Abstract
In a recent, important series of papers,1Prof. Donald Davidson has proposed and illustrated a theory about theories of meaning. The theory, to put it very much more crudely than Davidson does, is simply that a theory of meaning for the (natural) language L ought to take the form of a truth definition for L. That is, such a theory ought to recursively associate each truth-valuable sentence of L with a representation of its truth conditions. Davidson says “what we require of a theory of meaning for a language L is that without appeal to any (further) semantical notions it place enough restrictions on the predicate ‘is T’ to entail all sentences got from schema [1] when ‘s’is replaced by a structural description of a sentence of L and ‘p’by that sentence.” (T & M 309)
Thanks are due, inter alia, to Janet Dean Fodor, Prof. Judith Thomson, Prof. Helen Cartwright and Prof. Richard Cartwright, all of whom helped with the hard bits. But none of the preceding is guilty of any of the following.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
Similar content being viewed by others
References
I shall be referring to ‘T & M’ and ‘LFAS’. The former is cryptic for Donald Davidson, ‘Truth and Meaning’, Synthese 17(1967) 304–323;
the latter for Donald Davidson, ‘The Logical Form of Action Sentences’ in The Logic of Decision and Action(ed. by N. Rescher), University of Pittsburgh Press, Pittsburgh, Pa., 1967, pp. 81–95.
It should be mentioned, however, that there are plenty of troubles with this (or any other known) view of anaphora. See, inter alia,Chapter III of David Reibel and Sanford Schane (eds.), Modern Studies in English,Prentice Hall, Englewood Cliffs, N.J., 1969. For discussion of proforms for verbs,
see J. A. Fodor, ‘Three Reasons for Not Deriving “Kill” from “Cause to Die”’, Linguistic Inquiry 1(1970).
Davidson explicitly excludes adverbs like ‘clearly’ from his discussion on two grounds. First, they function similarly to comparatives (John spoke clearly for a man with his mouth full of marbles, but unclearly for a Yale man) and the logical form of comparatives is, in general, unknown territory. Second, unlike such adverbs as ‘at twelve noon’ or ‘in New York’, ‘clearly’ introduces no “new entity” into sentences like (8/9). However, neither of these peculiarities of ‘clearly’ is the one on which our discussion of (8/9) will turn; and we will see presently that some sentences which have neither of these properties are nevertheless counterexamples to Davidson’s analysis.
In fact, I think the right account is that (10) is ambiguous in the same way as (27) below.
This point has been widely recognized in the linguistic literature. For an extended analysis, see Noam Chomsky, Aspects of the Theory of Syntax,M.I.T. Press, Cambridge, Mass., 1965.
‘S’ for ‘sentence’, ‘NP’ for ‘noun phrase’, ‘V for ‘verb’, ‘VP’ for ‘verb phrase’, and ‘D’, indiscriminately, for ‘adverb’ or ‘adverbial phrase’.
That this case is not isolated, and not restricted to verbs which govern opaque contexts, can be seen from the existence of rather subtle ambiguities like ‘John left the train at the station’. On the most obvious reading (where ‘left’ has the force of ‘disembarked from’ the Davidsonian paraphrase is the natural one since the presumed deep structure is (at the station (John left the train)), i.e. ‘at the station’ functions as a sentence modifier. But consider the alternative reading where the train is a toy train and John left it, say, at the station baggage counter. Here ‘John left the train at the station’ does notentail ‘John disembarked from the train’, nor can we paraphrase this reading by ‘At the station, John left the train’ (cf. the patent ungrammaticality of ‘At the station, John left his umbrella’). On this second reading, the presumed deep structure is (John ((left) at the station) the train) or, possibly, John ((left the train) at the station); in any event, a constituent modifier rather than a sentence modifier seems to be at issue. The point is, of course, that the two readings are not equivalent, so that (∃x) (leave (John, the train, x)) & At (the station, x) can represent at most one of them.
There may be some temptation to suppose that ‘leave’ is lexically ambiguous, but I think the temptation should be resisted. ‘John left his friend at the station’ has the same ambiguity as ‘John left the train at the station’, but here neither reading can be captured by rendering ‘left’ as ‘disembarked from’. Indeed, in this case the relevant difference between (at the station (John left his friend)) and (John ((left) at the station) his friend) is just that the former entails that both John and his friend were at the station when they parted, while the latter would be true if, say, John drove his friend to the station and then drove off.
It should be added, however, that the problem of modeling the presuppositions of negative sentences in a language which contains no constituent negation is not quiteso worrying as the problems about adjectival and adverbial constituent modification. This is because, if we assume the mechanisms of set theory, and in particular a complement operator, we can get something of the force of constituent negation. At, by the way, a corresponding sacrifice of naturalness in the mapping from English into the canonical language. (It is important to remember that this mapping is, ex hypothesisto be accomplished by a mechanicalprocedure. It is thus a rule of thumb that the more structural dis-similarity we permit between natural language expressions and their meta-linguistic translations, the more trouble we are likely to have in characterizing the mechanism that performs the translation). For discussion of the relations between constituent negation and presupposition, see Noam Chomsky, ‘Deep Structure, Surface Structure, and Semantic Interpretation’ in Semantics: An Interdisciplinary Reader in Philosophy, Linguistics, Anthropology and Psychology(ed. by Leon Jakobovits and Danny Steinberg), Cambridge University Press, Cambridge 1971.
