24. I will ignore the subtle difference between "language" and "parole" in Saussure. Baskin Wade translates the former with "speech" and the latter with "speaking". See Saussure (1916), pp. 9 and 13. In my discussion, I will continue just to speak colloquially of speech, but I usually will be talking about Saussure's "parole".
25. All page references in this section are to Saussure (1916).
26. See pp. 10 and 14 which I quote below, but also throughout much of the Course, where this talk of language as conventional is maintained. But see below for why he takes the sense in which language is conventional to be limited.
27. One might have expected that some of the sources of linguistic change and evolution should have been found in the use of metaphor or other non-literal usage. But in his discussion of linguistic change (in Part Three of the Course), Saussure doesn't even mention these. See next footnote.
28. Phonology for Saussure is about the "physiology of sounds" while phonetics is "the study of the evolution of sounds" [p. 33]. "Phonetics" Saussure tells us, "- and all of phonetics - is the prime object of diachronic linguistics" [p. 140]. He does suggest that there are meaning phenomena in diachronic linguistics as well: "...sounds are not the only things that change with time. Words change their signification...." [p. 141]. But his considerations of issues concerning meaning in his discussion of diachronic linguists (Part Three of the Course) are minimal.
29. Saussure attributes something like this view to "neogrammarian" German linguists of the last quarter of the 19th century (who included Hermann Paul): "The new school...fought the terminology of the comparative school, and especially the illogical metaphors that it used. One no longer dared to say, 'Language does this or that,' or 'life of language,' etc. since language is not an entity and exists only within speakers" [p. 5, footnote].
30. See pp. 98-100 for some general comments on how Saussure thinks of language change, and also the third part of the book called "Diachronic Linguistics" in which he discusses what he takes to be the mechanisms of language change.
31. It is interesting to compare Saussure with Welby here. For Welby, as we saw, also respected greatly the apparent mutability of language and its source in the individual speech behaviors of the linguistic community. But, since she had such a strong aversion to the notion of central meanings, a notion like that of Saussure's language could never have been at issue for her: a Saussurean language essentially pairs sound-images with meanings, that is, central meanings. But Saussure must struggle with a language that is stable enough to study empirically, and yet founded somehow on the constant flow of everyday speech.
32. We might look at his talk of convention as a limited metaphor and note that Saussure also defends the use of metaphor in linguistic studies right after the comment quoted in footnote 33 above apparently praising the neogrammarian condemnation of certain metaphors: "One must not go too far, however, and a compromise is in order. Certain metaphors are indispensable. To require that only words that correspond to the facts of speech be used is to pretend that these facts no longer perplex us. This is by no means true, and in some instances I shall not hesitate to use one of the expressions condemned at that time" [p. 5, footnote].
33. See Saussure's discussion of and arguments about the relation between language and other sorts of social institutions, Saussure (1916), pp. 72-74.
34. See the treatment given to Gardiner and his influence by Nerlich in Nerlich (1990) and Nerlich (1992). But with respect to Gardiner's influence on Oxford philosophy Nerlich says nothing, her interest being primarily in the history of the discipline of linguistics and not the history of ideas more broadly.
35. Unless otherwise indicated, all page number references in this section are to Gardiner (1951).
36. As with most of the themes in The Theory of Speech and Language, this point can be found in earlier publications of Gardiner. See Gardiner (1919), p. 5: "Language...is a mechanism for the communication of thought, not a duplicate, or kind of soul in it."
