ABSTRACTS |
|
The paper is a discussion the controversial claim that metaphorical utterances have no meaning. I examine, first, Davidson's arguments against meaning-shift theories of metaphor and, second, two literalist accounts of figurative speech: Fogelin's elliptical-simile and Davidson's causal theories. It is concluded that, although metaphorical utterances in general can be said to have a meaning in a Gricean sense, and conventional metaphors have both a meaning and a propositional content. Davidson's discredited claim that metaphors have no propositional content does hold for non-conventional novel metaphors.
(Universidad de Murcia, Spain)
Concepts may be construed as identification procedures, i.e., procedures identifying (if successful) objects in a very broad sense. These 'procedures' can be exactly defined in terms of s.c. (logical) construction (Tichý's term). A conceptual system is a set of concepts a proper subset of which is the set of 'primitive concepts'; the complement of this subset consists of concepts definable in terms of primitive concepts.
Any language is a system of encoding concepts. Thus we must distinguish between an expression and the concept (or the set of concepts) encoded by it. The standard (part of a natural) language establishes an 'official' association of expression with concepts. Idiolects do not respect this association.
Kripke's 'puzzle' can be easily explained away as soon as we recognize that Pierre's association of 'Londres' and 'London' with concepts is based on a conceptual system belonging to a 'personal idiolect'. The s.c. disquotational principle is false, since it mixes up standard language and (personal) idiolects.
(Brno)
In 1966, Keith Donnellan spotted a fact. A description, he claimed, is sometimes used to speak of things, independently of whether or not they satisfy the description. At a party I could indicate you Bob Smith by saying "The man overthere drinking a martini is Bob Smith", even though he
ABSTRACTS* MEANING/ KARLOVY VARY '93 |
51 |
is actually drinking tonic water. Donnellan avoided to account for the fact, and called that use "reverential", and called "attributive" the use of a description to pick out the things that satisfy it. Donnellan alleged that fact as a criticism of both Russell's and Strawson's analyses of descriptions, and avowed some other relevant concomitant facts.
After 1966, there have been many accounts of the Donnellan case. Some other facts have been brought to hear on it, and some theoretical machines have been alleged to explain for it.
What was exactly the Donnellan case? What fact exactly did he spotted? Though so very careful, wasn't his report also misleading? Wasn't some proper theoretical machinery applied to the case inappropriately manouvered?
Briefly, about two of these questions. Donnellan seems concerned with descriptions and their uses. It is generally hold that a description picks out what satisfies the predicate at its core, and that, as a consequence, a sentence with an improper description is always not true. Yet, remarks Donnellan, sometimes when we use an improper description, we happen to state something true of what the description is claimed to describe. Donnellan labels the case as a reverential use of a description. What is peculiar to the case is that we give (or are given) an item and use a description to call attention to the item. That makes irrelevant whether or not the item referred to satisfies the description. The case is close to that where we use a proper name -- the use of a name is standardly reverential. But it has been claimed that appropriateness is so irrelevant that in such a case not only we can use an improper description but also an improper name. An improper name, as much as a proper one, is used referentially. The mechanisms at work cannot be reference rather than satisfaction.
Indeed, to speak of something it isn't required that we refer to or otherwise designate it. We are in a public place with rather few other people. Luckily, there is a strikingly beautiful girl, and I say "Bella!" You understand I am speaking of her, and I know you do, though I haven't said anything which refers to or otherwise designates her. She is relevant and that predication fits her perfectly. If you were not already alert, you would, by surveying the situation, figure out immediately what I am speaking of. I believe that a more faithful report of Donnellan's case would make it close to the last one, and that the machinery evoked by his labels, "reverential" and "attributive", suggested a misleading picture.
(Venezia)
Grice's project to give a analyses of 'meaning' is complicated by the fact that there are two versions of it, the first contained in "Meaning", and the second, a revision of the first, in "Utterer's Meaning. Sentence Meaning and Word Meaning". Both versions share the following general strategy. The first step is to give an analysis of speaker meaning in terms of a
ABSTRACTS* MEANING/ KARLOVY VARY '93 |
52 |
complicated intention of the speaker to bring about a certain effect on the hearer. The idiolect meaning(s) of an expression, in turn, is defined in part by an appeal to this intention. Finally, the timeless meaning(s) of an expression in a language is (are) defined in terms of its idiolect meaning(s). I concentrate my attention on the second, revised version in which characterizations of expressions meaning have the form "X means p". What I wish to show is that any attempt to give an analysis of 'mean' as it occurs in "X means p", which appeals only to the intentions of speakers, no matter how complicated they are, is bound to fail. The reason is that intentions can only give us contents, but what the "p" in, "X means p", requires are characters.
