Find counter arguments by entering NameVs… or …VsName.
The author or concept searched is found in the following 3 entries.
| Disputed term/author/ism | Author |
Entry |
Reference |
|---|---|---|---|
| Compositionality | Schiffer | I XVIII SchifferVsCompositionality: we must reject it because we must also reject the theory of relation (without which we cannot have the compositionality). >Frege principle, >Relation theory/Schiffer. Understanding/Schiffer: understanding must be explained otherwise: Solution: Schiffer thesis: conceptual role in neuronal lingua mentis without compositionality. >Conceptual role, >Lingua mentis, >Language of Thought. I 183 SchifferVsCompositionality: verbs for propositional attitudes can hardly be put into a compositional semantics. In addition e.g. "is a picture of", "true", "big", "toy"(soldier) - adverbs, evaluative terms like "should", "good", pronouns and demonstrative pronouns "everyone", "all" are problematic. Also counterfactual conditional and modal expressions represent a problem for compositional semantics. >Counterfactual conditionals, >Adverbs, >Adjectives. I 183 Compositional truth theoretical semantics/Schiffer: attributes truth conditions to sentences. >Truth conditions. I 184 Compositionality/SchifferVsCompositionality/SchifferVsFrege: natural language does not need any compositional semantics for understanding. >Understanding. For new sentences, we are not confronted with new words and even only with known constructions. Pro Frege: meaning theory must determine compositional mechanisms, but this does not lead to the fact that the meaning theory must be truth-theoretical (must determine truth conditions). >Meaning theory, >Truth-conditional semantics. I 208 SchifferVsCompositionality/SchifferVsFrege: E.g. "and": the everyday linguistic meaning is not captured by the truth value table. >Truth table. Compositional semantics would require that there is a non-logical axiom for each non-logical expression. - This is not possible. Propositions by E. Harveys spoken language receive their representational character via the connection with mental representation. >Mental representation. Therefore Mentalese does not need compositional semantics. >Mentalese, >Language of thought. |
Schi I St. Schiffer Remnants of Meaning Cambridge 1987 |
| Context/Context Dependence | Hintikka | II 108 Context Dependency/context/compositionality/Frege principle/Hintikka: problem: context dependency violates the Frege principle. ((s) The meaning of a sentence can change then, although no component changes.) >Frege principle. Any/every/he/a/Hintikka: bad solution: it is not a good solution to analyze (16) (16) (Ex) George knows, that (w = x) as (20) John does not believe Mary likes him. Problem: (16) says that it is compatible with John's beliefs that Mary does not love one while (20) is compatible with the fact that John does not believe Mary likes him (John). This is then compatible with the fallacy of (17). (17) ~John believes, that (Ex)(x is a boy & Mary likes x) II 109 Any/context dependency/context/Hintikka: what we need is an explanation of how the interpretation of "any x" depends on the context. II 109 Frege principle/compositionality/Hintikka: if we proceed from the outside to the inside, we can allow that the Frege principle is violated (i.e. the semantic role of the constituents in the interior is context-dependent). II 110 HintikkaVsFrege/HintikkaVsCompositionality: thesis: meanings (meaning entities) should not be produced step by step from simpler ones in tandem with syntactic rules. They should instead be used as rules of semantic analysis. >Syntax, >Semantics. |
Hintikka I Jaakko Hintikka Merrill B. Hintikka Investigating Wittgenstein German Edition: Untersuchungen zu Wittgenstein Frankfurt 1996 Hintikka II Jaakko Hintikka Merrill B. Hintikka The Logic of Epistemology and the Epistemology of Logic Dordrecht 1989 |
| Sense | Dummett | II (b) 72ff Sense/Dummett: The sense is a property of a single sentence, not of the full use of language. I 122 Sense/Capture/DummettVsFrege: The thought is not presented directly to the consciousness - rather grasping the sense: set of skills. I 123 The fact that the glasses are in the other jacket cannot be content of consciousness. I 124 Such examples (glasses) lead to an opposite direction according to which language is explained by thoughts that are considered to be language-independent, and not vice versa (> DummettVsEvans). - Saussure s conception of language as a code will avoid such a declaration - VsFrege: sense of the word is not the same as a part of the thought. EvansVsCompositionality. III (a) 25 Sense/Dummett: from division of states of affairs: 1) where the statement could be misused. 2) where it could be not misused. ad 1) Statement: false - conditional: false - atomic sentence or without truth value. ad 2): statement: true, Conditional: true or without truth value, atomic sentence: true. III (a) 28 Sense/Dummett: entirely determined by one knowing when it has an designated truth value and when a non-designated - finer distinctions only needed in complex sentences with operators. >Truth value/Dummett. III (b) 74 Sense/Dummett: not only through verification method, but understanding what circumstances must be realized. >Circumstances. |