37. Gardiner acknowledges Saussure's influence to a limited extent in The Theory of Speech and Language, but the extent of the influence seems much greater in many respects than what Gardiner acknowledges there. And in later discussions he gives more credit to Saussure. Gardiner (1933) begins: "To Ferdinand de Saussure belongs the merit of having drawn attention to the distinction between
speech
and language , a
distinction so far-reaching in its consequences that in my opinion it can hardly fail, sooner
or later, to become the indispensable basis of all scientific treatment of grammar." In
Gardiner (1937), Gardiner gives the following account of Saussure's influence on him:
"I read de Saussure's now famous Cours de linguistique shortly after its first appearance,
and at that time found it extraordinarily obscure. It was not until my own volume was
out that I studied de Saussure again consecutively, and only then did I discover how
much, though quite subconsciously, I really owed to him. It is a matter of justice to a
great thinker that I should now openly admit my indebtedness, and express the regret that
I did not earlier become aware of it." Gardiner throughout his work is generally not at
all ungenerous in citing influences, especially with respect to the linguist Philipp Wegener
(to whom The Theory of Speech and Language is dedicated), and so I find this account
of Gardiner's wholly believable. It would be wrong to suppose, in addition, that Gardiner
did not do a lot of creative and wholly independent work based on the distinction between
speech and language.
38. It may look from this passage as though Gardiner disagrees with Saussure in taking a sign to be a sound which has a meaning rather than a pairing of a "sound-image" and a meaning. But the difference is only apparent. In 25 of The Theory of Speech and Language Gardiner argues that words cannot be understood to be sounds and claims that they are instead "psychical entities" (Gardiner (1951), pp. 69-71). And at one point he uses the term "sound-image": "The articulate sounds appear to be physical, audible, copies of one aspect of their psychical originals. It is only the sound-image connected with words which can be reproduced in a physical copy" (Gardiner (1951), p. 70). I am chiefly concerned with issues of meaning however, and I will generally ignore these sorts of questions, however interesting.
39. This idea is anticipated by Stout: "Let any one pick out at random a sentence from a book and then let him look for another exactly like it; it is a hundred to one that he will succeed in finding one, even by a long and diligent search. Sentences, indeed, are, for the most part, composed of mere elementary conjunctions of words which have become familiar through custom. But the point of a sentence which interests and attracts attention usually finds expression in the relative novelty which belongs to it as a whole" (Stout (1893), p. 123). Something like this idea is picked up on later by Ryle when he argues in Ryle (1953) against the idea that each meaningful sentence must have a unique role (or use). (See Landesman (1972), pp. 19 ff. on Ryle on this point.) And, of course, Chomsky makes much of the novelty of most sentences in various of his works.
40. Gardiner uses the expression "act of speech" frequently throughout his book in ways that seem to anticipate the use of the term "speech act" associated with Austin. I will discuss Gardiner's influence on Austin below.
41. Gardiner, in a footnote to this passage, cites Erdmann as the source of this distinction between word-meanings and 'ideas'.
42. I do not know whether Gardiner knew anything of Welby's works directly. He does refer to significs (Gardiner (1951), p. 85) and he equates it with "semasiology" and "semantics". Given that, as we saw, Gardiner sees his primary concern as "semasiological" (Gardiner (1951), p. 13), there is at least this much evidence, then, that he saw himself as doing just the work that Welby called for. Other than this, I have no evidence of any influence of Welby on Gardiner.
43. I will discuss Gardiner's understanding of what today we call speech-acts below, but it is probably good to note here that Gardiner saw speech-act notions as central to his theory, so "communicative intention" here includes for him not just intentions to say things, but to command things, request things, etc. as well. In the present context, however, I am focussing not on the speech-act that needs to be discovered by a listener, but the content of the speech-act.
44. See my chapter 4 below for a discussion of compositional-meaning theories.
45. The last sentences of this passage in connecting specific acts of speech (questioning and commanding) to responses from the listener might call to mind Austin's notion of securing of uptake being essential to speech acts (see, of course, Austin (1962) and Strawson (1964)). I will mention this connection again below.
46. Gardiner cites Hermann Paul in his section on metaphor. He also mentions at one point Paul's distinction between occasional and usual meaning which it was noted above (note 6) Stout cited and Welby quoted: "...[Paul] distinguishes between usuelle Bedeutung, the generally accepted meaning of a word, and okkasionelle Bedeutung, the meaning which a speaker attaches to a word at the moment of utterance..." [p. 58].
