THE SEMANTIC FUNCTORPAPER FROM THE PROJECT LEXICAL STYLISTICS |
In this paper I will introduce a semantic feature that I have called a functor.(1)
My hypothesis is that the semantic functor is a feature different from other semantic components and features in that it is not in itself referential or expressive (attitudinal), but rather strengthens or amplifies another feature in the meaning of a word.
The hypotheses put here consist of (1) the introduction of the semantic functor and (2) the proposition that attitudinal features should be included in the meaning of words and not be considered as contextually evoked so-called connotations (see below). As an heritage from Charles Osgood's theories, attitudinal features and the semantic functor may be assumed to be of great relevance for the effect of the words in a text and hence for the style of that text.
One incentive for this study is a project, lexical stylistics, the main goal of which is to find out whether it is possible to discern and isolate stylistically relevant features in words (or in word meanings). It is then a quite natural consequence that work on the project started with the fundamentals of lexical semantics. On the basis of our work since 1992, I have come to the conclusion that word meaning may be described satisfactorily by means of the following components and features:
1. Referential (2) components (sometimes less adequately labeled cognitive (3));
2. Attitudinal features
3. Roles or cases and frames (as in Fillmore's and post-Fillmore analysis; especially valid in the analysis of verbs and verbal nouns and adjectives). For a complete (or nearly complete) analysis of verbs (and nouns that are possible to regard as derived from actions) one comes quite a ways with a combination of this model and the componential analysis. (4)
4. The semantic functor
All words certainly do not have all these features, as I indicate above, and some functional words may not have any of them. (5)
Referential components are those, as established for instance in structural analysis of concrete nouns, such as in the well-known analysis of chair, having as components prototypical features, as in that case four legs and support for the back. Such components are sometimes even presented with a plus or a minus -- a legacy from structuralist phonology. (6)
A different kind of referential component is one with positive and negative value components (as opposed to attitudinal features -- see below); these components are excluded from semantics by certain scholars who -- wrongly -- deny them distinctive force. (7) As an example of a word with a referential evaluative component I propose the Swedish ruckel ('shack'). The definition in Svensk ordbok is 'dåligt byggt ... fallfärdigt hus' ('badly built ... dilapidated house'). Ruckel thus has a referential negative semantic feature. If you take away the negative component, a distinctive part of the definition is lost.
In Anglo-American semantics, terms such as emotive or expressive meaning are often used to refer to `evaluation',(8) but I don't know of anyone who has made a distinction between what I call referential evaluative components on the one hand and attitudinal ones on the other. Attitudinal components do not refer to any object spoken of; they express the attitude of the speaker towards the object spoken of (or towards the addressee). As an example of a word with an attitudinal factor I propose the Swedish profit. In Swedish this word has a clear negative meaning (of exploitation or making money from others' work) as distinguished from vinst, which is neutral in this respect. The words share the referential components, but profit has a further, attitudinal, feature. (In philosophical semantics that feature might be one of the connotations.) This means that the negative feature in profit is not referential in the same way as the negative component in ruckel: The negative feature in profit expresses the attitude of the speaker, it does not refer to his atittude. It is important to note that this does not mean that this negative feature is not part of the meaning of the word: the negative attitude is lexicalized, i.e. not evoked by context, since there is a distinct difference in meaning between profit and vinst in Swedish.
The referential negative value in ruckel, however, is easily transformed into an attitudinal one if you use the word in a metaphorical sense, for instance calling a perfectly decent building a ruckel. Only if the word is used in the metaphorical fashion does it receive stylistic relevance in a text. Lexically, it has the potential to become relevant, but, unlike profit, ruckel has no style value in its lexical meaning.
The semantic functor
Because attitude (in a broad sense) and strength are important ingredients in what is usually meant by style, attitudinal features might be an important part of the style value in words. Dictionaries nearly always fail to provide information about these subtle aspects of meaning (9) and, in general, linguists and semanticists try to avoid these phenomena as much as possible, obviously because they are so difficult to handle.(10) The reluctance to make evaluative referential elements and non-referential elements of signification the subject of thorough investigation is evident. This fact is the more deplorable because the information called emotive and the like is usually the information we want to focus upon when we use words with emotive meaning. (11)
One of the very few scholars to actually have studied what other semanticists tend to call 'additional' -- and thereby not interesting -- features of meaning was the late American psycholinguist Charles Osgood. In The Measurement of Meaning from 1958, he elaborates three main factors in meaning: evaluation, potency and activity.
