¹ When I set out to challenge anew the conventional uses of meaning as the basic tool of analysis in sentence structure and syntax—the area of linguistic study in which it has had its strongest hold—I felt very keenly an obligation to state as fully and as accurately as I could just what uses I did make of meaning in my procedures. This paper represents the result. Although the materials presented here use general terms, I should like to point out that my experience has dealt primarily with English. This statement of the principles and assumptions that have underlain and guided my studies of English may not have equal relevance to the problems presented by other languages.

² J. R. Firth, General linguistics and descriptive grammar, Transactions of the Philological Society 1950.82 (London, 1951).

³ John B. Carroll, A survey of linguistics and related disciplines 15 (Cambridge, Mass., 1950).

⁴ Leonard Bloomfield, Language 139, 140, 162 (New York, 1933).

⁵ Bernard Bloch, A set of postulates for phonemic analysis, Lg. 24.5 note 8, 24.6 (1948).

⁶ George Trager and Henry Lee Smith Jr., An outline of English structure 54 (Oklahoma, 1951).

⁷ Zellig S. Harris, Methods in structural linguistics 365 note 6 (Chicago, 1951).

⁸ Bloomfield, Linguistic aspects of science 12 (Chicago, 1944).

⁹ Bloomfield, Language or ideas?, Lg. 12.92 note 6. Bloomfield followed Weiss in objecting to the term behaviorism and believed that physicalism indicated much better the essential quality of the kind of descriptive statements he sought.

¹⁰ The meaning ‘performer of the action’ is one of the meanings signaled by the structure we call ‘subject’. We cannot, however, expect to define the structure ‘subject’ as ‘performer of the action’, for this meaning is signaled by a variety of other structures that are not ‘subjects’. For example, in each of the following sentences, the word committee has the meaning ‘performer of the action (of recommending)'; but only in the first sentence is this word in the structure of ‘subject': The committee recommended his promotion; His promotion was recommended by the committee; The recommendation of the committee was that he be promoted; The committee's recommendation was that he be promoted; The action of the recommending committee was that he be promoted. The structure of ‘subject’, on the other hand, signals at least five different meanings—four in addition to that of ‘performer‘—each distinguished by special formal arrangements. See C. C. Fries, The structure of English 176-83 (New York, 1952).

¹¹ To accuse linguists of deliberately refusing to treat meaning at all is to ignore the facts. The number of thoroughly trained linguists is very small indeed, and, although there are more positions for such linguists in the academic world than there were twenty years ago, there are not enough such positions to support a sufficient number of linguists to carry on the linguistic studies that are needed in every part of the field. For the past twenty-five years the really live issues that have claimed the attention of linguists have centered about linguistic structure. The new views of the significance of structure and the success of new procedures of structural analysis applied to various aspects of language have aroused such enthusiasm that most linguists have devoted their studies to these matters. Although the present center of liveliest interest in linguistics is structure rather than meaning, some portions of the general problems of meaning have received attention. (See for example Bloomfield's Linguistic aspects of science, 1939; Philosophical aspects of language, 1942; Language or ideas?, Lg. 12.89-95 [1936]; Meaning, Monatshefte für deutschen Unterricht 35.101-6 [1943].)

Even with respect to the lexicon, we should, for the record, note that the American linguists whose center of interest has been the English language have given constant support to the labors of those who have been struggling with the problems of producing the various period dictionaries proposed by Sir William Craigie. The Oxford dictionary itself was an effort to apply practically what was at that time called the ‘new philological science’.

¹² See Leo Abraham, What is the theory of meaning about?, The Monist 46.228-56 (1936). In this article are gathered more than fifty typical quotations from philosophical and psychological writers, in each of which the term meaning is used in a different sense. At the end of these quotations he concludes, ‘There is clearly nothing both common and peculiar either to all the various disparate senses, or to only the more familiar among them, which itself bears, or should bear, the name “meaning” ... A subject matter for the “theory of meaning” cannot, accordingly, by obtained by abstraction from all or some of the entities revealed by the linguistic phenomenology of the term “meaning”.‘

¹³ A. Gardiner, The definition of the word and the sentence, British journal of psychology 1922.361.

¹⁴ N. Campbell, Physics: The elements 132 (1920).

¹⁵ C. S. Peirce, Collected papers 5.165 (1934).

¹⁶ F. Anderson, On the nature of meaning, Journal of philosophy 1933.212.

¹⁷ A. J. Ayer, Demonstration of the impossibility of metaphysics, Mind 1934.333.

¹⁸ C. W. Morris, Pragmatism and metaphysics, Philosophical review 1934.557.

¹⁹ H. L. Hollingworth, Meaning and the psycho-physical continuum, Journal of philosophy 1923.440.

²⁰ W. E. Hocking, Philosophical review 1928.142.

²¹ M. R. Cohen, Reason and nature 107 (1931).

²² C. W. Morris, Signs, language, and behavior 19 (New York, 1946).

²³ W. B. Pillsbury, Meaning and image, Psychological review 1908.156.