Dummett I M. Dummett The Origins of the Analytical Philosophy, London 1988 German Edition: Ursprünge der analytischen Philosophie Frankfurt 1992 Dummett II Michael Dummett "What ist a Theory of Meaning?" (ii) In Truth and Meaning, G. Evans/J. McDowell Oxford 1976 Dummett III M. Dummett Wahrheit Stuttgart 1982 Dummett III (a) Michael Dummett "Truth" in: Proceedings of the Aristotelian Society 59 (1959) pp.141-162 In Wahrheit, Michael Dummett Stuttgart 1982 Dummett III (b) Michael Dummett "Frege’s Distiction between Sense and Reference", in: M. Dummett, Truth and Other Enigmas, London 1978, pp. 116-144 In Wahrheit, Stuttgart 1982 Dummett III (c) Michael Dummett "What is a Theory of Meaning?" in: S. Guttenplan (ed.) Mind and Language, Oxford 1975, pp. 97-138 In Wahrheit, Michael Dummett Stuttgart 1982 Dummett III (d) Michael Dummett "Bringing About the Past" in: Philosophical Review 73 (1964) pp.338-359 In Wahrheit, Michael Dummett Stuttgart 1982 Dummett III (e) Michael Dummett "Can Analytical Philosophy be Systematic, and Ought it to be?" in: Hegel-Studien, Beiheft 17 (1977) S. 305-326 In Wahrheit, Michael Dummett Stuttgart 1982 |
| |||
| Disputed term/author/ism | Author Vs Author |
Entry |
Reference |
|---|---|---|---|
| Analyticity | Fodor Vs Analyticity | IV 185 Analytic/synthetic/gradual analyticity/Block/Fodor/Lepore: some authors have concluded from "Two Dogmas" that a certain "gradual analyticity" is not excluded. IV 185 Fodor/LeporeVs: this then presupposes equality of meaning rather than identity of meaning. But we have already seen that for inferences analyticity and compositionality are the same. Then one must live with gradual compositionality as well. Question: is this also possible together with systematicity (systematics: believing related attitudes), isomorphism (see above), and productivity? Would gradual compositionality not only include a finite acquaintance with (infinite) language? So that you only "kind of" understand new concepts? E.g. if you understand aRb, then you "kind of" understand bRa. E.g. the constituents of the sentence S "kind of" express the constituents of the proposition P?. E.g. "John loves Mary" "kind of" expresses that John loves Mary, but only because "John" refers "approximately" to John? 29 IV 185 Analytic/synthetic/Quine/Fodor/Lepore: you may wonder how we agree with Quine about the a/s distinction (group), but still stick to compositionality including analyticity and that languages are compositional. This is not a paradox: compositionality licenses structurally determined analyticity: IV 245 E.g. "brown cow", "brown" but not "cow" >Animal. Quine: "Logic is chasing truth up the tree of grammar". IV 178 Fodor/Lepore/QuineVsKant/QuineVsAnalyticity/QuineVsCompositionality of Inference: (external): it must be possible for conclusions to turn out to be wrong. IV 178/179 VsFodor/Lepore: then one might have a reformulated CRT (conceptual role theory): this one has compositional meaning, but the inferential role is not compositional, only within analytical conclusions? Fodor/LeporeVsVs: there is a risk of circularity: if you assume analyticity at all, compositionality, analyticity and meaning spend their lives doing the work of the others. Quine would say: "I told you!". Inferential Role/Fodor/Lepore: the present proposal also threatens their naturalisability ((s) that they are ultimately explained in physiological categories): originally, their attractiveness was to provide a causal role as a basis for the solution of Brentano’s problem of irreducibility to the neurophysiological (>Computation). |
F/L Jerry Fodor Ernest Lepore Holism. A Shoppers Guide Cambridge USA Oxford UK 1992 Fodor I Jerry Fodor "Special Sciences (or The Disunity of Science as a Working Hypothesis", Synthese 28 (1974), 97-115 In Kognitionswissenschaft, Dieter Münch Frankfurt/M. 1992 Fodor II Jerry Fodor Jerrold J. Katz Sprachphilosophie und Sprachwissenschaft In Linguistik und Philosophie, G. Grewendorf/G. Meggle Frankfurt/M. 1974/1995 Fodor III Jerry Fodor Jerrold J. Katz The availability of what we say in: Philosophical review, LXXII, 1963, pp.55-71 In Linguistik und Philosophie, G. Grewendorf/G. Meggle Frankfurt/M. 1974/1995 |
| Collins, A.W. | Fodor Vs Collins, A.W. | Cresswell II 157 Sentence/reason/mental object/Collins/Cresswell: (Collins 1979, 225f) thesis: sentences are mental particulars ((s) VsCompositionality). Problem: but everything that can have a truth value (true value) must be a universal. Mental Events/Collins: here we need temporality. Truth/Collins: the carriers of truth and falsehood need propositionality instead of temporality (CollinsVsFodor). Cresswell: this corresponds to Frege’s distinction between idea and thought. FodorVsCollins: Collins is right, but if we believe something, then there is a representation in us that has semantic properties. CresswellVsFodor: Fodor makes use of a confusion of object and content. Belief/relation theory/Fodor/Cresswell: his proof that belief is relational (1981, 178-181) is in fact a proof that "believes" relates a person with a content (not an object). Belief Object/Fodor/Cresswell: Fodor also has other arguments for belief objects. Object/Content/Cresswell: I just want to say that once this distinction has been made, it does not answer the question what the "content" is that objects are described (order/distinction: if A and B are different, a description of A does not help to understand B). II 159 Belief/Collins: (1979, 420): thesis: a belief can be no internal state, because if I want to find out if I believe p, this is indistinguishable from the procedure that I would use to determine p and different from the procedure I would use if I’m in a particular internal state or not. Semantics/stages/McGinn/Cresswell: McGinn (1982) thesis: semantics has several stages. Lately, this thesis has found several followers. Cresswell: this certainly involves a distinction between the object and content. Because then it is about two things: the explanation of truth conditions and the explanation of the role of linguistic thinking in our mental life. |
Fodor IV Jerry Fodor Ernest Lepore Holism. A Shoppers Guide Oxford GB/Cambridge USA 1992 Cr I M. J. Cresswell Semantical Essays (Possible worlds and their rivals) Dordrecht Boston 1988 Cr II M. J. Cresswell Structured Meanings Cambridge Mass. 1984 |