(Philosophy Department, Simon Fraser University, Burnaby, B.C.)
According to Michael Dummett, a speaker's knowledge of meaning must be fully manifested in his ability to use language. This is why meaning cannot be equated straightforwardly with realist truth conditions, and a speaker's understanding with knowledge of those conditions, since in the case of sentences concerning the past, infinities and remote regions of space-time, there is nothing in the behaviour of the speaker to warrant knowledge of how things stand there. However, John McDowell has proposed that we display our knowledge of such things by our comprehending use of the sentences that mention such states of affairs. This modest theory of meaning offers no explanation of our grasp of such contents, and hence no substantial constraints on what those meanings are. Yet for any theoretical account which postulates meanings for our words and sentences there must be an answer to the question of what makes that theory correct. Does is specify the meanings that speakers actually use and understand? Some manifestation constraint is required. However, the behaviour in which speakers display their understanding, according to Dummett, doesn't interpret itself. So what does reveal speakers' meanings in linguistic use? Either that behaviour is described question-beggingly, as the modes theorist proposes; or is characterised theoretically, in which case what is there to block the re-introduction of truth conditions; or Dummett must revert to some non-linguistic theory of thought to provide an account of the contents of linguistic assertion. Is there any way for Dummett to avoid this difficulty? I shall suggest that a proper cognitive characterisation of the notion of a linguistic ability can preserve much of Dummett's full-blooded account of meaning, but that it calls for revision to the manifestation condition.
(Dept. of Philosophy, Birkbeck College, London)
ABSTRACTS* MEANING/ KARLOVY VARY '93 |
53 |
Many writers, among them Hilary Putnam and John Etchemendy, have recently argued that a Tarskian truth-theory for a language L cannot illuminate the semantical properties of L. Scott Soames and Donald Davidson have in different ways attempted to defend Tarski.
I shall in this paper briefly review this criticism of Tarski, and show that it ignores the role of convention T as this is stated. Convention T requires, minimally, a reference to conditions for the application of the ordinary truth-predicate, or something whit that role. It is because convention T holds that it might be possible to see a Tarskian truth-theory as expressing semantical facts. I shall in my paper focus on the conditions for seeing a Tarskian truth-theory as conveying semantic information.
Firstly I shall argue that standard ways of formulating convention T presuppose an individuation of languages which takes semantical properties as essential properties of languages. This conception of languages is shown to be necessary if we are to be guaranteed that theorem in truth-theories are candidates for expressing semantic information.
Secondly I shall argue that this in itself makes it impossible to regard Tarskian truth-theories as empirical theories as Davidson does. In fact there is a possible conflict between regarding truth-theories as empirical and regarding them as guaranteed semantically informative.
I shall then look at Scott Soames's defence of Tarski. I shall show that taking semantical properties as essential properties of languages, as he also does, is not sufficient to give a full defence of Tarski at this point unless it is correct to hold a deflationary concept of truth. If a more substantial concept of truth is operative as the ordinary truth-predicate referred to in convention T, then Soames's defence fails. In this latter case there is semantical information (truth-conditions) which we would like a Tarskian truth-theory to express which Soames's account cannot give us. (We can still see the concept given us by a Tarskian truth-theory as a concept which can replace the intuitive truth-predicate for other purposes than doing semantics, and also think that this is all we should see a Tarskian truth-theory as providing.)
Can a conception of a Tarskian theory as expressing semantical facts still be defended if one does not accept a deflationary view on truth, or is a defence of Tarski connected to a defence of a deflationary view? I shall show that Tarski can still be defended. One natural way implies modifying one's conception of syntax. (An indication of this possibility is to be found in Tarski.) The price of changing the conception of syntax is not high, however, and there are several technical advantages in doing this.