Osgood shows, with a rather elaborate statistical apparatus, that these are the factors most often connected with words and concepts. One of the difficulties in Osgood's theory for a full-blown linguist is that, when people are asked to react to items in his tests, it is really not possible to discern whether they react to words or to meanings or to concepts or, in fact, to reality. So when, for instance, people consider the test word train as having a high degree of ACTIVITY, this seems to have more to do with reality than with the word. On the other hand, we have sometimes felt that this might not really have to be a drawback: actually it seems to be embarrassing only when you try to isolate 'linguistics' from what language is all about or isolate 'words' from text.
From a semantic point of view, words such as explosion or earthquake might include a factor of intensity, since the phenomena they denote are so intense, i.e. in Osgood's terms, they have so much POTENCY. That factor would however not be stylistically relevant at all in a news report -- as little as ruckel, as we have seen above, might have any stylistic relevance as long as the word is used in its literal meaning. If, on the other hand, I tell you that the mere sight of a certain person causes a formal explosion in my heart, the potency of the word will have a stylistic impact for the reason that I could have used a less explosive metaphor - perhaps stir, affect, move or the like. (12)
In Osgood 's original model, there is strong evidence for the existence of an evaluation factor, less, but still satisfactory, evidence for a POTENCY factor, and fairly inconclusive evidence for a factor of activity. Our project began with the hypothesis that these three features form a certain quality in words that we call the style value of a word, considering that the attitude and intensity expressed in words have stylistic impact on the text which they form.
It was my disbelief in the existence of a discrete factor of ACTIVITY that first brought me to the idea of a semantic functor, this functor corresponding to Osgood's variable POTENCY. The amplified feature that the functor strengthens may be a scalar one.
I believe that in go, walk, run, rush, to choose a word field at random, there is a basic feature CHANGE that may have different specifications (such as TIME, MOVE(MENT) and the like) which can be operated upon by a functor POTENCY or STRENGTH; the feature MOVE influenced by this functor may result in FAST. The stronger the functor, the faster the movement. NB that I have omitted many features that characterise the movement in the different words!
WORD:
-----------------------------------------------------------------------------------------------
go
----------
CHANGE
MOVE
...
________________________________________________________________
walk
-----
MOVE
...
_________________________________________________________________
run
-----
functor: strength
v
MOVE
v
speed
or: functor potency(MOVE) = speed
_______________________________________________________________
rush
-------
functor: great strength
v
MOVE
v
great speed
functor much potency(MOVE) = great speed
________________________________________________________________
(The v-sign means 'turns into')
The feature MUCH (or whatever you want to call it) in a word such as big can thus be amplified by the functor, and the result is lexicalized as enormous or giant(like) etc. In Swedish and Norwegian we have jätte-, kämpe as amplifying prefixes.(13) In this case the functor is lexicalized. In "strong words" it is very often the evaluation that is strong. An example of how the functor works on an attitudinal feature is shown in the near- or quasi-synonyms we seem to need in language to express and vary our dislike and contempt for a prostitute; this dislike is expressed by the functor working on the negative factor in the words together with minor changes in meaning and/or reference. (14)
Style
A few remarks about style may be useful. There are actually few things stylisticians agree upon -- even the very concept of style has been challenged! (15) -- but we do agree that style is a concept that applies to texts (in a semiotic definition, i.e. communicative units). (16) The aim to isolate stylistically relevant features in words might thus sound paradoxical, but we can observe stylistic effects on many levels. Some are achieved by phonological, morphological and, not least, syntactical devices; but the most pertinent effects are no doubt caused on the lexical level by lexical units - words. Hence it might be valuable to study words in isolation, even if words in isolation must be seen as some kind of abstraction. (17)
Style level and style value
There are two aspects from which words can be viewed stylistically: one being style level (18)
(often called 'usage' in dictionaries) and the other being what the project has called style value. (19)
For those of us who believe in a rhetorical definition of style, style is a relation between what is said and how it is said, i.e. a relation between form and content, if you choose these terms. (20)
Style is thus neither form only, nor content per se. Stylistic relevancy (21)
depends on the effect of the relation between what is said and how it is said and hence is a scalar concept. This also implies that if there is only one way to express a fact or a thought, this way is stylistically neutral. Style thus presupposes alternative possibilities to express more or less the same idea.