| Compositionality | Fodor Vs Compositionality | IV 64/65 Truth Conditions/tr.cond./holism/Fodor/Lepore: (Fodor/LeporeVsCompositionality as a solution:) "Snow is white", has the truth condition it has, because it belongs to a language that contains "This is snow" and "This is snow", and an indefinite number of other sentences with "is white" and "snow". Semantic Holism/SH: Now, of course, it would be a good argument for semantic holism if only compositionality were really necessary to exclude sentences such as W. Problem: if it is really only because of the structural similarity between "snow is white" and "This is snow" that the former means that snow is white (and not that grass is green), then it would look like an a priori argument against the possibility of non-compositional language, i.e. the expressions of such a language could not have truth conditions! But: Non-Compositional Language/non-recursive/recursive/Fodor/Lepore: e.g. suppose a child has mastered the entire non-recursive apparatus of German. It can say things like: "It’s raining, snow is white, grass is green, that’s snow, that is frozen, everybody hates me, I hate spinach etc.", but not: "Snow is white and grass is green" or "Everyone hates frozen spinach", etc. We assume that the dispositions of the child towards the sentences that it has mastered are exactly the same as those of a normal adult who uses these sentences. It is very plausible that this child, when it says "snow is white", it actually says that snow is white. So far, the compositionality principle of holism is not in danger if we assume that the child has "snow is white" and "this is snow" in its repertoire (idiolect). IV 66 E.g. suppose a second child who uses the unstructured expression "Alfred" instead of "Snow is white". For "This is snow": "Sam", and for "This is cold": "Mary". 1st child: infers from "this is snow" to "this is cold" 2nd child: infers from "Sam" to "Mary". We assume that the translated verbalizations of child two do not differ from the verbalizations of child 1. Nevertheless, if compositionality were a necessary condition for content, then there would be an a priori argument that child 2 could not mean anything specific with his/her statements. Meaning/Vs: what someone means with their statements depends on their intentions! ((s) and not on the sound clusters.) Which a priori argument could show that the child could not make its statement "Sam" with the intention to express that snow is cold? T-sentence: perhaps the T-sentence: "Alfred" is true iff. snow is white is to be preferred over the T-sentence "Alfred" is true iff. grass is green. Important argument: but this cannot be a consequence of the compositional structure of "Alfred", because it has none. It can also be doubted that compositionality is sufficient for the solution of the extensionality problem: IV 178 QuineVsKant/QuineVsAnalyticity/QuineVsCompositionality of Inference: (external): it must be possible for inferences to turn out to be wrong. IV 178/179 VsFodor/Lepore: then one might do with a reformulated CRT (conceptual role theory): the compositional meaning, but not the inferential role is compositional, only within analytical conclusions? Fodor/LeporeVsVs: there is a risk of circularity: if you assume analyticity at all, compositionality, analyticity and meaning spend their lives doing the work of the others. Quine would say: "I told you!" Inferential Role/Fodor/Lepore: the present proposal also threatens their naturalisability ((s) that they are ultimately explained in physiological categories): originally, their attractiveness was to provide a causal role as a basis for the solution of Brentano’s problem of irreducibility to the neurophysiological (>Computation). |
Fodor III Jerry Fodor Jerrold J. Katz The availability of what we say in: Philosophical review, LXXII, 1963, pp.55-71 In Linguistik und Philosophie, G. Grewendorf/G. Meggle Frankfurt/M. 1974/1995 Fodor IV Jerry Fodor Ernest Lepore Holism. A Shoppers Guide Oxford GB/Cambridge USA 1992 |
| Compositionality | Harman Vs Compositionality | Schiffer I 188 Mentalese/HarmanVsCompositionality/Understanding: (Harman 1975, 271): (He was not confused by puns about "use"): even for the understanding of public language compositionality is not needed, provided that thinking takes place in their own public language. He needs two premises: 1) we would only need a compositional semantics if spoken utterances were a matter of decoding language into public Mentalese. 2) But because we think in German, there is no such decoding. Compositionality/Compositional Semantics/Representation/CS Theory/Schiffer: we must also recognize that the compositional semantics theory is not threatened by what I assume about mental representations: (A) Even if we think in Mentalese, there is no true theory of intentionality or of representation as such, which implies that Mentalese has a compositional semantics (B) The truth of (A) does not imply that the public language has no compositional semantics. |
Harman I G. Harman Moral Relativism and Moral Objectivity 1995 Harman II Gilbert Harman "Metaphysical Realism and Moral Relativism: Reflections on Hilary Putnam’s Reason, Truth and History" The Journal of Philosophy, 79 (1982) pp. 568-75 In Theories of Truth, Paul Horwich Aldershot 1994 Schi I St. Schiffer Remnants of Meaning Cambridge 1987 |