(Dept. of Philosophy, University of Oslo, Norway)
ABSTRACTS* MEANING/ KARLOVY VARY '93 |
54 |
Eliminative materialism is in the first instance a denial of propositional attitude ascriptions such as third person belief sentences e.g. Ralph believes that It is snowing in New York. Propositions (or contents) are then denied secondarily, when one assumes that the logical form of belief sentences commits us ontologically to propositions. An alternative explored in this paper is to put aside or reject eliminativism on belief sentences - propositional attitude ascriptions, and to, so to speak, be an eliminativist on propositions - contents, i.e., provide a logical form for such sentences that doesn't rely on propositions. The project has its origins in Categorial Grammar (Ajdukiewicz-Prior) and substitutional quantification. An attempts is made to explain in a uniform fashion the logical form of sentences containing that clauses without quantifying over propositions.
(New York)
That-clauses refer to meanings: It was Frege's genius to link these two apparently disparate ideas:
1. the meanings of sentences and other parts of speech
2. the semantics of indirect contexts, such as that-clauses
This link gets its plausibility from the truism that
3. 'Snow is white' means that snow is white.
Suppose that 'means' is a perfectly ordinary verb, such as 'hit', except that it relates sentences and their meanings. The phrase to the left of 'means' in (3) refers to a sentence, and the phrase to its right must refer to a meaning. So (3) requires that the meaning of the sentence be referred to by its own that-clause form. This entails Frege's link.
Verbs versus Operators: The above reasoning applies only if there is not something special about the logical form of the verb 'means'. But conventional practice of the last few decades of work in modal, epistemic, and deontic logic suggests that a verb followed by 'that' always has a form that is special. In these enterprises, the role of the word 'that' is to adhere to a verb or adjective to produce an operator, such as 'It is possible that' or 'x knows that', the whole of which combines with sentences in non-truth functional ways. This practices sees 'Mary believes that snow is white' as having the structure
Mary / believes-that / snow is white,
where the that-clause has dissolved, much as definite descriptions dissolve under Russellian analysis. I disagree; the sentence has the structure that one naively attributes to it:
Mary / believes / that-snow is white.
As many authors have noted, this view gets semantic credibility from the fact that valid arguments link that-clauses and ordinary NP's. Anderson 1984 gives the examples
Kant know that 5 + 7 = 12.
ABSTRACTS* MEANING/ KARLOVY VARY '93 |
55 |
It is necessary that 5 + 7 = 12.
\ There is something which Kant knows and which is necessary.
Nobody knew before Gödel proved it that arithmetic is incomplete.
\ There is something which nobody knew before Gödel proved it.
If the verbs are normal, and if that-clauses are referring units, and if the quantifiers behave normally, then these arguments are valid in the predicate calculus. If the that-clauses are torn apart, the validity is a mystery.
Formulation of an Explicit Theory of Meaning: The point of this paper is to explore what an explicit theory of meaning would be like, based on the above observations.
Grammatical Structure: One theme of the enterprise is that identifications of "indirect context" are based on grammatical structure, not on the symptoms of referential opacity. So the context 'An infallible omniscient being knows that ____ is clever' is an indirect one, in spite of its transparency, and the context 'It is not the case that _____ is white' contains and indirect context, again, in spite of the fact that it is transparent. In both cases the transparency is due to a complex of conditions applied to an indirect context.
A Material Adequacy Condition for Meaning: In the theory of reference, we have a clear guide to adequacy: the construction of a theory of truth that meets Tarski's criterion of material adequacy. This criterion states that any adequate theory of truth should yield as output every sentences of the form:
s is true iff p
where 's' is a structural name of p. I suggest a similar criterion adequacy for the theory of meaning: A good theory of meaning should yield all such instances of:
s means that p.
This is not a goal of most work in semantics, such as Montague Grammar, but in can be sought. In the main of part of the paper I give a sample theory that satisfies the material adequacy condition for meaning.
Differences in Meanings: The adequacy condition itself does not specify what the meaning of a sentence should not be. The theory I give, in conjunction with ordinary data, addresses such cases. For example, given the data that Agatha believes that Grenada is a U.S. state but that she does not believe that snow is purple, the theory deductively concludes:
'Grenada is a U.S. state' does not mean that snow is purple.
This phenomenon is compared with work in intensional logic, where the theory of meaning is supposed to yield such consequences without appeal to such data.
Idiolect Problems: Finally, I talk about how to extend the theory to deal with idiolect problems, which arise when we are called upon to discuss the beliefs of a person who speaks a non-standard idiolect. The extension is applied to cases such as Pierre's believing that Londres is pretty.