²⁴ C. L. Stevenson, Ethics and language 33 (New Haven, 1944): “The emotive meaning of a word is the power that the word acquires, on account of its history in emotional situations, to evoke or directly express attitudes, as distinct from describing or designating them.‘ Id. 73: The independence of emotive meaning can be roughly tested by comparing descriptive synonyms which are not emotive synonyms. Thus to whatever extent the laudatory strength of ”democracy“ exceeds that of ”government where rule is by popular vote“, the emotive meaning of the former will be independent.‘

²⁵ William Fankena, Cognitive and non-cognitive aspects of language, Language and symbolism [unpublished tentative report] 5.27, 28 (1952).

²⁶ These ‘sanies’ must not be taken as the engineer's ‘norms’ with margins of tolerance—statistical norms clustering around averages; see Martin Joos, Language design, Journal of the Acoustical Society 22.701-8 (1950). They are ‘sames’ as the various types of ‘strike’ in baseball are functionally the same; see Fries, The structure of English 60-1.

²⁷ I take recognition here to mean not a conscious act of identification, but rather an automatic conditioned response connecting the patterns of vocal sound with recurrent features of experience. Recognition is itself a ‘meaning’ response. I am assuming that every kind of meaning has this kind of process. On every level, it seems to me, shapes, colors, sizes, smells, tastes have meaning only as they fit into patterns that connect them in some way with recurring features of experience. When stimuli do not fit such patterns of recurring experience they are ‘meaningless’, and confuse us. As a matter of fact it is usually the case that only features that do fit such patterns are reacted to at all; the others do not become effective features of stimuli. For adults there seems to be no such thing as ‘raw’ observation unrelated to any pattern of experience.

²⁸ A linguistic community consists of those individuals that make the ‘same’ regular and predictable responses to the ‘same’ patterns of vocal sounds. The language function is fulfilled only in so far as it is possible to predict the response features that will regularly be elicited by the patterns of vocal sound. For the discussion here I am not concerned with what might be called ‘personal’ meaning—the special non-recurrent or not regularly recurring response features that mark individual differences.

²⁹ For many others the meaning of a text or a sequence of utterances has often been considered a function of (a) the ‘words’ (as items of sound patterns which experience has connected in some way with reality), and (b) the ‘context’. This context has included both the so-called ‘verbal context’ or linguistic context (not specified further) and the ‘context of situation'—the circumstances in which the utterance occurs. Firth has pushed the analysis of ‘context’ much farther in his dealing with ‘formal scatter’ and ‘meaning by collocation’. See his Modes of meaning, Essays and studies (English Assn.) 1951.118-49, and General linguistics and descriptive grammar, Transactions of the Philological Society 1951.85-7; cf. his earlier Technique of semantics, Trans. Philol. Soc. 1935.36-72, and Personality and language in society, Sociological review 1950.

³⁰ In the study of the language records of a former time we have, because of the nature of the evidence, usually had to try to arrive at the meanings of the language forms by connecting them with recurring elements of the situations in which they were used. In the study of living languages it is often possible to observe directly the responses which particular language forms elicit in a speech community. We assume that if a particular response regularly follows the utterance of a language pattern, then this pattern ‘means’ this response. Upon such regular recurrences rests the kind of prediction that makes possible the social functioning of language.

³¹ ‘Complete utterance unit’ here means the total span of talk of one person in a single conversation or discourse.

³² As we record more specifically the details of the experience of language learning, we realize increasingly that we ‘learn’ not only the shape of a lexical item and the recurrent stimulus-response features that correlate with it, but also the sets of other lexical items with which it usually occurs. Perhaps when psychologists explore the ‘free association’ of words for an individual, they are really dealing with these sets of lexical distribution.

³³ In English the functioning units of the contrastive arrangements that signal meanings are not lexical items as such, but rather classes of these items. A variety of formal features make possible the classification of lexical items into a very small number of form classes the members of which each function as structurally the same. Linguistic analysis must discover and describe these form classes as a means of dealing with the structures themselves.

³⁴ The term ‘social-cultural meaning’ is not wholly satisfactory, but it is the best I have found to cover all the varieties of predictable meaning other than linguistic meaning. As indicated above, I have excluded from the discussion here the personal meaning of individual differences.

³⁵ This is true even of many of the varieties of social-cultural meaning—for example, the set of deviations from the norm of the sound segments that signal the meaning that a speaker is drunk, the whispering of an utterance that signals the meaning that the content of it is secret, and the unusual distribution that is the cue to a metaphor.

³⁶ Sometimes it is insisted that we use ‘differential’ meaning, not ‘referential’ meaning. Perhaps this statement means that the linguistic analyst seeks basically to establish the fact whether two instances differ in meaning content or not. He does not need to know what that content is or in what ways the two may differ. If they differ in meaning he assumes that there must be some difference in formal features, and sets out to find, prove, and describe that difference.

³⁷ [The foregoing article was offered as a contribution to the Edgerton number of Language (29:3), but reached the Editor too late to be included in that number.]