47. Interestingly, Gardiner does not cite Saussure at all in this section of his book, but he does cite the linguist Hermann Paul. Paul also, it will be recalled, had some influence on Welby and on Saussure. See above, footnotes 6 and 29.
48. It should be noted that Gardiner often enough says things that seem to contradict this claim. For example, he says at one point quite plainly that "every act of speech is the purposeful performance of a 'speaker' employing 'words' in order to draw the attention of a 'listener' to some 'thing'" [p. 187]. But I think that Gardiner's view is such that he would allow non-conventional signs to be words sometimes. He speaks notoriously little in terms of conventions of language even though he stresses the necessity of more than one person in order for there to be language. Because of this, his notion of language sometimes gets applied to pre-conventional speech situations. But I think that his overall intention was to hold that language is, as in the quote in the text above, the result of repetition of what were originally language-less speech episodes. I won't try to iron out this matter more nicely here. Gardiner's interests were not primarily with working out the details of the metaphysics here and so, as with Saussure, it can be expected that he will have some metaphysically puzzling threads hanging out at the edges of his thinking.
49. Gardiner makes it clear throughout his work that his attitude on this score owes much to the linguist Philipp Wegener. He provides the following quote from Wegener: "the purpose of our speech is always to influence the will or the perception of someone in a way which the speaker considers to be of importance" [p. 237]. He also discusses a number of other earlier authors on this score; see p. 182 and pp. 237-239.
50. Austin's gloss of the truth of a statement in his famous paper "Truth" is reminiscent of the present definition of a statement by Gardiner: "A statement is said to be true when the historic state of affairs to which it is correlated by the demonstrative conventions (the one to which it 'refers') is of a type with which the sentence in making it is correlated by the descriptive conventions" (Austin (1950), p. 122). Strawson in his famous criticism of Austin (1950) notes that the thesis "that all statements involve both demonstration and description is, roughly, the thesis that all statements are, or involve, subject-predicate statements..." (Strawson (1950), p. 193) and goes on to find fault with Austin for holding this thesis (Strawson (1950), p. 211). Strawson's characterization of Austin's view nicely emphasizes the features of the latter that are like Gardiner's definition of a statement: Austin's characterization of truth merely repeats Gardiner's characterization of a statement.
51. "Special sentence-quality" is a technical term for Gardiner: "...the utterance itself is called a request, question, statement, or exclamation, and these descriptions constitute its special sentence-quality" [p. 197].
52. On Grice on recognition of intention, see my first chapter above. On Strawson on recognizing intention and securing of uptake, see Strawson (1964).
53. There are a number of passages in Gardiner's book that illustrate this anticipation of the notions of intention of recognition of intention and of securing of uptake. Another is the following: "Now the speaker may have had the best of intentions, but the listener may nevertheless fail to understand what was meant. In that event the sentence does not function, and its special quality has been in vain. The act of speech desiderates an intelligent act of understanding as its counterpart, and this, however much mechanized, is always a deduction from both the words and the situation" [p. 198].
54. This distinction does have things in common with Grice's distinction between what is said by an utterance and what is implicated by it, though I cannot pursue the matter here. But see pp. 202 ff., 213, 223, 299, 306, and 330 where this distinction is discussed or illustrated. See especially pp. 230 ff. for an interesting example by Gardiner.
55. Urmson in his paper "Performative Utterances" (Urmson (1979)) discusses a little bit of how and when Austin developed his thoughts on performative utterances, but he mentions no historical influences. Still, as mentioned above, there is evidence that Austin knew of Gardiner's work and there might have been some influence here. I cannot investigate the matter more fully at this time unfortunately. See below in the text when I discuss Furberg's observations concerning the influence of Gardiner on Austin.
56. His influence in Egyptology is still widely felt today, I understand, and it has been argued that he has had an important influence on the discipline of linguistics. See Nerlich (1990) and Nerlich (1992).