(U.C. Irvine)
ABSTRACTS* MEANING/ KARLOVY VARY '93 |
56 |
The sentence connectives 'and' and 'but' differ in meaning, but not in a way that shows up truth conditionally: there is no possible world in which 'P and Q' and 'P but Q' differ in truth-value. Moreover, according to Frege, not only does the semantic difference not show up at the level of reference, it does not show up at the level of sense: 'P and Q' and 'P but Q' express the same thought. The words 'but' and 'and' differ only in respect of what Frege calls illumination: 'but' illuminates the shared sense in a way that 'and' does not.
The contributions that 'but' makes to the meanings of sentences containing it ought to be predicted, at least in part, by a compositional semantics for English. (The matter is not purely a question of pragmatics because it involves the conventional meaning of a particular word.) So the question arises how, precisely, within a semantic theory is the difference between 'and' and 'but' to be characterized.
This question would not take on very much importance if it were not for two facts. First, as Grice observes, the case of 'but' is one of a good number involving ordinary expressions (some of which are of philosophical interest): 'therefore', 'consequently', 'so', 'still', 'yet', 'even', and others. It would seem, then, that a rather general theory of illumination (or conventional implicature, as Grice calls it) should be a component of and adequate semantic theory.
Second, as Frege observes, the case of 'and' and 'but' seems to have something in common with cases involving pairs of general terms that may have the same sense such as 'horse' and 'steed'. So there is some hope that by shifting our attention from singular and general terms to connectives, light might be thrown on puzzles concerning substitution in oblique contexts. Certainly this would be so if it could be shown that illumination of sense (a) is a general phenomenon, and (b) involves illumination of reference.
In this paper I will (i) bring out a range of distinct problems that illumination presents for semantic theories, (ii) rebut the idea that some semantically significant notion of "presupposition" is involved, (iii) present an account of illumination that goes some way toward resolving most of the difficulties, and (iv) apply the fruits of the inquiry to the matter of substitution puzzles.
(Berkeley)
We should agree with David Lewis that a language is an abstract object that may or may not be used by anyone. On this conception of language, the big question is: What is it for a language to
ABSTRACTS* MEANING/ KARLOVY VARY '93 |
57 |
be used by, or to be the actual language of, a given person or population of persons? In other words, there is some relation R--call it the actual-language relation--such that a language L is a language of a population P just in case P bears R to L, and the problem--call it the actual-language-relation problem--is to say what R is. A solution to this problem would tell us in what sense, and how, meaning supervenes on the practices and psychological states of language users. After arguing against answers proposed by Lewis, Chomsky, and certain Griceans, I offer what I think is the correct answer.
(New York)
Functional theories of content have been developed mainly to show how mental contents can satisfy certain physicalist constraints. In this paper, we plan to argue, by contrast, that functional theories of content cannot play such a role because there is a significant tension between the multiple realizability of functional properties and such physicalist constraints.
The paper is structured as follows. In the first part, we discredit the notion of narrow content by pointing to a conflict between the presumed explanatory character of such a notion and the idea of multiple realization. Arguments against narrow content tend to be interpreted as an indirect case for the epiphenomenalism of the mental. But such a conclusion will only be legitimate on the assumption that any causally relevant property of an organism must strongly supervene upon its physical properties. We devote the rest of the paper to explore the extent to which functional properties are consistent with this strong supervenience demand. Thus, in the second part, we challenge the metaphysical relevance of such a demand by showing that the only sense in which the principle of strong supervenience is trivially satisfied by functional properties is a sense in which this principle is irrelevant for the current issue about mental causation, the externalist dispute included. Complementarily, we claim, in the third part, that functional properties cannot meet a more demanding principle of strong supervenience, which mental contents might be at pains to satisfy.
(Dept. of Metaphysics, Valencia, Dept. of Philosophy, Murcia).
Saul Kripke's argument defending Russell's theory of (definite) descriptions against the possible charge that the distinction between attributive and referential uses of descriptions marks a semantic ambiguity has been highly influential. The argument is alleged (by Kripke himself) to involve the same line of reasoning as his 'schmidentity' argument. In this paper I argue that the
ABSTRACTS* MEANING/ KARLOVY VARY '93 |
58 |
two arguments in fact draw quite different conclusions from similar thought experiments: the former is purely defensive whereas the latter, if it is sound, actually repudiates a theory of semantic content. My main claim is that the latter, 'offensive', argument is unsound and I attempt to locate the error. Finally, I suggest a moral one might draw from these considerations concerning the semantics - pragmatics distinction.