A rhetorical definition of style presupposes that effect is a feature of style, (22) as rhetoric is the art of persuading and influencing people by effective delivery, and, finally, because style (elucutio ) is one of the most important means, in turn, to achieve this end. A relation between form and content that has no effect whatsoever is consequently stylistically irrelevant. Since the semantic functor achieves the effect of amplification, one facet of style value seems in any case to arise from the combination of two types of semantic features, i.e. the functor working on a feature, which may be referential or attitudinal. Since furthermore the different ways in which speakers can express meanings may expose or reveal the speaker's attitude to what he speaks about and to whom he speaks, the attitudinal feature in word meaning is also highly important for stylistics.
The project has had to discuss whether stylistic or expressive features should (and can) be categorized as specific, discrete elements in the structure of isolated words, whether it is possible to analyse all features in semantic terms or whether stylistically important features should be defined in terms of relations between semantic features. We think it is quite unimportant as to whether these features should be labeled stylistic or semantic.
These considerations have been the theoretical platform for our project and we have come to believe that if there are 'stylistic features' in the prototypical meaning (23) of a word, or semantic features that may have stylistic implications, the definition of this word in a dictionary must include this information. This should be self-evident, but it by no means is to all semanticists (24) and neither apparently to lexicographers. (25) The reluctancy to include style value in lexical definitions is certainly partly due to the contextual sensitivity and hence variability of some of these aspects, which is the reason that they are called connotations in linguistics. (26)
The term connotation is interesting in the perspective that style is often metaphorically circumscribed by words such as overtones . (27) The simile is well founded, as real overtones, i.e. second harmonics, can be distinguished from the cluster of the fundamental frequency and the third and so forth harmonics only by spectrographical analysis. The same goes for stylistics: overtones, values (in Swedish valör), sound (Swedish klang(färg)or just färg 'colour') in a metaphorical sense are some of the words used to described stylistic effects.
Let's look at a passage from 'As You Like It' for clarification of the concepts style level and style value since this famous passage shows some of the richness of language that makes it possible to express more or less the same information on different levels of style (or in different registers) and at the same time demonstrates aspects of style value.
As you see, Shakespeare makes a distinction between perish, on a high style level, and die on a lower, and between kill , make away and translate life into death; kill is on a lower level than translate life into death, but on a higher level than make away. These differences remain, as you see, uncommented in the text, but Shakespeare labels woman as being common in comparison with female, leave as vulgar in comparison with abandon, and company as boorish in comparison with society . These labels not only mark a certain level, they also mark a derogatory attitude. (29)
Quite naturally, style in the lexicon (in our mind) is closely related to synonymy, as synonyms are - if they 'are' at all (30) - words that mean more or less the same thing. (31) Only if one and the same referent can be expressed in more than one way can there be any question of style. (32) But synonyms or so-called near-synonyms actually differ, and they may for instance differ in regard to the attitude expressed by the speaker towards the listener and the object spoken about.
