4. fa fymt: SEMANTICS SCOPE, DEFINITIONS, METHODS. If this conference is going to he a conference on SEMANTICS the first question to be answered will be what we understand by "semantics*'. I understand that we have to deal exclusively, or primarily with vocal speech and language, and more especially with the problem of MEANINGS, The study of meanings has two, and only these two aspects : (1) meanings in language, i.e. the study of the meanings of significants of a given language (or more than one language), the significants being : words, parts or elements of words, combinations of words (a black swan, a white crow), sontenco intonations, and parts or elements of sentence intonations fYes, but does he know ?). (2)' meanings in the use of a language, i.e, speech; this os the study of the use made of meanings of significants in actual speech, or, better, the use made of significants as far as this use is determined by their meanings. 2, In order to study meanings we don't need a perfectly satisfactory defi- nit ion of "meaning” any more than e,g. mathematicians need a definition of a point and a line to build up the science of geometry. At least at the first stage, maybe in all following stages, the only thing wo need is a more or less provisional agreement on what we mean by "the meaning of a word" or any other significant. So we need a kind of ostensive definition^ Cp.my Structurele Syntaxis. 1949, 161; W.E.Johnson, Logic I, ans L.Susan 3tabbing, A Modern Introduction to Logic. 2nd ed., 1933, p,423. We may start from the assumption (hypothesis) that a word (the term word is here used loosely to denote any or several types of significants) has a meaning, and that this meaning is something different from "things" or "some things" in the so-oalled practical world around us (supposed, or denied, or fictional reality). By moans of the use of a word in a given situation a speaker may refer to one or more of such "things". This assumption may be based - among other things - on the following observations, a. The same "something in reality" may be referred to by different words having different meanings ; a girl may be referred to by the use of the words Jane, she, girl. president. darling, cat. beauty, perfection, angel, etc, b, The same word (in language), i.e. words having the same form (marks) and the same meaning (in speech) may be used to refer to different "somethings in reality". This can cause ambiguities. Using the word cat the speaker may leave his audience in doubt which cat ho is referring to ; a cat in the room, or on the roof, his own cat, his neighbour's; or oven whether ho rofors to an animal - or a woman. The word used by the speaker, or its moaning, is a clue to the some- ^ o thing he is referring to. This clue has to be used by the hearer in order to cr identify the something referred to, r— =3 T” 3. The study of meanings is one of the most essential parts of the study of c~ a language, consequently of linguistics. Linguistics is here used to denote the study of speech and languages, A significant is - to use a term of Roman Jakob- ^5 son's - bipartite. This does not mean, of course, that it has two parts, but it w t* means that a significant is :m indivisablo unit of form and moaning. The form is a collection or combination of marks • Form and meaning of a significant are O correlative. A form (of a significant) is a form of, or for a definite meaning, vsj ^ A moaning is a moaning of a definite form. To study forms of significants ^ "V without taking into account their meanings as such, is a theoretical contradic- tion, and practically impossible. Cp.f.i. Kennoth L.Pike, Grammatical prerequi- .sites to phonemic analysis. Word 3.155-72 (1947),
2- 2
It should be kept in mind, however, that such terms as word, morpheme, phrase, sentence, often are used to refer to the form of a word, a morpheme, etc. 4. A language may be called a system of significants in that it is both a system of symbol (i.e.significant) forms, and of symbol meanings.
System of word marks : (phonemes:)
System of word forms :
- cup kin kip
son, sun
p t k b d g m n ng
s in soon seen
- coon keep coop
etc
etc
System of word meanings son daughter
son
PARENTS father mother grandfather grandmother GRANDCHILD grandson granddaughter
(morning, afternoon, evening! day.week,month,year,
etc.
