosophical Studies, 59, pp. 1-30.
-- and Green, Mitchell 1993: "Indeterminism and the Thin Red Line", in J. Tomberlin, ed., Philosophical Perspectives 8: Logic and Language. Atascadero, California: Ridgeview Press, pp. 365-88.
Bertolet, Rod 1994: "Are There Indirect Speech Acts?", in S. Tsohatzidis 1994, pp. 335-49.
Brandom, Robert 1983: "Asserting". Nous, 17, pp. 637-50.
-- 1994: Making It Explicit: Reasoning, Representing, and Discursive Commitment. Cambridge, Mass.: Harvard University Press.
Clarke, Romaine 1981: "Moods and Modalities". Acta Philosophica Fennica, 32, pp. 44-59.
Davidson, Donald 1975: "Thought and Talk", in Davidson 1984, pp. 155-70.
-- 1979: "Moods and Performances", in Davidson 1984, pp. 109-23.
-- 1982: "Communication and Convention", in Davidson 1984, pp. 265-80.
-- 1984: Inquiries into Truth and Interpretation. Oxford: Oxford University Press.
Dudman, V.H. 1970: "Frege's Judgment-stroke". The Philosophical Quarterly, 20, pp. 150-61.
Dummett, Michael 1973: Frege: Philosophy of Language. London: Duckworth.
-- 1993: "Mood, Force and Convention", in his The Seas of Language. Oxford: Oxford University Press, pp. 202-23.
Evans, Gareth 1982: The Varieties of Reference. Ed. J. McDowell, Oxford: Clarendon Press.
Frege, Gottlob 1879: Begriffsschrift, in van Heijenoort, ed., 1976: From Frege to Godel: A Sourcebook in Mathematical Logic. Cambridge, Mass.: Harvard University Press, pp. 1-82.
--  1964: "On the Scientific Justification of a Concept-script", trans. by J.M. Bartlett. Mind, 73, pp. 155-60.
--  1968: "On the Purpose of the Begriffsschrift", trans. by V.H. Dudman, Australasian Journal of Philosophy, 46, pp. 89-97.
--  1964: Basic Laws of Arithmetic, partial trans. by M. Furth. California: University of California Press.
-- 1979: Posthumous Writings. Eds. H. Hermes, F. Kambartel and F. Kaulbach. Oxford: Basil Blackwell.
-- 1980: Philosophical and Mathematical Correspondence. Eds. G. Gabriel et al. Chicago, Illinois: University of Chicago Press.
-- 1984: Collected Papers on Mathematics Logic and Philosophy. Ed. B. McGuiness. Oxford: Basil Blackwell
Geach, Peter 1960: "Ascriptivism", in his 1972, pp. 250-4. Originally published in 1960 in The Philosophical Review, 69, pp. 221-6.
-- 1965: "Assertion", in his 1972, pp. 254-69. Originally published in The Philosophical Review, 74, pp. 449-65.
-- 1972: Logic Matters. Oxford: Basil Blackwell.
-- 1976: "Saying and Showing in Frege and Wittgenstein". Acta Philosophica Fennica, 28, pp. 54-70.
Green, Mitchell 1993: Illocutions and Attitudes. Ph.D.-Dissertation, University of Pittsburgh.
-- 1995: "Quantity, Volubility, and Some Varieties of Discourse". Linguistics and Philosophy, 18, pp. 83-112.
Grice, H.P. 1975: "Logic and Conversation". in his Studies in the Way of Words. Cambridge, Mass.: Harvard University Press, pp. 1-144.
Hare, R.M. 1970: "Meaning and Speech Acts". The Philosophical Review, 79, pp. 3-24.
-- 1989: "Some Subatomic Particles of Logic". Mind, 98, pp. 23-37.
Harnish, Robert 1994: "Mood, Meaning and Speech Acts", in Tsohatzidis 1994, pp. 407-59.
Harrah, David 1980: "On Speech Acts and Their Logic". Pacific Philosophical Quarterly, 61, pp. 204-11.
Hooper, Joan 1975: "On Assertive Predicates". Syntax and Semantics, 4, pp. 91-124.
Huddleston, Rodney 1984: Introduction to the Grammar of English. Cambridge: Cambridge University Press.
McArthur, Robert and Welker, David 1974: "Non-assertoric Inference". Notre Dame Journal of Formal Logic, 15, pp. 225-44.
McCawley, James 1981: "The Syntax and Semantics of English Relative Clauses". Lingua, 53, pp. 99-149.
-- 1982: "Parentheticals and Discontinuous Constituent Structure". Linguistic Inquiry, 13, pp. 91-106.
Nunberg, Geoffrey 1990: The Linguistics of Punctuation. Palo Alto, California: Center for the Study of Language and Information.
Pendlebury, Michael 1986: "Against the Power of Force: Reflections on the Meaning of Mood". Mind, 95, pp. 361-72.
Recanati, Francois 1987: Meaning and Force: The Pragmatics of Performative Utterances. Cambridge: Cambridge University Press.
Reichenbach, Hans 1947: Elements of Symbolic Logic. London: Macmillan.
Searle, John and Vanderveken, Daniel 1985: Foundations of Illocutionary Logic. Cambridge: Cambridge University Press.
Segal, Gabriel 1990: "In the Mood for a Semantic Theory". Proceedings of the Aristotelian Society, 91, pp. 103-18.
Stenius, Eric 1967: "Mood and Language-game". Synthese 17, pp. 254-74.
-- 1981: Wittgenstein's Tractatus: A Critical Exposition of Its Main Lines of Thought. Westport, Connecticut: Greenwood Press.
Tsohatzidis, Savas, ed., 1994: Foundations of Speech Act Theory: Philosophical and Linguistic Perspectives. London: Routledge.
Urmson, J.O. 1952: "Parenthetical Verbs", reprinted in A. Flew, ed., 1956: Essays in Conceptual Analysis. London: Macmillan. Originally published in Mind, 61, pp.480-96.
-- 1976: "Fiction". American Philosophical Quarterly, 13, pp. 153-7. Vanderveken, Daniel 1990: Meaning and Speech Acts, Vols. I and II. Cambridge: Cambridge University Press.
Walton, Kendall 1990: Mimesis as Make-Believe. Cambridge, Mass.: Harvard University Press.
Wittgenstein, Ludwig  1961: Tractatus Logico-Philosophicus. London: Routledge.
MITCHELL S. GREEN Corcoran Department of Philosophy University of Virginia 521 Cabell Hall Charlottesville, VA 22903 USA msg6m@,virginia.edu
|Printer friendly Cite/link Email Feedback|
|Author:||Green, Mitchell S.|
|Date:||Apr 1, 1997|
|Previous Article:||How Are We to Live?|
|Next Article:||Explaining referential/attributive.|