| Compositionality | Schiffer Vs Compositionality | I 220 SchifferVsCompositionality: my rejection is based all the time on the rejection of the theory of relations for belief. Here it is difficult to speculate about what kind of conditional sentences for "believes" would require a meaning theory that would not be a truth-theoretic semantics. How could such m.th. look like at all?. E.g. Conceptual Role Semantics: (Schiffer Vs: see section 4.3). Bsp Game Theoretical Semantics/game theory/Hintikka/Schiffer: (Hintikka 1982): this is not an alternative to the conventional theory. PeacockeVsHintikka: (1978) has shown that game theoretical rules provide corresponding truth-theoretical or model theoretical axioms. I XV SchifferVsCompositionality/SchifferVsFrege: natural languages do not have any compositional meaning theories (m.th.). I 137 Paul and Elmer/SchifferVsQuine: Quine: there are no countable belief objects. E.g. if John believes that snow is white, and Mary believes that snow is white, there must be something that both believe. Schiffer: this conditional is false: I 138 Either that or the alleged quantification through belief objects is not what it appears to be the Quine eye. I 144 SchifferVsQuine: harmless apparent quantification. SchifferVsCompositionality: we can now conclude that no natural language has a compositional truth-theoretic semantics (comp.tr.th.Sem.). Otherwise the theory of relations would be correct. In addition, it also has no compositional m.th. because then it has to be a compositional semantics. Understanding/SchifferVsFrege: So compositional semantics are not required to explain speech understanding! I 182 SchifferVsCompositional Semantics: it is false, even regardless of the falsity of the theory of relations of belief. ((s) Compositional Semantics/(s): does not consider the truth conditions but speaks only of the contributions of the meaning of words for the meaning of the proposition.) Schiffer. 1. t is not plausible that languages have a compositional truth-theoretic semantics unless it follows from the stronger assertion that they have compositional truth theories, which themselves are truth-theoretic. (> stronger/weaker; >Strength of Theories). I 192 SchifferVsCompositionality/public language/Mentalese/Schiffer: if I'm right, that no public language has a compositional semantics, I have to find a mistake in (U). It is not my goal to show that speech comprehension does not imply that the natural languages have compositional semantics, the explanation of our understanding would be an empirical task. I rather want to give a counter-E.g. VsCompositionality. E.g. (1) Harvey understands an indefinite number of new propositions of a language E1, which itself contains infinitely many propositions. (2) an explanation of his capabilities does not require compositional semantics. E1: is not a fully-developed natural language. I 193 Harvey: is in this considered possible world an information-processing machine that thinks in machine language: "M": Belief/conviction: Harvey has it if it is in a certain computational relation to an embodied (tokened) proposition of M. ((s) Mentalese: so there is still an internal relation to one's own thought language). B: is a box in Harveys head in which a proposition of M (tokened) exists exactly then when a token from the proposition occurs in B. (Assuming, Harvey has only a finite number of convictions). Belief: for each there is exactly one proposition in Mentalese whose occurrence in B realizes it. µ: is a formula in M so that Harvey believes that snow is white. Realisation/"meaning"/Schiffer: as propositions of M (machine language, Mentalese) realize belief, they also have ipso facto semantic or representational properties. Then it is fair to say that μ "means" that snow is white. And also, that a component of μ references as inner counterpart of the word to snow in the public language. I 195 Speech comprehension/Understanding/Schiffer: without compositionality: E.g. (Continuation: E1: spoken language (without ambiguity and indices) M: Mentalese for Harvey conceptual role: to explain the transition from (1) to (2). (and any others that correspond to it). Propositions in internal code: (or representations thereof: (3) Nemrac derettu "sum"-"sno"-"iz"-"pör-pol" ((s) English backward, [phonetic language], metalanguage (ML) and object language (OL) mixed) (4) Nemrac dias taht emons wons si elprup ((s) English backward, but explicit language, ML) and (5) Nemrac ecnarettu si eurt ffi emos wons si elprup ((s) ML and OL! "true" and "iff" in machine language, but without everyday linguistic meaning or "eurt" does not have to mean "true"! Conceptual role instead of meaning). I 196 Conceptual Role/c.r./SchifferVsCompositionality: we hereby show that "dias taht" and "eurt" can have conceptual roles that a) do not require any compositional semantics, b) explain the transition from one occurrence of (3) in Harveys B-Box to an occurence of (4) and (5) We do not need to specify the full meaning role! I simply assume that (4) and (5) have a role ("whichever"), which by virtue of their formula in Harvey triggers this belief. And none of this makes a compositional semantics necessary: Justification: E.g. you could just have a mapping relation for propositions between two different languages, with which a person who does not understand the other language, knows when a proposition of the other language is true. (…+…) I 200, 202f, 208. |
Schi I St. Schiffer Remnants of Meaning Cambridge 1987 |