Style and function
As we have seen, semanticists ascribe different functions to different features in the meaning of a word or in a different realm of meaning. It might be rewarding to contrast this standpoint with one that Michael Halliday launched in a paper in 1976, 'Functions and universals of language'. Halliday disagrees with the Czech structuralists Novák & Sgall, (33) who had proposed that functions in language should be ascribed to utterances (parole), whereas Halliday proposes that functions belong to language (langue). As function is a key word in stylistics, there is good reason for reflection upon the use of terms such as function & language. Both terms are in fact ambiguous: the meaning of function in connection with language depends on the preposition between the nominals: functions in language mean 'functions of language in language' , i.e. in the system, whereas functions of language ought to mean 'functions of language in communication and in society'. (34)
The idea expressed in this paper, namely that different kinds of semantic features that together make up the meaning of a word might have different functions, indicates that function might also be found on this micro-semantic level of language. What we want to communicate or, in other words, which function we intend our utterances to have is the main topic of Karl Bühler's Organon model (35) - the original, which seems more or less forgotten for the modifications that have been done: the Jakobsonian one dominated until recently (and perhaps still does in the USA), whereas today's stylistic papers very often quote Michael Halliday's extension or reduction - whatever - of Bühler's model of the Organon. In his paper (1976) Halliday modifies it to combine or collapse Bühler's Ausdruck- and Apell- functions into one, the interpersonal function, and adds a function in language that Bühler no doubt presupposed: the ability to form texts, which Halliday calls the 'textual function'. In the child's phases in learning the mother tongue, the textual function is well motivated. I am not sure that this is the case if we use the model as Bühler intended. Neither does the linguistic argument for combining Bühler's 'conative' and 'expressive' functions convince me: the argument that Halliday proposes for his standpoint, that the difference between Bühler's 'conative' and 'expressive' functions is (linguistically) 'very tenous', is that there is no surface (syntactic or morphological) difference between questions with different functions. Halliday asks: "is an interrogative, for example, a demand to be given information (conative), or an expression of a desire for knowledge (expressive)?" (1976: 27). (36)
In Bühler's Organon model, the function of Darstellung (Symbolfunktion) (i.e. representation or ideation or whatever you would like to call it) means that there is something in the world that you (can) speak about. (37) There is also the interpersonal (38) axis, where some utterances focus on the speaker - the term Bühler uses for this function in Sprachtheorie (1978 [1934]) is Ausdruck and the function is called Symptomfunktion, which seems to me to be a more adequate term than 'expressive'. This function has a prototypical manifestation in the word class of interjections (39), and thus, contrary to Halliday's claim that there is no grammatical equivalent to this function, there is, and is one for the Apell- or Signalfunktion ('conative' in anglosaxon terminology) in Bühler's model as well, namely the subjunctive and imperative moods. (40)
Therefore Halliday's view that the difference between Bühler's 'conative' and 'expressive' functions is 'very tenous', even if we constrict ourself to a systemic aspect of language, might be questioned. Nevertheless, there can be little doubt that the two functions could be seen as distinct from the symbolic function; in fact, Jakobsons 'phatic function' is the interpersonal function in its absolute, since the importance of what you speak about is here minimalized. An observation, which no doubt is one of the fundamental achievements in the original Bühler Organon model and which might disagree with Halliday's argument against dividing symptom and signal, is Bühler's emphasis on the three fundamental functions usually all being represented, although in different proportions or, in Bühler's own word, dominance (Dominanz) (1978: 30). This is actually stressed by Halliday himself in the same article (1976: 30). What you might then observe is that surface structures do not, according to Halliday, distinguish between functional dominance types in questions. But this also goes for propositions: 'It's cold in here' might have any of Bühler's functions as dominant and certainly is always referential and symptomatic. It might, in favourable circumstances, also have the Apellfunktion as dominant, for instance if I am of a high social rank addressing myself to someone else where there is an open window. In communication there can be a huge difference in the function of seemingly identical utterances -- we all know about the presuppositions or felicity conditions that influence the meaning and function of utterances in pragmatics. There is obviously a difference in the meaning of function: "In my opinion [says Halliday] a functional theory is a theory of language, and is an essential aspect of any theory that attempts to explain the nature of language" (1976: 27). It is important to bear in mind that this functional theory which Halliday advocates is an intrinsic theory of language, not an extrinsic one, nor a theory of communication. But even though Bühler called his book Sprachtheorie, there is no doubt that his Organon model is part of a theory of communciation; Roman Jakobson's expansion already obscures this vital difference, which seems to be primarily a difference in purpose and perspective. It regards your aim: whether you use communication to cast light on linguistics or whether your point of departure is communication which can be explained by means of language.
It might be of interest that, early on, Hobbes noted in Leviathan four uses of speech:
What's especially interesting here is the last function, rather close to the poetic function proposed by Jakobson!
The original Bühler model is as I said more rarely seen and in my opinion far too little known.(41) There are several reasons for a re-reading of Bühler apart from the Organon model: his criticism of behaviourism (1978: 27), his stressing of process and function instead of what he calls Stoffdenken (thinking only of substance) and his remarkable closeness to modern natural sciences. (42) For the stylistician, the original Bühler Organon model is actually extremely important (43), as it is a basis for interpretative stylistics and hermeneutics. The difference between the circle ('das konkrete Schallphänomen') and the triangle ('das sprachliche Zeichen') is very important: the triangle is both smaller than the circle ('Prinzip der abstraktiven Relevanz') (44) (p. 28) and greater ('um anzudeuten, daß das sinnlich Gegebene stehts eine apperzeptive Ergänzung erfährt') (ibid.).