(Manchester)
Everyone would like to have a satisfactory account of how to understand the content sentence in an attitude report. However, in this paper, I ask a different question about such sentences, namely where do they come from? More specifically, I want to know how the contents of beliefs are related to those of desire, fear and the rest. And, most importantly, I want to understand where knowledge comes into all this. In a somewhat speculative way, I suggest that we may have focused too closely on belief, when in fact knowledge is more central, and I illustrate how this new perspective would affect Kripke's puzzle about belief.
(London)
Within the general framework of the theory of direct reference, there is no agreement as to whether unactualised possible objects can be referred to by means of directly referential singular terms. Three kinds of arguments at least can be provided in order to deny any such possibility, by appealing either to the "causal chain" - version of the doctrine, or to alleged "transcendental conditions of naming" based on the (actual) existence as much as the givenness of the nominatum, or finally to the (either epistemic or ontological) indistinguishability of an unactualised possible entity from its alleged individual essence. None of these arguments is really sound, however. First, there are a lot of cases of directly referred although not causally available entities, or else of causally available but not directly referred entities. Second, if reference is an intentional relation the actual existence of the referent is not required; besides, the only givenness one may ultimately make sense of as a condition for direct reference is not the epistemic, but the modal keeping track of a certain object though different possible worlds. One might retort that with possible unactual objects there cannot be such a keeping track insofar as the third argument above is viable. However, this is precisely not the case, since regardless of its (actual) existence or nonexistence no object can be identified with its individual essence. Rather, insofar as we may have such an essence at our disposal be the only individual which may satisfy it actually existent or not, we may fix the reference of a directly referential term to an unactualised possible object
ABSTRACTS* MEANING/ KARLOVY VARY '93 |
59 |
via a de facto rigid definite description expressing a certain individual essence, as some of the sustainers of the doctrine of direct reference have also maintained.
(University of Palermo, Italy)
The paper is an exercise in translation between conceptual schemes, but it hopes to obtain a surprising result from such a boring undertaking. First, two objections to the causal efficacy of content the "explanatory exclusion" problem and the "supervenience" problem, the second being the claim that Twin-Earth consideration show that externally individuated content properties cannot supervene on intrinsic properties, and therefore their alleged causal efficacy is "screened off" by that of those - are disentangled, and it is claimed that Dretske's attempt at a solution in his Explaining Behaviour can only be rendered as confronting the latter. Then I offer my translation to his account of the causal efficacy of content properties, as "structuring causes" of "behaviour" - 'behaviour' meaning here causal processes starting with neurological occurrences and ending, in the basic cases, with body movements, but not these body movements on their own. According to my translation, "structuring causes" are seen to be partial supervenience explanations (or, better put, plausible outlines of such explanations) for teleologically viewed functional properties. This way, Dretske's dualism of causal explanations ("triggering" versus "structuring" causes) is rendered superfluous, reduced to the already accepted dualism of "horizontal" (causal) explanations and "vertical" explanations of macro-processes by lower level mechanisms. Most of the paper attempts to develop and to justify these points. The surprising (not anymore, in view of recent developments of Fedor's views) result is a novel reply to the Twin-Earth based arguments against the causal efficacy of externally individuated content properties. The reply has it that, in the presence of a (contingently plausible) explanation of how the having of a meaning-property could be (physically and metaphysically) constituted by the having, say, of a neurological property, sheer thought-experiments to the effect that doppergängers might be imagined, having the neurological property but lacking the meaning-property, are irrelevant to the point in dispute, which was the screening-off, or otherwise, of the higher-level property by the lower-level one.
(Departamento de Lógica, Historia y Filosofía de la Ciencia,
Universidad de Barcelona)
ABSTRACTS* MEANING/ KARLOVY VARY '93 |
60 |
Meaning holism is not the frightening doctrine that it is often said to be. It is not, in particular, the form of "radical" or "constitutive" doctrine that the meaning of a single expression or sentence is network dependent of the meanings of all other expressions of a language, with its supposed implications that the individuation of semantic contents and the reality (or at least objectivity) of meanings are impossible. Although these doctrines are indeed frightening, they are not implied by meaning holism properly conceived. This claim implies that one is clear about the various formulations of meaning holism, and its relationships with other forms of holism and other doctrines. I do not propose to dwell too much on this. I shall try instead to formulate meaning holism in the weakest form possible, as a set of trivialities about the ascription of content, such as: in general no behaviour can be considered as the expression of the understanding of a particular statement unless it can be also considered as the expression of the understanding of other contents, that a theory of meaning must suppose certain systematic relations between the meanings of sentences, etc. The status of the trivialities is similar to the status of the trivialities about truth, meaning and truth conditions, in deflationary or minimalist conceptions of truth and meaning. Just as accepting these trivialities does not prevent us for fleshing out more substantial principles about meaning and truth in the realist-anti-realist debate, a form of minimalistic holism does not prevent us from formulating more local substantive accounts of contents, just as it does not prevent us from accepting local holism as well. In the last part of the paper, I shall examine how this connects to the relationship between holism and normativity.