uncle aunt
man wife CHILD
etc, 5. The meaning of a word or any other significant is not a feature common to all the situations in which the word is used or may be used. The speaker may choose a wrong word. He may use it to refer to something that is not covered by the meaning of the word, e.g, when he refers to the stars by the eyes of night. The feeling tono of a word like pig is part of its meaning, but it is absent in many situations. It may be asked whether the meaning of a word has anything to do with the situation in cases where the speaker refers to the form of the signi- ficant itself, e,g, the genitive of the second declension in Latin has the phoneme -i. It depends on what we understand by the situation,, It may bo that K.Kuipers is right in maintaining that the only foature common to all situations in which a word is used is tho presence of the word itself. Such facts as the possibility of referring to an animal, as well as to a lady or a girl, or to the word cat as such, i.e. as a unit of sound and meaning, a si- gnificant (The word cat has three phonemes.) t by using tho v/ord cat; or of rofor- ring to a part of the body of an animal as well as to a part of a ship, or to the word eye as such, suffice to prove that the meaning of a significant is not the common feature to all situations in which the significant may be effectively used, i.e. so that the hearer is able to understand, to identify the something referred to, or the speakers attitude expressed. Remarkable is also the use of ''wrong** words resulting in perfect understanding on tho part of the hearer. So the meaning of a word is not a greatest common divisor, nor is it a least common multiple. The meaning of a significant has a more or loss definite number
3- 3 -
of aspects (referential or, and attitudinal), Except for special cases in each separate speech situation only some of its elements or aspects are functioning as a due for the hearer. In some situations or combinations even none of them. An example is obligatory rection of a definite grammatical category by a proposition, a noun, or a verb, or any other word, e.g. the ablative case in ab urbe, de urbe, ex urbe, or gender, number and oas of bonarum in b on arum feminarum, or the ablative in ut~i~gladio in Latin. Here the grammatical categories, which certainly have meanings of their own, are functioning as more indicators of syntactic rank, as marks of determiners. So it would be an illusiong to assume that the study of meaning sis just a matter of summing up a large number of situations. On the other hand it would be unscientific to avoid the study of meanings either because of its complexity and difficulty, or because meanings ar not perceivable, but thinkable. 6, Consequently, the study of the meaning of a significant : a word, a mor- pheme, a combination ofs words, a sentence intonation, or part of a sentence in- tonation, is much more delicate than making a comprehensive inventory of situa- tions. Among other problems we have carefully to distinguish independent and com- binatory meanings of the same word or other significant; correct and incorrect use which are not identical with effective and not effective use in a given situation; reference to the significant itself, and to something else present in the situa- tion (suppositio materialis, and non materialis); cases where the meaning of the significant actually covers the something referred to, or the speaker's attitudo expressed, and whore it does not (individual and bold so-called metaphores), and whore there even is contradiction between meaning and actually meant (e.g. in case of lying, or irony); cases of ambiguity, and others where there is no ambiguity. For each of these distinctions the method of distinguishing and the criteria to be applied may be determined and formulated. There may be other such distinc- tions to be made prior to the actual analys i3 of a me an ing, in order to eliminate cases where nothing or little can bo inferred from the situation as to the meaning itself. This priority, of course, is more theoretical than practical, In practice, the question whether the meaning discussed is independent or combinatory; whether in a given situation the use of the word or other significant is correct, or not; whether the significant refers to itself, or not; whether there is mataphore, irony, or ambiguity, usually cannot be answered till one has pretty well, esta- blishod its meaning, or, at least, has conceived a very probable working hypothe- sis about it. However, the_ only way to study a meaning is to start from situations in which the significant seems to be used correctly and effectively. In case of a language which is one's own mother-tongue, or of which one has a very good speaking and hearing knowledge, the linguist may apply the method of introspection, or, better, use himself as a test person. This procedure is valuable especially when dealing with one's own mother tongue. The results, however, should be checked by using other test persons having the same language for their mother tongue. In case of a foreign language other informants are indispensable. 7, A meaning of a significant is just 7/hat it has in common with the meaning of one or more other significants of the same language, plus in what it differs from them. So the definition of a meaning of a word nosessarily is done by means of words or word groups having the same or opposite meanings. The only alternative is an ostensive definition, i,e, an illustrative collection of sentences (or parts of sentences) with indication of the situations in which they may be used correctly end effectively.