57. Ryle (1961), in Parkinson (1968), p. 109.
58. Furberg (1971), p. 57 and above, note 48.
59. Wright (1976), pp. 513-514. The work by George Campbell that Wright cites here, The Philosophy of Rhetoric, was originally published in 1776, so it was over 150 years prior to Gardiner's work (Wright cites an 1823 edition; but see Campbell (1988)). The suggestion that hooks Gardiner and thereby Grice to the tradition of rhetorical studies is perhaps a historically interesting one. Gardiner himself speaks of the "rhetorical intentions" of the speaker a number of times in his book. In a very early paper of his, "Some Thoughts on the Subject of Language", he says the following: "Besides the content of the statement, command, question, wish, negation, maxim, epigram, or curse that is expressed, there is always involved an unexpressed relation to a listener, which by a curious paradox, is in practice the decisive factor in determining what words are actually used and the order in which they are arranged". He then asks in a footnote, "May it not, then, be said that what is now most required for the progress of linguistic science is the study of comparative Rhetoric?" (Gardiner (1919), p. 6). The connection between the tradition of rhetorical studies, which, of course, goes back to Aristotle and earlier, and the tradition that I see as flowing through Welby and Gardiner to Grice, unfortunately cannot be further examined here.
60. Lyons (1977), p. 34. "Gardiner, 1932" is, of course, The Theory of Speech and Language (Gardiner (1951)).
61. The emphasis is in Gardiner's original (Gardiner (1922), p. 360) and as it is quoted in Ogden and Richards (1923), p. 193. It is interesting to note that in the full quotation that Ogden and Richards use, Gardiner anticipates one of Searle's (Searle (1965)) and Ziff's (Ziff (1967)) famous, if misguided, criticisms of Grice's theory: "Is the meaning of a sentence that which is in the mind of the speaker at the moment of utterance...? ...[N]ot that which is in the mind of the speaker, for he may intentionally veil in his utterance the thoughts which are in his brain, and this, of course, he could not do if the meaning of the utterance were precisely that which he held in his brain" (Gardiner (1922), p. 360). Gardiner thinks the formulation he presents - the one quoted in Ogden and Richards - is not subject to this criticism. See the footnotes on pp. 83 and 101 of Gardiner (1958) for responses by Gardiner to the criticisms of Ogden and Richards.
62. Ogden and Richards (1923), p. 195. It is to this sentence that Gardiner responds in his footnote on p. 101 of Gardiner (1951). It is also interesting to note that Ogden and Richards included a supplemental essay by Bronislaw Malinowski in their book in which Malinowski makes the following claim: "Dr A. H. Gardiner, one of the greatest experts in hieroglyphic script and Egyptian grammar ... has published some remarkable articles on Meaning, where he approaches the same problems as those discussed by Mr Ogden and Mr Richards, and solved by them in such an interesting manner, and their respective results do not seem to me to be incompatible" (Malinowski (1923), p. 298; in a footnote Malinowski indicates that the "remarkable articles" he has in mind are Gardiner (1919) and Gardiner (1922)). But, whatever the compatibilities between Gardiner's and their own, Ogden and Richards do seem to present the intention-based view of meaning as something rather misbegotten.
63. See note 59 above.
64. I don't at all include Saussure in the intention-based tradition - even though he is clearly important to it - because Saussure was not concerned directly with notions of speakers' intentions, no less with the identification of meaning with such notions. Saussure's importance cuts across the various traditions that I discuss in this chapter.
65. See note 62 above where Malinowski is noted as having remarked that he takes Gardiner's approach to be compatible with that of Ogden and Richards.