| Compositionality | Cartwright Vs Compositionality | Horwich I 57 Relations theory/Belief object/Truth/Easy/Russell/Moore/Cartwright: one question remains: should the doctrine that truth is a simple unanalysable property go together with a relation theory of belief? ((s) Adopted relation to internal objects). I 58 Solution/Cartwright: (ultimately not entirely satisfactory): E.g. the complicated proposition: (12) Brown is taller than Smith and Robinson is taller than Smith. ... + ... CartwrightVsCompositionality: but this is not a case of compositionality in the strict sense. And that applies equally to the above case with propositions. Triadic relation: standing in the K (8) and (9) is a function whose arguments are ordered triples and have a value O(p,q). CartwrightVsRussell: but the fact that (12) is such a function of K, p, and q, does not justify to regard these entities as "components". Russell/Moore/Cartwright: would probably have replied that a proposition is in a certain way an entity or a unit that has parts in an indefinable sense. (Principles Ch. 16) They would have said that it’s one thing for (12) to be a function of K, (8) and (9), but another for these entities to be part of (12). Cartwright: with that they would have been right. But we do not need more than functionality. I 60 You might think functionality is too cheap, there is always some function, and that K, (8) and (9) are linked more intimately in (12) than in a function. But the alternative to strong compositionality is not mere functionality, it is rather the determination of what is required to assert the proposition (12)! Thus, a comparison of (12) is prepared with other propositions. |
Car I N. Cartwright How the laws of physics lie Oxford New York 1983 CartwrightR I R. Cartwright A Neglected Theory of Truth. Philosophical Essays, Cambridge/MA pp. 71-93 In Theories of Truth, Paul Horwich Aldershot 1994 CartwrightR II R. Cartwright Ontology and the theory of meaning Chicago 1954 Horwich I P. Horwich (Ed.) Theories of Truth Aldershot 1994 |
| Davidson, D. | Harman Vs Davidson, D. | Field II 59 Meaning/HarmanVsCompositionality/HarmanVsDavidson: (1075, p.286): Davidson’s theory would be circular if the speaker, who somehow has to represent that snow is white, used the words "snow is white". If that is to be solved with Mentalese, the problem occurs again with the question about the meaning of Mentalese. Solution: not knowing the truth conditions is important for the semantics of representations, but we must attribute the truth conditions, not their knowledge. Field: we also want to be able to distinguish between beliefs about Caesar and those about quarks. Field: but it is a serious question why we should distinguish these contents of representations at all!. For the purposes mentioned above we only need syntax, no semantics of representations. E.g. Suppose a super-simple psychological theory that only seeks to explain the link ">" in representations legally: II 60 E.g. for all sentences S1 and S2 in a system: if a person believes [S1> S2] and wishes S2, then they also want S1. Important argument: although this is explained by a truth value table, we do not need it for our psychological laws (of the super-simple theory). I.e. we do not need to know the meaning of the sentences S1 and S2. ((s) Is this about attribution? Then it is okay. Otherwise it would correspond approximately to: "I wish that the conditions for positive events are realized." And that is too complex and indirect.) Field: however, we must be able to introduce degrees of belief here. The super-simple theory could also include laws such as this: E.g. There is a special class of observation sentences in the representation system, with the property that each of them is connected with a particular type of stimulus. Whenever the stimulus occurs, the organism believes the observation sentence. Important argument: here we do not need to know the meaning of the observation sentence. The psychological theory does not need to assume that the sentence E.g. "there are rabbits in the vicinity" is true. Scientifically nothing is lost if the relation R is assumed to be one between people and meaningless sentences. E.g. in RI: RI: E.g. the native lifts his rifle in the face of the rabbit. This is an overwhelming reason to assume that he beliefs that there are rabbits in the vicinity. Solution: he beliefs a sentence in his language, which in his psychology II 61 plays about the same role as the sentence "there are rabbits in nearby" plays in mine. Semantics/Field: Important argument: is this really a semantic concept? Translation: Is a semantic concept, but a weak one. |
Harman I G. Harman Moral Relativism and Moral Objectivity 1995 Harman II Gilbert Harman "Metaphysical Realism and Moral Relativism: Reflections on Hilary Putnam’s Reason, Truth and History" The Journal of Philosophy, 79 (1982) pp. 568-75 In Theories of Truth, Paul Horwich Aldershot 1994 Field I H. Field Realism, Mathematics and Modality Oxford New York 1989 Field II H. Field Truth and the Absence of Fact Oxford New York 2001 Field III H. Field Science without numbers Princeton New Jersey 1980 Field IV Hartry Field "Realism and Relativism", The Journal of Philosophy, 76 (1982), pp. 553-67 In Theories of Truth, Paul Horwich Aldershot 1994 |
| Frege, G. | Schiffer Vs Frege, G. | I XV Belief/Schiffer: Is no relation to what is believed SchifferVsCompositionality/SchifferVsFrege: Natural languages have no compositional meaning theories (m.th.). propositional attitude/Schiffer: late: thesis: cannot be reduced or explained! "No theory-theory"/Schiffer: all present philosophies of meaning and intentionality are based on false premises. Thesis: there cannot be any meaning theory. I 144 SchifferVsCompositionality: We can now conclude that no natural language has a compositional truth-theoretic semantics. Otherwise the relation theory would be correct. In addition, it also has no compositional meaning theory, because it would have to be a compositional semantics. Understanding/SchifferVsFrege: So compositional semantics are not required to explain speech understanding! Schiffer: so far the arguments are not yet very stable. We still have work to do. I 182 Compositionality/SchifferVsFrege/Problem: Intentional expressions like E.g. "is a picture of" E.g. "true" - adjectives like eg "large" E.g. "toys" (soldier). - E.g. adverbs - evaluative terms such as "should", "good", - E.g. pronouns and demonstrative pronouns - e.g. ordinary language quantifiers such as "everybody", "all", "some". Also counterfactual conditionals and modal expressions contain difficult ontological problems for a compositional semantics. I 183 Solution/Schiffer: Maybe we should give up the idea that there is something to do to give the semantics of these expressions. 3. (most important point): Thesis: Natural languages need no compositional semantics at all. |