(CREA, Paris)
In indicator systems one can distinguish modes of indications (of properties and objects) from the item indicated. E.g. if an indicator indicates height by indicating air pressure, then height is being indicated "through" pressure - as a property correlated with pressure. In living indicator systems the mode is dictated by the mechanics of indication, and the object indicated by its biological functionality. (The systems are detecting indirectly important sources of food, danger, reproductive affordances and the like). The distinction modes - thing itself mimics quite well the fregean distinction of sense and reference. It is plausible to suppose that the first couple is the biological ancestor of the second - simpler structures of indication initiate and support more sophisticated structures of thought. The distinction can be used to defend teleosemantic theories from the threat of indeterminacy, underscored by authors like Schiffer and Fodor - e.g. is the neural firing in frog's head indicating flies, small black moving objects, dots on retina or something else? On our view the firing indicates flies (referents) under the mode of small-black-moving objects. The distinction significantly reduces the indeterminacy in a systematic and conservative (neo-fregean) way.
(51000 Rijeka, I.G. Kovaćića 26, Croatia)
ABSTRACTS* MEANING/ KARLOVY VARY '93 |
61 |
There seem to be two major schools of philosophy of language in the present century: the logistic school of logical positivists and analytical philosophers, such as Russell, Carnap, Wittgenstein, and Quine; and the structuralistic school connected with the names of de Saussure, Hjelmslev, Jakobson, or Derrida.
However, this kind of classification is rather superficial: the really significant classification issues from considering the way individual philosophers view the link of expression to its meaning; and this classification cuts across standard philosophical schools. There are semiotic approaches which consider the interconnection of expressions in the system of language as derived from the primary fact that expressions stand for objects, and there are structural approaches, which grant primacy to the systemic relations of language and see links of expressions to their meanings as that which is derived. Thus, the semiotic approach is characterized by considering language as a means of codification of the pre-existing, language-independent world; whereas the structural approach sees meaning of an expresion as its value that arises out of the reification of the position of the expression within the system of language.
The semiotic approach is the product of misunderstanding of the real functioning of language; it is the structural approach that is adequate. However, the structural approach is far from being the matter of those who are usually considered as structuralists; on the contrary, its most consequential representatives are the most outstanding analytical philosophers, especially Wittgenstein and Quine.
It is truth, or the opposition between truth and falsity, that is primary; meaning is our means of "implementation" of this opposition. The "implementation" cannot be direct, it must take the detour via meaning, because it must be accomplished compositionally, i.e. respecting the part-whole structure of our language. Meanings are thus products (or by-products) of the compositionalization of truth.
(Prague)
Model-theoretic conventionalism is the view that logic (any logic) is neither more nor less than a theory in the model-theoretic sense, i.e., the result of conventionally selecting a certain class of meaning structures as the only "admissible" models for a given language. In this paper I scrutinize
ABSTRACTS* MEANING/ KARLOVY VARY '93 |
62 |
some basic facts and arguments supporting this view as well as some consequences and complications arising from it.
(Instituto per la Ricera Scientifica e Technologica-IRST, I-38050 Povo,
Trento, Italy)
It is often claimed that normativity is the mark of the meaningful: that the fact that meaningful expression can be true or false, correct or incorrect, is what marks the place of meaning in the world. The same has been said of the contents of intentional states, which will be my concern in this paper. In the dispute over whether intentional content is broad or narrow, many philosophers thing that only a broad theory of content can account for normativity. In this paper, I critically examine this claim. I defend a truth-conditional theory of narrow content, and I argue that not only does this view not have all the problems of normativity that its critics claim, but also that certain broad theories have problems accounting for normativity too. I conclude that if normativity is a problem, it is a problem for reductive theories of content, and not for narrow theories as such.