66. See Peirce (1955a), and Welby (1977) (the Welby/Peirce correspondence). For an overview of Peirce's theory of signs, see Rorty (1961). Peirce's theory is extremely complicated and his writing is replete with all sorts of unfamiliar technical jargon. Langer put it nicely: "Charles Peirce, who was probably the first person to concern himself seriously with semantics, began by making an inventory of all 'symbol-situations,' in the hope that when all possible meanings of 'meaning' were herded together, they would show empirical differentia whereby one could divide the sheep from the goats. But the obstreperous flock, instead of falling neatly into a few classes, each according to its kind, divided and subdivided into the most terrifying order of icons, qualisigns, legisigns, semes, phemes, and delomes, and there is but cold comfort in his assurance that his original 59,049 types can really be boiled down to a mere sixty-six" (Langer (1942), pp. 55-56). (Langer's source here is the portion of the Welby/Peirce correspondence that was published in Ogden and Richards (1923).) If there is a really tendentious aspect of the present history it is surely in my not giving Peirce a more central place. But my reason for not giving him a more central place is simply that I have not been able to convince myself that his influence over English-language philosophy in the twentieth century has been as great as Welby's and Gardiner's. Though Welby and Gardiner are little known today, they actually did have quite an influence, I believe, in their time as I tried to show to some extent in the main text above. But Peirce's contribution to the theory of meaning was not as widely read and known as either Welby's or Gardiner's. Recall also that Peirce implicitly acknowledged Welby's work on meaning to be earlier than his own; see his letter to Welby, discussed above, of March 14, 1909 (Welby (1977), pp. 109-110).
67. Most important for the present discussion are Russell (1919), Russell (1921), Lecture X, Russell (1923), and Russell (1926).
68. Ogden and Richards (1923), chapter 3. See Max Black's discussions of Ogden and Richards theory, which are actually helpful for understanding causal theories generally as well; Black (1949c), Black (1949d), and Black (1968).
69. Stevenson (1944), chapter 3.
70. Morris (1946). See Max Black's article on Morris, which, again, is helpful in understanding causal theories generally; Black (1949b).
71. I would like to mention however that Ferdinand Tönnies also provides something of a causal theory of meaning in his paper "Philosophical Terminology" (Tönnies (1899)). It is not unreasonable to wonder whether Tönnies might have been influenced in his causal theory by Hobbes since Tönnies was the editor of Hobbes's Behemoth and The Elements of Law as well as the author of a book on Hobbes's life and work (Tönnies (1925)). For a little about Hobbes's view on language, see the section on Hobbes in Kretzmann (1967) and see Landesman (1972), especially the quote on p. 2.
72. Ogden and Richards put this nicely; see Ogden and Richards (1923), pp. 10-11, especially the diagram on p. 11.
73. This isn't quite Stevenson's theory though it might sound a little like it here. As I said, though, I can't get into the subtleties here. See Stevenson (1944), chapter 3.
74. Of course, theorists who spoke in terms of "tendencies to behavior" might not be so happy with my referring to such things as mental states. But causal theories have been around longer than behaviorists and it seems that the behaviorists merely took the older theories and replaced the allusions to mental states with allusions to things like tendencies or dispositions to behavior.
75. See Russell (1919) and Russell (1921), Lecture X.
76. In Russell (1921), Russell seems already less committed to images than he was in Russell (1919) and towards the end of the former he says of the meanings of public-language expressions and mental images that "the relation which constitutes meaning is much the same in both cases. A word, like an image, has the same associations as its meaning has" (Russell (1921), p. 210). This sort of claim is significant because it shows that though Russell was using images in his theory, he wasn't depending on their being images to do semantic work: it is not because images resemble things that they mean, but because they cause the same associations as their meanings. One can wonder why Russell bothered to use images if their semantics is determined causally, but that is a matter for another time. Note, though, that in Russell's second review of Ogden and Richards (1923) he says, "These authors urge - rightly, as I now think - that 'images' should not be introduced in explaining 'meaning'" (Russell (1926), p.140). But see Russell (1948), p. 96 and my discussion in chapter 1 above, note 49, for indications that Russell seems later to have not only embraced images again, but to have let them mean in virtue of their resemblance to what they are images of. Go figure! Unfortunately, I can't discuss this any further here either.