Schi I St. Schiffer Remnants of Meaning Cambridge 1987 |
| Frege, G. | Hintikka Vs Frege, G. | Cresswell I 148 Compositionality/Cresswell: It has long been known that it fails on the surface structure. (Cresswell 1973 p 77). HintikkaVsCompositionality/HintikkaVsFrege: H. says that it is simply wrong. In saying that, he ignores the deep structure. And indeed you can regard the difference of the two readings of (39) (Everybody loves somebody) in the context of the game theory as changing the order in the choice of individuals. Then you could say that the only linguistic object is the surface structure. CresswellVsHintikka: but when it comes to that, his observations are not new. Compositionality/Cresswell: fails if we say that the two readings depend on the order in which we first process "and" then "or", or vice versa. Nevertheless, the Frege principle (= compositionality) is in turn applicable to (44) or (45). It is treated like this in Montague. (see below Annex IV: Game-theoretical semantics). I 149 HintikkaVsCompositionality/HintikkaVsFrege: fails even with higher order quantification. CresswellVsHintikka: this is a mistake: firstly, no compositionality is effective in the 1st order translation of sentences like (29). But authors who use higher-order entities (Montague and Cresswell) do not see themselves as deniers of the Frege principle. Hintikka seems to acknowledge that. (1982 p 231). I 161. "is"/Frege/Russell: ambiguous in everyday language. HintikkaVsFrege/KulasVsFrege: (1983): not true! Cresswell: ditto, just that "normal semantics" is not obliged to Frege-Russell anyway. Hintikka II 45 (A) Knowledge/Knowledge Objects/Frege/Hintikka: His concern was what objects we have to assume in order to understand the logical behavior of the language, when it comes to knowledge. Solution/Frege/Hintikka: (see below: Frege’s knowledge objects are the Fregean senses, reified, >intensional objects). Hintikka: For me, it is primarily about the individuals of which we speak in epistemic contexts; only secondarily, I wonder if we may call them "knowledge objects". Possible Worlds Semantics/HintikkaVsFrege: we can oppose the possible worlds semantics to his approach. (Hintikka pro possible worlds semantics). II 46 Idea: application of knowledge leads to the elimination of possible worlds (alternatives). Possible World/Hintikka: the term is misleading, because too global. Def Scenario/Hintikka: everything that is compatible with the knowledge of a knower. We can also call them knowledge worlds. Set of All Possible Worlds/Hintikka: we can call it illegitimate. (FN 5). Knowledge Object/Hintikka: can be objects, people, artifacts, etc. Reference/Frege/Hintikka: Frege presumes a completely referential language. I.e. all our expressions stand for some kind of entities. They can be taken as Fregean knowledge objects. Identity/Substitutability/SI/Terminology/Frege/Hintikka: SI is the thesis of the substitutability of identity ((s) only applies with limitation in intensional (opaque) contexts). II 47 E.g. (1) ... Ramses knew that the morning star = the morning star From this it cannot be concluded that Ramses knew that the morning star = the evening star (although MS = ES). II 48 Context/Frege/Hintikka: Frege distinguish two types of context: Direct Context/Frege/Hintikka: extensional, transparent Indirect Context/Frege/Hintikka: intensional, opaque. E.g. contexts with "believes" (belief contexts). ((s) Terminology: "ext", "opaque", etc. not from Frege). Frege/Hintikka: according to his own image: (4) expression >sense >reference. ((s) I.e. according to Frege the intension determines the extension.) Intensional Contexts/Frege/Hintikka: here, the picture is modified: (5) Expression (>) sense (> reference) Def Systematic Ambiguity/Frege/Hintikka: all our expressions are systematically ambiguous, i.e. they refer to different things, depending on whether they are direct (transparent, extensional) contexts or indirect ones (intensional, opaque). Fregean Sense/Hintikka: Fregean senses in Frege are separate entities in order to be able to work at all as references in intensional contexts. E.g. in order to be able to restore the inference in the example above (morning star/evening start) we do not need the identity of morning star and evening star, but the. identity of the Fregean sense of "morning star" and "evening star". II 49 Important argument: but Frege himself does not reinterpret the identity in the expression morning star = evening star in this way. He cannot express this fact, because there identity occurs in an extensional context and later in an intensional context. Identity/Frege/Hintikka: therefore we cannot say that Frege reinterprets our normal concept of identity. Problem: It is not even clear whether Frege can express the identity of the senses with an explicit sentence. For in his own formal language (in "Begriffsschrift"(1) and "Grundgesetze"(2)) there is no sentence that could do this. He says that himself in: "Über Sinn und Bedeutung": we can only refer to the meanings of our expressions by prefixing the prefix "the meaning of". But he never uses this himself. (B) Knowledge Objects/Possible World Approach/HintikkaVsFrege: Idea: knowledge leads us to create an intentional context that forces us to consider certain possibilities. These we call possible worlds. new: we do not consider new entities (intensional entities) in addition to the references, but