4- 4 -
The opposition of a meaning to another may be simple, i.o. that they have just one element different, e.g. bo£ and girl, or night and da£. It may be com- plex, e.g. clock and thermometer. Two meaning may have no common feature, e.g. Charios and between. The features of a meaning always, or often, are more or less central or peripheric in relation to each other. In girl the feature of human being is more central than that of young and of female. Analogously the marks of a significant are arranged in a similar hierarchic way. The phonemes of a word and their order are more central than the number of syllables, which, in its turn, is more central than absence or presence, or the plhce of word ac- cent. Analogously, the distinctive features of a mark of a significant always, or very often, are arranged hierarchically. Evidently the pitch formants of a vowel in Latin are more central than length. Features of meanings, of forms of significants, and of marks of significants, in combinations are of different rank. The structure of a system in language is a collection of relations bet- ween the members of the system. These relations are not only oppositions, but also a matter of rank. The opposition between the meaning of the words boy and girl is the same as that between the word groups young human male and young human female, i.e. the identical and the different features are tho same. The rank, howovor, within the same meanings, is different. Methods and criteria to establish ranks of features of meanings should be found and applied. In case of the word groups mentioned word order is a mark of rank. There is no such mark or criterion in case of the words boy and girl. In this case other criteria should be found, or it should be proved that in this and similar cases there is no linguistic, and, if any, only a logical rank. The usual criterion, especially in case of nucleus and determiner in a subordinate word group, is based upon the definition that if a combination a plus b is such that a without b can be used in a similar function, e.g. syntactically, but not b without a, a "is called more central than b, For instance There were small is not usual or "correct, There were birds is. So in the group small birds the word birds is more central than small. In applying this criterion, however, distur- bing factors of various kinds have to be taken into account. For instance in many birds the word birds is more central than many, though There were many is correct, The same problems exist as to the rank of features of forms, and of marks of significants. 8. As a meaning is something opposed to meanings of other significants in the same language, and a significant in an actual situation functions as a clue (term of Alan H, Gardiner) owing to the fact that there are such oppositions, the method of studying meanings is analogous to the study of forms of signifi- cants, and of foatures of marks. It is the procedure of comparing in order to establish common elements and difforont olemonts. Thus by comparing the forms man and can, or (tho) import and (to) import marks of forms of significants are established ; phonomos and accents. Thus by comparing p,b.,_t,d,k,£ distinguishing features of phonemes are established. Thus by comparing boy and girl as to their meanings, common and different features are established, There seem to be two different kind of things that should be compared at the same time. First, in case of referential meanings, the things referred to them- solves (either objects, or persons, or qualities of objects, etc., or processes, or relations, etc.), and, incase of attitudinal meanings, such as meanings of sentence-words : interjoctionV and yes, no, etc., the states of mind or speaker- attitudes in the actual situations, Second, the contextual use of the signifi- cants themselves, in case of words the valence of the words, i.e. their ability