77. "A sign, or representamen, is something which stands to somebody for something in some respect or capacity. It addresses somebody, that is, creates in the mind of that person an equivalent sign, or perhaps a more developed sign. That sign which it creates I call the interpretant of the first sign" (Peirce (1955a), p. 99).
78. "A sign is always a stimulus similar to some part of an original stimulus and sufficient to call up the engram formed by that stimulus" (Ogden and Richards (1923), p. 53). "An engram," they tell us, "is the residual trace of an adaptation made by the organism to a stimulus" (ibid.).
79. "...[A] sign's descriptive meaning is its disposition to produce cognitive mental processes, where 'cognitive' is to be taken as a general term designating such specific kinds of mental activity as believing, thinking, supposing, presuming, and so on" (Stevenson (1944), p. 62). True, Stevenson does speak of tendencies to behavior also. But he seems not to want to be taken as a full-blown behaviorist: "In stressing dispositions to action, the present account is not presenting an uncompromising defense of behavioristic psychology" (op. cit., p. 66).)
80. "If anything, A, is a preparatory-stimulus which in the absence of stimulus-objects initiating response-sequences of a certain behavior-family causes a disposition in some organism to respond under certain conditions by response-sequences of this behavior-family, then A is a sign" (Morris (1946), p. 10, emphasis Morris's). See note G on p. 250 of Morris (1946) where Morris cites Tarski as the one who gave him the idea to try to state only a sufficient condition for something's being a sign, rather than a necessary and sufficient condition.
81. See, for example, Grice's criticisms of Stevenson's theory in Grice (1957), Black's criticisms of Ogden and Richards's theory in Black (1949c), and Black's criticisms of Morris's theory in Black (1949b).
82. It is perhaps important to note that various philosophers who I have mentioned in this history have at least vaguely broached the subject of meaning in the sense in which a speaker is said to mean something by an utterance. What is important here is that though such philosophers addressed this notion, it seems not at all to have occurred to them that perhaps a theory of meaning could be based on this notion as it did to Welby. Langer is a good example of such a philosopher. She speaks of "the ambiguous verb 'to mean'" and says of it that "sometimes it is proper to say 'it means,' and sometimes 'I mean.'" She continues, "[o]bviously, a word - say, 'London' - does not 'mean' a city in just the same sense that a person employing the word 'means' the place" (Langer (1942), p. 55). But though Langer notes the notion of speaker meaning, it is only to put it aside: "In the further analyses that follow, 'meaning' will be taken in the objective sense, unless some other is specified; that is to say, I shall speak of terms (such as words) as 'meaning' something, not of people as 'meaning' this or that" (Langer (1942), pp. 57-58). The fact that Langer didn't make use of a notion of speaker-meaning doesn't makes her work any less interesting, by the way. But Langer represents a clear case of a philosopher who identified something like a notion of speaker-meaning only to let it go. Welby did not do this.
83. I would also like to say that I would not be horribly shocked if it should somehow be revealed that Grice read Gardiner. There is good evidence that both Ryle and Austin had read at least something of Gardiner since they both make reference to him (see above in my section on Gardiner). But, in fact, I have found no direct evidence that Grice read Gardiner.
84. Interestingly, Stenius seems not at all to have been influenced by the intention-based tradition in his "Mood and Language-Game", but rather, he was explicitly influenced by Ogden and Richards and Russell and their causal theories (Stenius (1967), p. 259). Yet Stenius's paper has had a pretty big affect on the Gricean tradition largely through its influence on David Lewis (see Lewis (1969) and Lewis (1975)). This is just one place that the causal and intention-based traditions dovetail in more recent work.
85. Carnap (1928), pp. 34-35.
86. Or, slightly less ambiguously, the theory is that there is a relation such that a sentence means what it does among a group because each member of that group stands in the relation to a compositional-truth theory for the language to which that sentence belongs. See the next note for more on this theory.