(University College London)
An argument for the thesis that propositional attitudes must be associated with internal representations (having semantical and syntactical properties) is based on the intuitive idea that a theory of mental content should account for one's conception of the world. What a person believes, desires or hopes is based on her particular outlook on the world. The nature of a mental state is 'outward-looking', but it does not matter to psychology which world, if any, the state actually 'sees'. The meaning or content that must be ascribed for the purposes of psychological explanations is to be abstracted from wide truth-conditional meaning. What is outside the skin cannot determine the content of an attitude. Furthermore, content supervenes on what is inside a person's skin. If not, it would cease to play a causal role in the explanation of actions. The psychological realm is thus autonomous. Mind can be studied in vitro.
Numerous thought-experiments in the philosophy of mind and the philosophy of language which are designed to account for particular forms of externalism, notably Putnam's infamous Twin-Earth-Experiment and Burge's Arthritis-Argument, rely on the assumption that, if we change the environment, we change the truth-conditions of sentences but not necessarily one's conception of the world. The role of external events and objects is thus confined to make beliefs or sentences true; the content it is supposed to determine must not be accessible to and ordinary
ABSTRACTS* MEANING/ KARLOVY VARY '93 |
63 |
believer. This form of externalism identifies in the in situ-conditions with a possible world that does not necessarily relate to mind so as to confer content upon mental states or events.
Based on arguments derived from, externalism in action theory (the idea that (a) actions are identical with bodily movements, and (b) descriptions under which they are intentional assume non-deviant causal relations with external effects), I will reject the assumption that externalism involves a bifurcation of content. My talk will develop three themes: (a) It is not some possible world but the actual world - the world with which a mind is causally related through beliefs and actions - which determines content. This approach does not eliminate truth, on the contrary: truth plays a central explanatory role in a theory of content; (b) Thought experiments that 'spin the possible worlds' don't prove anything about an alleged gap between wide content and narrow content. (c) When mind is studied in situ, representations disappear from the scene. The paper will focus on points (a) and (b).
(Dept. of Philosophy, University of Tilburg, The Netherlands)
The authors aim is to clarify the foundation and the potential of the institutional
account of meaning presented by him at the first Karlovy Vary symposium. The obligatory
starting point is criticism of the intentionalist account. The author argues that the
intuitively based elements of the Gricean-style definitions of utterance-meaning and
speech act types exploit our acquaintance with communicative rules which in fact do not
assign any genuine role to actual speaker's intentions. Other elements of those
definitions reflect rather internal problems of intentionalist theories arising from the
inability to capture the communicative transparency or overtness in terms of speaker's
intentions. The author also questions the wide-spread presumption that regularities in
intentions accompanying communicative behaviour are relevant for the functions of speech
act types and for the fixation of utterance-type meanings. The only link between speech
acts and utterance types on one side and speaker's intentions on the other side, which is
relevant in this respect, is normative. The author suggests a definition of promise in
terms of commitments incurred by the speaker which
(1) purports to give an empirically testable account of what communicatively happens by
the act of promise;
(2) purports to capture the real nature of communicative overtness;
(3) implies a specific account of utterance meaning, sentence meaning and of the
actual-language-relation;
(4) provides, according to the author, a better basis for the analysis of communicative intentions than intentionalist definitions.
ABSTRACTS* MEANING/ KARLOVY VARY '93 |
64 |
The institutional account of meaning need not put any special weight
on linguistic conventions: it just presupposes that an utterance cannot acquire meaning
otherwise than on the background of some general principles with normative power. These
principles cannot be conventions of Lewisian or Schifferean type since these do not
include normativity.
(Prague)
This paper shows how to formulate theories of meaning and truth for natural language. This requires a way to treat "meaning sensitive" contexts; contexts in which the truth of the whole depends on the meaning, not just on the reference, of the part; examples are the contents of 'that'clauses. I treat 'that' clauses as units (as urged by Anderson, Bealer, and others), instead of supposing that 'that' goes with the verb or adjective to form an operator. A second essential idea is to use a metalanguage that itself contains shifted contexts. Theories of truth and meaning can be formulated for a language that contains quantification into meaning sensitive contexts. These theories entail all instances of the schemas:
'S' is true iff S.
'S' means that S.
The theory of truth is parasitic on the theory of meaning, rather than vice versa, as Davidson suggests. The theory of meaning does not rely on the theory of truth. The theory of truth appeals to Frege's insight that the reference of a 'that' clause is the meaning of its contained sentence.