we look at the same references in different possible worlds. Morning Star/Evening Star/Possible Worlds Semantics/Hintikka: Solution: "morning star" and "evening star" now single out the same object, namely the planet in the real world. II 50 (C) Possible Worlds Semantics/HintikkaVsFrege: there is no systematic ambiguity here, i.e. the expressions mean the same thing intensionally as extensionally. E.g. Knowing what John knows means knowing those possible worlds which are compatible with his belief, and knowing which are not. II 51 Extra premise: for that it must be sure that an expression singles out the same individual in different possible worlds. Context: what the relevant possible worlds are depends on the context. E.g. Ramses: here, the case is clear, On the other hand: E.g. Herzl knew Loris is a great poet Additional premise: Loris = Hofmannsthal. II 53 Meaning Function/Possible Worlds Semantics/Hintikka: the difference in my approach to that of Frege is that I consider problems locally, while Frege considers them globally. Fregean Sense/(= way of givenness) Hintikka: must be considered as defined for all possible worlds. On the other hand: Hintikka: if Fregean sense is construed as meaning function, it must be regarded as only defined for the relevant alternatives in my approach. Frege: precisely uses the concept of identity of senses implicitly. And as meaning function, identity is only given if the mathematical function works for all relevant arguments. Totality/Hintikka: this concept of totality of all logically possible worlds is now highly doubtful. Solution/Hintikka: it is precisely the possible worlds semantics that helps dispense with the totality of all possible worlds. ((s) And to consider only the relevant alternatives defined by the context). Fregean Sense/Hintikka: was virtually constructed as an object (attitude object propositional object, thought object, belief object). This is because they were assumed as entities in the real world (actual world), however abstract. II 54 Meaning Function/M. F./HintikkaVsFrege/Hintikka: unlike Fregean senses, meaning functions are neither here nor elsewhere. Problem/Hintikka: Frege was tempted to reify his "senses". Knowledge Object/Thought Object/Frege/Hintikka: Frege, unlike E.g. Quine, has never considered the problem. Existential Generalization/EG/Hintikka: entitles us to move from a sentence S(b) with a singular term "b" to the existential statement (Ex) S(x). This fails in intensional (epistemic) contexts. Transition from "any" to "some". E.g. epistemic context: (10) (premise) George IV knew that (w = w) (11) (tentative conclusion) (Ex) George IV knew that (w = x) II 55 Problem: the transition from (10) to (11) fails, because (11) has the strength of (12) (12) George IV knew who w is. EG/Fail/Solution/Frege/Hintikka: Frege assumed that in intensional (opaque) contexts we are dealing with ideas of references. HintikkaVsFrege: Problem: then (11) would follow from (10) in any case ((s) and that’s just what is not desired). Because you’d have to assume that there is definitely some kind of sense under which George IV imagines an individual w. Problem: "w" singles out different individuals in different possible worlds. II 56 Possible Worlds Semantics/Solution/Hintikka: E.g. Suppose. (13) George knows that S(w) to (14) (Ex) George knows that S(x) where S(w) does not contain expressions that create opaque contexts. Then we need an additional condition. (15) (Ex) in all relevant possible worlds (w = x). This is, however, not a well-formed expression in our notation. We have to say what the relevant possible worlds are. Def Relevant Possible Worlds/Hintikka: are all those that are compatible with the knowledge of George. Thus, (15) is equivalent to (16) (Ex) George knows that (w = x). This is the additional premise. I.e. George knows who w is. (Knowing that, knowing who, knowing what). Knowing What/Logical Form/Hintikka/(s): corresponds to "knows that (x = y)" ((s) >single class, single quantity). E.g. knowing that "so and so has done it" does not help to know who it was, unless you know who so and so is. ((s) i.e. however, that you know y!) Solution/Hintikka/(s): the set of possible worlds compatible with the knowledge) II 57 Meaning Function/M. F./Possible Worlds Semantics/Hintikka: in order to be a solution here, the meaning function (see above) needs to be a constant function, i.e. it must single out the same individuals in all possible worlds. Frege/Identity/Opaque Context/Hintikka: Frege had to deal with the failure of the SI (substitutability in case of identity) ((s) i.e. the individuals might have a different name), not with the failure of the Existential Generalization (EG). ((s) I.e. the individuals might not exist). Hintikka: therefore, we need several additional premises. Possible Worlds Semantics: SI: here, for substitutability in case of identity, we only need on the assumption that the references of two different concepts in any possible world can be compared. Existential Generalization: here we have to compare the reference of one and the same concept in all possible worlds. Frege/Hintikka: now it seems that Frege could still be defended yet in a different way: namely, that we now quantify on world-lines (as entities). ((s) that would accomodate Frege’s Platonism). II 58 World Lines/Hintikka: are therefore somehow "real"! So are they not somehow like the "Fregean senses"?. HintikkaVs: it is not about a contrast between world bound individuals and world lines as individuals. World Lines/Hintikka: but we should not say that the world lines are something that is "neither here nor there". Using world lines does not mean reifying them. Solution/Hintikka: we need world-lines, because without them it would not even make sense to ask at all, whether a