87. The theory that the argument now under discussion attributes to Frege is very close to that which is associated with Davidson in his various writings since Davidson (1967) (or maybe to some extent since Davidson (1965)). For, to know the logical form of a sentence is to know how that sentence is accommodated in what is sometimes called a compositional-truth theory. Of course, one aspect of Davidson's thought on language is a rather disparaging view of the idea of giving an analysis of the concept of meaning. But putting this aspect of his thought aside, we can say that insofar as there is something like a theory of meaning suggested in Davidson's work, it is something like this: for a sentence to mean something among a group of people is for the sentence to be part of a language that has a compositional-truth theory and for the members of the group to stand in a certain relation to that truth-theory. (See also Peacocke (1976) for something like this gloss of Davidson's views.) I will discuss the idea that a theory of meaning will require that language users know a compositional-truth theory below in chapter 4.
88. If I don't discuss this tradition here it is because it will be dealt with, but conceptually, not historically, in chapter 4 below. See Davidson (1963), Davidson (1967), and Davidson (1973). See also Evans and McDowell (1976).
89. To speak of the psychological grounds of meaning here is not to speak of the psychological grounds of meanings. Looking for the psychological grounds of meaning, in the sense in which this is intended here, is not inconsistent with Frege's platonism, that is, his anti-psychologism, with respect to meanings.
90. Wittgenstein used the term logical form in a somewhat special sense (see 2.18, 4.12 and 4.121, for example). I am using the term here as I did above in the discussion of Frege. Wittgenstein used the term "complete analysis", more-or-less, like I use the term "logical form". Thus, he says: "A proposition has one and only one complete analysis" (3.25). (Quotations from the Tractatus are all taken from Wittgenstein (1961) and I will cite their locations, as I do here, by just giving the number from the Tractatus of the passage.)
91. Note that the term "proposition" in translations of Wittgenstein usually refers to sentences, not to the supposed abstract objects that are taken to be the referents of "that"-clauses as it does in Moore and many more contemporary philosophers. (See Moore (1953), a collection of lectures from 1910-1911 given by Moore; see especially chapter 3 - "Propositions" -, and chapter 14 - "Beliefs and Propositions".)
92. Wittgenstein in the Tractatus does believe that natural languages are fully compositional since he believes that all languages are wholly extensional, nay, truth-functional. See 5.101 and 5.54-5.422. Also, see Smart (1986) where Smart tries to make something of a case that what Wittgenstein was trying to do in the Tractatus is akin in motivation and execution to some of what Davidson was interested in.
93. "Man possesses the ability to construct languages capable of expressing every sense, without having any idea how each word has meaning or what its meaning is - just as people speak without knowing how the individual sounds are produced. [¶] Everyday language is a part of the human organism and is no less complicated than it. [¶] It is not humanly possible to gather immediately from it what the logic of language is. [¶] Language disguises thought. So much so, that from the outward form of the clothing it is impossible to infer the form of thought beneath it, because the outward form of the clothing is not designed to reveal the form of the body, but for entirely different purposes. [¶] The tacit conventions on which the understanding of everyday language depends are enormously complicated." [4.002] But, this passage doesn't suggest an actual theory of how the meanings of the expressions of everyday language actually have their meaning due to tacit conventions.
