way defined and limited by the system. The result may be schematically represented as follows: St consists of the sound-combination [wud] which is in sign-context with S [wud]. S symbolizes the very general, common and complex reference RAB linking S to its two consciously related referents rA and rB. RAB is decomposable into the common complex reference RA linking S to rA and into the common complex reference RB linking S to rB. RA and RB respectively are decomposable into a certain number of more simple and specified references R1, R2. . . . and R'1, R'2.. When I utter the sentence 'I break this piece of wood' the symbolization process may be pictured thus: St may consist of the phonetic combination [aibreikðispisəvwud], which is in sign-context with the corresponding S. The latter symbolizes the sentence-reference R and is thus linked to its referent r. A comparison of the two processes reveals some striking differences. (1) a) The word is polysemous, that is, the same symbol is simultaneously in context with a variety of common and more specific references. This does not mean that all these references are consciously present with great or equal clarity; one or more may be set off, leaving the others in the penumbra of consciousness. b) The sentence-symbol has only one specified and well located reference, and, of course, only one referent. Again this does not exclude the possibility of ambiguous sentences. But if they are so, they have to be expanded or otherwise located until they are unambiguous in all relevant respects, or they cannot serve their natural purpose of adequate expression or communication. As to intentional equivocating as well as Psittacism, these outgrowths of speech may be ignored in a short sketch like this. (2) a) In the case of the single word both the symbol and its various references are engraphically or mnemonically predetermined.18 The voluntary aspect of the whole process is limited, on the part of the speaker, to the possible evocation of some part of the context which then brings forth the other parts automatically, and to the act of utterance; on the part of the listener, all the phases of the process are primarily passive. b) In the case of the sentence the following phases may be voluntary: on the part of the speaker, 1) the evocation of some contextual part, 2) the choice and the grouping of the various referential components, 3) the choice between a variety of possible symbols, 4) the act of utterance. On the part of the hearer the process is primarily passive. (3) a) The referential contexts of the word-symbol are engraphically present in all the individuals who know the language concerned, whether the word is uttered or not, provided, of course, that the word belongs to their acquired vocabulary. These individual contexts may vary within certain limits as to their number, both actual and potential, their clarity, their complexity, their adequacy, etc., but in their general and basic features they are alike. b) The referential context of the sentence-symbol is only present in the speaker and the hearer(s). 18 Cf. H. Ammann 45. (4) a) The spoken single word is exclusively or at least primarily a referential utterance. b) The sentence may be 1) a purely or primarily emotive expression, or 2) a purely or primarily referential communication. This is what we mean. The single word, as we have suggested before, never exclusively or primarily expresses an attitude (emotion, desire, etc.) of the speaker. Where it apparently does, as for instance, in exclamations 'Oh!', 'Beautiful!', or in curses, 'Damn!', etc., we really have sentences. I may, of course, detach a sound-combination like damn, or beautiful, or donnerwetter from the rest of the vocabularysystem, quite apart from any sentence-context; but if I do so, these symbols lose their exclamatory character and become mere mnemonic and polysemous symbols for complex references. When a phonetic combination like sacrebleu becomes a word-symbol, it does not express an attitude of the speaker. Instead of being a sign in an emotive context it becomes a symbol, whose complex reference is derived from many personal and impersonal sign-contexts sacrebleu, and whose referent is a certain attitude of any speaker. A certain emotive colouring may undoubtedly accompany the utterance of such a word, but this is decidedly a by-product. All that we utter by the single word as such is the symbol for a compound reference. Thus it is differentiated not only from an emotive expression, but also from a referential communication. Communication is positively foreign to the single word.19 This remains true, even if we assume that one person may act in a dual function: as speaker and as listener. If we really communicate a word to ourselves as if we were the hearer(s), this word becomes a sentence, which when expanded would be something like this: "The word-symbol for the complex reference "house" is [haus].' Such a phonetic expansion is obviously not needed here, because the sentence-context is largely expressed by the situation known to us. The fact that we always hear or that others may hear the word we utter, is merely an accidental circumstance. It is not the function of the single word to be communicated or to communicate. From a hasty interpretation of figures 1 and 2 it might be concluded that the single word is really nothing else but the sum of a certain number of sentences. As a matter of fact each word-symbol can become the subject of a certain number of sentences in each of which one of 19 Cf. A. Gardiner, "The definition of the word and the sentence,' Brit. Journ. of Psychology 1922. 