resident of a possible world is the same one as that of another possible world. ((s) cross world identity). II 59 World Line/Hintikka: we use it instead of Frege’s "way of givenness". HintikkaVsFrege: his error was to reify the "ways of givenness" as "sense". They are not something that exists in the actual world. Quantification/Hintikka: therefore, in this context we need not ask "about what we quantify". II 109 Frege Principle/FP/Compositionality/Hintikka: if we proceed from the outside inwards, we can allow a violation of Frege’s principle. (I.e. the semantic roles of the constituents in the interior are context dependent). II 110 HintikkaVsFrege/HintikkaVsCompositionality: Thesis: meaning entities should not be created step by step from simpler ones in tandem with syntactic rules. They should instead be understood, at least in some cases, as rules of semantic analysis. 1. G. Frege, Begriffsschrift, eine der arithmetischen nachgebildete Formelsprache des reinen Denkens, Halle 1879, Neudruck in: Ders. Begriffsschrift und andere Aufsätze, hrsg. v. J. Agnelli, Hildesheim 1964 2. Gottlob Frege [1893–1903]: Grundgesetze der Arithmetik. Jena: Hermann Pohle Wittgenstein I 71 Def Existence/Wittgenstein: predicate of higher order and is articulated only by the existence quantifier. (Frege ditto). I 72 Hintikka: many philosophers believe that this was only a technical implementation of the earlier idea that existence is not a predicate. HintikkaVsFrege: the inexpressibility of individual existence in Frege is one of the weakest points, however. You can even get by without the Fregean condition on a purely logical level. HintikkaVsFrege: contradiction in Frege: violates the principle of expressing existence solely through the quantifier, because the thesis of inexpressibility means that through any authorized individual constant existential assumptions are introduced in the logical language. |
Hintikka I Jaakko Hintikka Merrill B. Hintikka Investigating Wittgenstein German Edition: Untersuchungen zu Wittgenstein Frankfurt 1996 Hintikka II Jaakko Hintikka Merrill B. Hintikka The Logic of Epistemology and the Epistemology of Logic Dordrecht 1989 Cr I M. J. Cresswell Semantical Essays (Possible worlds and their rivals) Dordrecht Boston 1988 Cr II M. J. Cresswell Structured Meanings Cambridge Mass. 1984 W II L. Wittgenstein Wittgenstein’s Lectures 1930-32, from the notes of John King and Desmond Lee, Oxford 1980 German Edition: Vorlesungen 1930-35 Frankfurt 1989 W III L. Wittgenstein The Blue and Brown Books (BB), Oxford 1958 German Edition: Das Blaue Buch - Eine Philosophische Betrachtung Frankfurt 1984 W IV L. Wittgenstein Tractatus Logico-Philosophicus (TLP), 1922, C.K. Ogden (trans.), London: Routledge & Kegan Paul. Originally published as “Logisch-Philosophische Abhandlung”, in Annalen der Naturphilosophische, XIV (3/4), 1921. German Edition: Tractatus logico-philosophicus Frankfurt/M 1960 |
| Soames, S. | Schiffer Vs Soames, S. | I 217 Compositional Semantics/Comp.sem./Understanding/Explanation/Scott Soames/Schiffer: (Soames 1987) Thesis: comp.sem. is not needed for explaining the language understanding, nevertheless natural languages have a comp.sem .: Language understanding/Soames: you should not look at the semantics to explain semantic competence. Instead one needs comp.sem you. for the explanation of the representational character of the language. The central semantic fact about language is that it is needed to represent the world. Propositions encode systematic information that characterize the world so and so. We need comp.sem. for the analysis of the principles of this encoding. SchifferVsSoames: Instead, I have introduced the expression potential. One might assume that a finally formulated theory should be able to formulate theorems for the attribution of expression potential to each proposition of the language. But would that then not be a compositional theory?. I 218 E.g. Harvey: here we did not need comp.sem. to assume that for each proposition of M (internal language) there is a realization of belief, that means (µ)(∑P)(If μ is a proposition of M and in the box, then Harvey believes that P). (s) Although here no connection between μ and P is specified). Schiffer: Now we could find a picture of formulas of M into German, which is a translation. But that provides no finite theory which would provide a theorem for every formula μ of M as If μ is in the box, then Harvey thinks that snow is sometimes purple. Propositional attitude/Meaning theory/Schiffer: Problem: it is not possible to find a finite theory which ascribes verbs for belief characteristics of this type. Pointe: yet the terms in M have meaning! E.g. "Nemrac seveileb taht emos wons si elprup" would realize the corresponding belief in Harvey and thus also mean trivially. SchifferVsCompositionality: when the word-meaning contributs to the proposition-meaning, then it is this. Then expressions in M have meaning. But these are not characteristics that can be attributed to a finite theory. We could find only the property to attribute to each proposition of M a particular belief, but that cannot happen in a finite theory. mental representation/Mentalese/Schiffer: the formulas in M are mental representations. They represent external conditions. Propositions of E, Harvey's spoken speech, received their representational character via the connection with mental representations. Therefore Mentalese needs no comp.sem. SchifferVsSoames: So he is wrong and we need the comp.sem. not even for an illustration of how our propositions represent the world. I 219 We had already achieved this result via the expresion potentials. Because: representational character: is indistinguishable from the expression potential. |
Schi I St. Schiffer Remnants of Meaning Cambridge 1987 |
| |||