Of course the fact that Bühler is primarily interested in communication and psychology while Halliday is primarily interested in systemic grammar can be one reason for the differences between the models. More important, it seems to me, is that Halliday's model is based upon how the child learns to mean: where Bühler presupposes knowledge of language, Halliday shows how the child begins to master language, therefore motivating the 'textual function'. At the moment that the child has learned the language, this function becomes becomes intrinsic and hence is not of the same kind as the other two, extrinsic, functions.
Conclusion
'Lexical meaning' should thus include features from all four different realms of meaning and of so-called co-meaning, where we can discern (at least) two types : a) the part of meaning that gives some 'additional information' about the speaker (i.e. a person who would call all types of houses shacks, which tells us he has a severely restricted code etc.) and b) collocational meaning. Information about word meanings that should be analysed and introduced in dictionaries tentatively could be of these kinds:
___________________________________________________________
STYLE LEVEL
| HIGH
scrutinize |
MEDIUM
see, look |
LOW
gape, peep |
STRENGTH
Of Osgood's three factors, we consider the EVALUATION factor to be a distinctive semantic feature which contributes to the referential as well as to the attitudinal meaning of a word. POTENCY is seen as a semantic functor that gives a word its strength and which makes it possible to reduce Osgood's three variables to two: Osgood's variable ACTIVITY is explained as a function of the semantic functor POTENCY and a component of movement. Strength in a word also enables it to influence other words and to remain uninfluenced by context. The last step in our project will be to try to cast light on how words interact in texts, a question that was actually the very incentive of the whole project. (46)
References
Aitchison, Jean (1987). Words in the Mind. Oxford UK & Cambridge USA:Blackwell.
Baldinger, Kurt (1980). Semantic Theory. Towards a Modern Semantics. (Transl. from spanish by William C. Brown and ed. by Roger Wright.) Oxford: Blackwell
Bar-Hillel, Yoshua (1964). On categorial and phrase structure grammars. In Language and Information. Selected Essays on their Theory and Application. Reading, Mass.: Addison-Wesely.
Beardsley, Monroe (1958). Aesthetics: Problems in the Philosophy of Criticism. New York.
Bühler, Karl (1978 [1934]). Sprachtheorie. Die Darstellungsfunktion der Sprache. Frankfurt/M etc.: Ullstein
Calvo, Clara (1992). Pronouns of address and social negotiation in 'As You Like It '. In Language and literature Vol 1 Number 1.
Cassirer, Peter (1970). Deskriptiv stilistik. Acta Universitatis Gothoburgensis. Nordistica Gothoburgensia 4.
Cassirer, Peter (1975 ). On the place of stylistics. In Style and Text. Studies Presented to Nils Erik Enkvist. Stockholm:Skriptor.
Cassirer, Peter (1977). Studier över ordförståelse. Rapport från projektet SVENSKARNA OCH DERAS ORD. Göteborg. Acta Universitatis Gothoburgensis. Nordistica Gothoburgensia 10
Cohen, Ted (1978). 'Metaphor and the cultivation of intimacy'. Critical Inquiry vol 5 number 1, Automn.
Danesi, Marcel (1994). From the 'Fantasia' through 'Metaphor' to 'Knowledge': A Vichian Perspective of Conceptualization. (Paper read at the Fifth Congress of the International Association for Semiotic Studies, Berkeley, Californa, June 17, forthcoming in the proceedings).
Enkvist, Nils Erik (1964). On defining style. In Enkvist, N.E., Spencer, J. and Gregory, M. Linguistics and Style. London: Oxford Univ. Press.
Enkvist, Nils Erik (1973). Linguistic Stylistics. The Hague etc:Mouton
Garza-Cuarón, Beatriz (1991). Connotation and Meaning. Transl. by Charlotte Broad. Berlin - New York: Mouton de Gruyter.
Geckeler, Horst (1971). Strukturelle Semantik und Wortfeldtheoire. München: Fink
Gray, Bennison (1969). Style - The Problem and its Solution. The Hague -- Paris: Mouton
Halliday, M. A. K. (1976). 'Functions and universals of language'. In System and Function in Language. Selected papers ed. by G.R Kress. London: Oxford Univ. Press
Jakobson, Roman (1960). 'Closing statement. Linguistics and Poetics'. In Style in Language, ed. by Thomas A. Sebeok. Cambridge, Mass.: M.I.T. Press.