5of being used as a nucleus of definite type of groups, or as determiner in a definite type of group, and as determiner of a certain rank,. For instance sub- stantives may be used as a nucleus in a group substantive plus adjective.. An adjective may be used as a determiner of a substantive (good wine), or of a verb (the wine is good), but of lower rank than a deictic word, as is shown by word order : this good winenot : good the wine. Similarly of a lower rank than a numeral : these two good boys, not these good two boys, or rarely, and in that case owing to a disturbing factor. In case of bound morphemes,, such as cases or tenses, the valence of the words in which they occur may be established, i.e., identities and differences. So in Latin a substantive in a genitive case, i.e, with the morpheme of the genitive, usually is determiner of another substantive (pars pecuniae, part of the money), in an ablative case usually of a verb (pecunia adiuvaro, to help with money). A very interesting example of applying the first method is offered by William E.Bull's paper AN EXPLORATORY STUDY OF THE NATURE OF ACTIONS AND THE FUNCTIONS OF VERBS IN SPANISH, and his QUEDAR and QUEDARSE : A STUDY OF CONTRAS- TIVE RANGES (Hispania 32, 64-73, 1949, and Language 26, 467-480, 1950). An interesting synthesis of the problem of meaning in language and its use in speech is the paper REFERENTIAL ADJUSTMENTS by Garvin, in 9. In applying the method of comparison in order to establish identities and differences use should be made of various working hypotheses which have already proved to be of considerable help, and which, in fact, have not been proved wrong on the evidence of actual facts of speech or language. The first is that, except for special cases, i.e, cases with disturbing factors, a significant has always the same meaning. The seeming contradiction that it may be used to refer to, or to express, different things or states of mind, is accounted for by the fact that in speech usually only part of the fea- tures of a meaning constitute the actual clue to what is meant. This is a hy- pothosis of Roman Jakobson's. The second, of Trubetwkoy and Jakobson's, is that the oppositions of mor- phological or grammatical systems, such as cases and tenses, are always simple, even bitermal, i.e. of the type a/a plus b, or : a without b/a with b, or : a plus b/a plus c (cumulative, privative, and contrarious oppositions). A third seems to be that there is always different rank of features of meaning within the same meaning, as has been illustrated before. As to morphological systems, it has not been settled as yet whether, e.g. in oase systems, oppositions of meaning, or oppositions of syntactical valence of the words in which the morphemes occur, are primary, the others secondary. Compare L,Hjelmslev, La categorie des cas, 1935-37; Roman Jakobson, Beitrag zur allgemeinen Kasuslehro, Travaux du Cercle Linguistique do Prague 6,240-88 (1936) A.W, do Groot, Les oppositions dans les systemos de la syntaxo et des cas, Melanges Bally, 107-27 (1939); and, recently, J.Kurylowicz, Le probleme du clas- soment des cas, Bull, do la Soc.Polonaise de Linguistique 9,20-43 (1949), 10. The most important distinction of typos of meanings, or of features of meanings, seems to bo the distinction of reforontial meanings. used by the speaker to refer to something in the outer world, i.e, in supposed, or fictional thought of reality, and att itudinal meanings, used by the speaker to oxpress hin state of mind or attitude to things referred to (by means of words of the same
6- 6 -
sentence), or implied by the word itself, e.g. an interjection : Ouch ! or i Yes. No. Attitudinal are also statements and questions, e.g. the attitudinal actor-action construction he comes, as opposed to the merely referential actor- action construction (i wantT him to come, or birds sing as opposed to singing birds Compare my paper STRUCTURAL LINGUISTICS AND SYNTACTIC LAWS, Word 5,1—12, esp.4-5, where I called them objective And subjective meanings. See also my book STRUCTURELE SYNTAXIS, 1949, Register sub "Betekenis". In matter of language, as opposed to speech, the first aim should be to establish the structure of the system of all words of the given langage, i.e. the structure of the system of word classes, with its subdivisions. Compare my paper STRUCTURAL LINGUISTICS AND WORD CLASSES, in Lingua 1,427-500 (1948), esp.468- 472 and 484-487. By "all words" is to be understood also compounds and inflec- ted words, not only the so-called vocabulary as it figures in dictionaries. A.W. DE GR00T
A comprehensive inventory of significants in language seems to be : 1, words; 2, elements of words, which may be called morphemes, e.g. stem, aspect, tense, mood and person, also number of person, in Latin amavissem; 3, combinations of words, e.g. a black swan, or a white crow; 40 sentence intonations ; Yes. Yes ? Yes ! 5. parts of sentence intonations, e.g. Yes, but will he come ? (Statement plus question.).
On the problem of ranks of destructive features of marks, e.g. of phonemes, see my paper Neutralization d'oppositions, in Neophilologus (about 1935). This paper, of courso, by now is out-dated in many rospects. The problem is discussod also by Y. Michel, La Phonologie.