352 f. the more general or more special references symbolized by the single word-symbol is the predicate. For instance, 'wood is a material', 'wood for building purposes is called timber,' 'a wood is a collection of trees', etc. But in the first place, in each of these sentences the reference of the symbol [wud] is specified by the sentence-context, whereas in the single word wood no such specification takes place. Secondly, the single word-symbol is in context with more than a mere sum of references; all the possible special references symbolized by the symbol have certain common features, however general, which represent the abstracted common reference of the word (RAB of fig. 1). Whenever the symbol is connected with a specified reference in a sentence-context, this reference is felt to be related in some way to the generalized common reference of the single word-symbol. In actual reality every new specification of a word-reference in a sentence results in a gradual transformation of the common abstracted reference; but psychologically the latter is static in the same way as the human body is a constant unit in spite of its biological changes.20 It is very difficult to define such a generalized common reference, but such difficulty is inherent to all definitions. It by no means follows that the object of the definition sought is less psychologically real.21 But even if we should succeed in defining it, we could only do so in a sentence. For instance, "The generalized reference of the word-symbol [wud] is x'. The reference of this sentence would not be the same, however, as that of the single word-symbol [wud] itself, which is x. The sentence-reference would be a relation of correspondence between word-symbol and word-reference; the sentence-referent would be the semanto-phonetic word wood. The referents of the single word are the material (wood), and a collection of trees. We can now see how apparently close the sentence and the word may come to be to one another, while actually they are just as removed from each other as the two poles of the earth. In the same respect a word is never a sentence and a sentence is never a word. The simultaneous presence of irreconcilable aspects in the same quantitative data is only superficially paradoxical. The same individual may at the same time. be father and son, although the aspects of fatherhood and offspring are diametrically opposed. But if we wish to characterize the son qua son, we must do so independently of his fatherhood. The latter may only 20 See H. J. Pos. 89, 90. 21 I do not believe that a specified sentence-reference is necessary in order to speak of a 'symbol-context'. See Ogden and Richards, 404. come in by way of antithetical and negative element of external demarcation, not as a positive element of referential definition. The expression 'sentence-word '22 in our conventional linguistic terminology is, therefore, a particularly unfortunate one and should be discarded. In linguistic science it is, at least as a dvandva-compound, just as sterile and misleading as the expression 'father-son' would be in anthropology. It is surprising that some inventive linguistic engineer has not yet thought of endowing our young science with such hybrids as 'sentence-syllable' and 'sentence-sound.' However, the term 'sentence-word' could be made useful by having it refer to the 'word-within-the-sentence'. As such it would form a most convenient contrast to such symbols as 'single word', or 'isolated word', or 'system-word'. The question as to what becomes of the word when utilized as part of the material of the sentence is a very interesting one and is able to throw more light upon the nature of the word. That we 'use' words in speech is beyond doubt, in spite of the fact that the sentence is the real and primary unit of speech and that we do not form sentences by merely juxtaposing words. As suggested before, the primitive word has developed from the speech-sentence;23 once, however, it had been recognized as a recurrent contextual entity, it was not only used in similar, though different, speech-contexts, but it soon also became the object of extra-speech combinations. Let us again resort to a schematic picture. The sentence 'I have not seen her' as pronounced in current speech may be represented as follows: S = the phonetic sentence-symbol; R = the sentence-reference; R1, R2, = the component references phonetically symbolized; r " Cf., e.g., L. Bloomfield, 64, 111. = the sentence-referent. 23 Cf. Rozwadowski, Wortbildung und Wortbedeutung 58 (1904). W. Wundt, Völker psych., Die Sprache I 609 f. |