It is worth mentioning the existence of a reading of (7) which attributes to John not a performancebut, roughly, a trait.This is, in fact, a kind of ambiguity that is exhibited quite pervasively by past tense action sentences. (Cf. ‘John was greedy’, ‘John was self-sacrificing’, ‘John spoke Latin’, etc.) I have ignored this reading of sentences like (7) because there is not a prayer of capturing it with Davidsonian paraphrases. In particular, ‘there are one or more events which consist of John’s speaking clearly’ surely does not entail ‘John (chronically) speaks clearly’ though it is, perhaps, entailed by it.
Jerrold J. Katz and Jerry A. Fodor, ‘The Structure of a Semantic Theory’, Language 39(1963) 170–210.
One price they have paid for this assumption is an unresolved problem about how to draw the distinction between inferences mediated by structure (for which the assumption might be true) and inferences mediated by (lexical) content (for which it quite certainly is not).
It is worth suggesting how this sort of approach might hope to handle entailment relations which turn upon the presence of adverbial phrases. Clearly, we shall need both rules which introduce adverbial phrases and rules for their elimination (i.e., rules which, respectively, determine when a sentence which does not contain a certain adverbial entails a sentence which does; and rules which determine when a sentence which does contain a certain adverbial entails a sentence which does not). The general character of the latter sort of rule is fairly clear. Roughly, given a sentence Swhose deep structure contains a verb phrase with its madverbial modifiers, this sentence will entail any sentence S′ which is identical to Sexcept for the deletion of the m–1thadverbial phrase. (I assume that the adverbs are ordered in deep structure and that the order is left-to-right by increasing scope. If this assumption is false, minor modifications will have to be made in the rule.) This rule explains why ‘John spoke clearly’ entails ‘John spoke’, etc. On the other hand, it does notpermit the (invalid) inference from ‘Probably John spoke’ to ‘John spoke’ since it does not apply to sentence operators. Two remarks may be made about this rule for ‘peeling off’ adverbial phrases. First, it applies to syntactically analyzedsentences (presumably, to deep structures). For example, the rule must ‘know’ that ‘except Sunday’ is not an independent adverbial in ‘John and Mary make love every day except Sunday’; otherwise it will allow us to infer from that sentence to ‘John and Mary make love every day’. I am assuming that the aspects of structural descriptions that may be mentioned in characterizing the domain of such rules may be quite abstract and are, in fact, precisely the sorts of syntactic properties of sentences that are investigated in modern grammars. Second, it is worth noting that the rule for peeling off adverbial phrases is strikingly analogous to the rule of simplification which allows us to infer P, Qfrom P & Q.(For example, both rules fail in the scope of opaque verbs, in the scope of negation, etc.) The primary difference is that since the peeling off rule operates on phrases, it allows us to capture the entailments of sentences containing adverbials without first translating them into conjunctions à la Davidson. The general character of the rule for shortening sentences by peeling off adverbial phrases seems to be fairly clear. What is more complicated is the rule which allows us to lengthen sentences by adding such phrases. Roughly, the following rule of addition would seem plausible. Assume, as Davidson does, that verbs have fixed numbers of argument places and that a verb may appear in well-formed sentences in which some of its places are unfilled. Suppose, in particular, that we have a verb Vof α places appearing in a sentence Swhere only ß < αof its places are filled. Then the rule of addition allows us to infer any well-formed sentence S′ which differs from Sonly by the addition of a ‘dummy’ phrase (‘sometime’, ‘somewhere’, ‘in some manner’, etc.) in the position of the appropriate unfilled place of V.Thus, the rule of addition lets us infer from ‘John bit Mary’ to ‘John bit Mary sometime’, ‘John bit Mary somewhere’ (in both senses), etc. Two points about the rule of addition. First, it is, as stated, grossly informal. But formalizing it would be quite straightforward given: a canonical form for representing base structures of sentences containing verbs with their argument positions, an enumeration of the lexical sequences that are to count as dummy phrases, and an enumeration of the well-formed sentences. There would appear to be no principled reason for believing that these conditions cannot be met. More serious is the following worry. The rule of addition assumes that certain adverbials (together, for that matter, with direct objects, indirect objects, etc.) appear as fillers for the argument positions of verbs, and this returns us to the problem of how we are to determine which and how many arguments a given verb should be assigned. For example, we presumably want the theory to allow us to infer from ‘John pointed the stick’ to ‘John pointed the stick at something’ but we do not want it to let us infer from ‘John waved the stick’ to ‘John waved the stick at something’. Clearly, the rule of addition will block the latter inference while permitting the former only if we assume that the phrase for what is pointed at occupies an argument position of ‘point’ but the phrase for what is waved at does not occupy an argument position of ‘wave’. This amounts to saying that the criterion for whether a phrase should be treated as occupying an argument position of a verb is simply whether the rule of addition applies. In particular, assume that an adverbial phrase occupies an argument position of the verb Vif and only if a sentence whose verb phrase contains just Ventails a sentence which differs only in that its verb phrase contains a dummy for the adverbial as well.
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 1972 D. Reidel Publishing Company, Dordrecht-Holland
About this chapter
Cite this chapter
Fodor, J.A. (1972). Troubles about Actions. In: Davidson, D., Harman, G. (eds) Semantics of Natural Language. Synthese Library, vol 40. Springer, Dordrecht. https://doi.org/10.1007/978-94-010-2557-7_3
Download citation
DOI: https://doi.org/10.1007/978-94-010-2557-7_3
Publisher Name: Springer, Dordrecht
Print ISBN: 978-90-277-0310-1
Online ISBN: 978-94-010-2557-7
eBook Packages: Springer Book Archive