94. Russell apparently wrote his introduction to the Tractatus sometime between December 1919 and April 1920. Wittgenstein and Russell met in The Hague in December 1919. In a letter to Paul Engelmann dated December 15, 1919, Wittgenstein wrote: "Russell wants to print my treatise, possibly in both German and English (he will translate it himself and write an introduction, which suits me)" (von Wright (1982a), p. 87; this was translated from German by L. Furtmüller). So, it seems, Russell hadn't written the piece as of that time. But, the piece was definitely finished by the beginning of April 1920 (op. cit., p. 93). This is significant because it means that Russell had already published Russell (1919) when he wrote the "Introduction". Russell, that means, was clear about the psychological nature of the theory of meaning: "Logicians, so far as I know, have done very little towards explaining the nature of the relation called 'meaning,' nor are they to blame for this, since the problem is essentially one for psychology" (Russell (1919), p. 7; I already quoted the beginning of this passage above in note 66). And in "The Philosophy of Logical Atomism", apparently before Russell had actually seriously attempted something of a theory of meaning: "I think that the notion of meaning is always more or less psychological, and that it is not possible to get a pure logical theory of meaning, nor therefore of symbolism. I think that it is of the very essence of the explanation of what you mean by a symbol to take account of such things as knowing, of cognitive relations, and probably also of association. At any rate I am pretty clear that the theory of symbolism and the use of symbolism is not a thing that can be explained in pure logic without taking account of the various cognitive relations that you may have to things" (Russell (1918), p. 45). So Russell clearly distinguished between logical and psychological sorts of issues with respect to meaning.
95. The story about Russell's views here, however, is probably more complicated. C. K. Ogden after reading the Tractatus for the first time, wrote the following in a letter which he sent to Russell in 1921: "I know that you are frightfully busy at present, but I should very much like to know why all this account of signs and symbols cannot best be understood in relation to a thoroughgoing causal theory. I mean the sort of thing in the enclosed: - 'Sign Situations' [chapter 3 of the published version of The Meaning of Meaning].... The whole book which the publishers want to call The Meaning of Meaning is now passing through the press; and before it is too late we should like to have discussed it with someone who has seriously considered Watson. Folk here still don't think there is a problem of Meaning at all, and though your Analysis of Mind has disturbed them, everything still remains rather astrological" (quoted in von Wright's introduction to Wittgenstein (1973), p. 3). In Russell's letter in response to Ogden, Russell says the following: "...I think probably the causal treatment of meaning does give the solution. It was because I thought so that I started working on 'Analysis of Mind', which grew out of the problem of meaning" (op. cit., p. 4). This is odd because Ogden seems to have been asking whether Wittgenstein's approach was misguided in a general way and whether the causal-theoretic approach was the right one. Russell appears to be assenting to Ogden. But if he separated the psychological task of providing a theory of meaning in the "Introduction" to the Tractatus from the logical task that he ascribed to Wittgenstein - whatever that was! -, then why would he assent to Ogden's question? I cannot sort through this now, but the question is there to complicate things.
96. In the latter case, whatever there is of Wittgenstein's interest in compositionality can be taken as I suggested above we take Carnap's interest in this.
97. I am indebted to Eleanor Williamson for her correspondence on the following matters and for much help with the research here. Though, of course, any mistakes about the later Wittgenstein are all my own fault.
98. Wittgenstein (1967), 16.
99. "Philosophers constantly see the method of science before their eyes, and are irresistibly tempted to ask and answer questions in the way science does. This tendency is the real source of metaphysics, and leads the philosopher into complete darkness. I want to say here that it can never be our job to reduce anything to anything, or to explain anything" (Wittgenstein (1958a), p. 18). This is by no means an argument against reductive analysis and it is not entirely clear how to take this. But it seems to illustrate Wittgenstein's negative view of reduction.
100. Against the mental-image picture, see, for example, the famous passage in Wittgenstein (1958a), p. 3, in which someone tries to fetch a red flower with a mental image. This passage isn't directly concerned with meaning, but is a general attack on the idea of mental-images playing useful explanatory roles in our psychology.
101. Here are some passages that indicate this dislike of the idea of meaning as a mental activity: "But - can't I say 'By "abracadabra" I mean toothache?' Of course I can; but this is a definition; not a description of what goes on in me when I utter the word" (Wittgenstein (1958b), § 665).
102. Likewise, I have not mentioned more recent work in causal-theoretic semantics. If I were giving a richer discussion, I would mention Stampe, Dretske, and Fodor, most notably, as well as a number of other philosophers. But such discussion is not possible here.