Kerbrat-Orecchioni, Catherine (1977). La connotation. 3e édition. Lyon: Presses Universitaire de Lyon.
Leech, Geoffrey (1987). Stylistics and functionalism. In The Linguistics of Writing . Arguments between Language and Literature, ed by Nigel Fabb, Derek Attridge, Alan Durant & Colin MacCabe, Manchester: Manchester Univ. Press.
Lewis, David (1970). General Semantics. Synthese 22 (18-67)
Longman Synonym Dictionary (1986)
Lyons, John (1963). Structural Semantics: Analysis of Part of the Vocabulary of Plato. Oxford.
Lyons, John (1977). Semantics. Cambridge etc. : Cambridge Univ. Press.
Malmberg, Bertil (1959). Nya vägar inom språkforskningen. Stockholm
Melcuk, Igor (1988). Dependency Syntax: Theory and Practice. Albany: State University of New York
Miller, George A. & Johnson-Laird, Philip N. (1976). Language and Perception. Harvard: Harvard Univ. Press.
Novák, P. & Sgall, P. (1968). 'On the Prague functional aproach'. Travaux Linguistiques de Prague 3. (Prague: Editions de l'Academie Tchecoslovaque des Sciences).
Osgood, Charles (1970). 'Interpersonal verbs and interpersonal behaviour'. In Studies in Thought and Language. Ed by J. L. Cowan. Arizona: Univ. of Arizona Press.
Pottier, Bernard (1963 ). Récherches sur l'analyse sémantique en linguistique et en traduction mécanique. Nancy .
Sansome, R. (1986). Connotation and lexical field analysis. In Cahiers de lexicologie. Revue internationale de lexicologie et de lexicographie, vol. XLIX.
Sanders, Willy (1973). Linguistische Stiltheorie. Göttingen: Vandenhoeck & Ruprecht.
Svensk ordbok (1986). Stockholm: Esselte.
Sperber, Dan and Wilson, Deirdre (1986 ). Relevance, Communication and Cognition. Oxford: Blackwell.
Ullman, Stephen (1962).Semantics. An Introduction to the Science of Meaning. Oxford: Blackwe
1 This mathematical term is used as far as I know in linguistics primarily in categorial grammar with a quite different meaning. See Bar-Hillel (1964: 99f.) and Melçuk (1988).
2 'Reference' is often used as having aproximately the same meaning as 'denotation' and 'extension' .
3 The term cognitive is widely used to indicate referentiality on the one hand and 'thought' as contrary to 'feeling' on the other, in order to distinguish 'cognitive meaning' from 'emotive meaning'. Cf Sansome (1986: 13-29) and Lyons (1977: 175).
4 I have tried to demonstrate this on the word introspection in Cassirer 1977.
5 This does not prevent them from having certain style characteristics: the Sw. relative pronoun vilken, for instance, is used only on higher style levels (see below!).
6 See for instance Pottier (1963 : 11-17) and Geckeler (1971: 215ff.)
7 For instance Geckeler (1971: 78): 'Konnotationen gehören, unserer Meinung nach, nicht zur Ebebe des Sprachsystems, und zwar nicht deshalb, weil sie ihnen die intersubjektive Gültigkeit fehlt, denn diese kann auch auf der Ebene der Norm der Sprache fehlen, sondern weil sie nicht die distinktiven Funktionen betreffen'. Cf also Lyons (1977: 175ff.).
8 Cf Jakobson (1960: 354f.): 'The so-called EMOTIVE or 'expressive' function, focused on the ADDRESSER, aims a direct expression of the speaker's attitude toward what he is speaking about. It tends to produce an expression of a certain emotion whether true or feigned; therefore, the term 'emotive', launched and advocated by Marty [in Untersuchungen zur Grundlegung der Allgemeinen Grammatik und Sprachphilosophie, vol. 1, Halle 1908] has proved to be preferable to 'emotional'. The purely emotive stratum in language is presented by the interjections'. The logic in this reasoning is not clear; if the 'emotive function' produces 'an expression of a certain emotion', 'expressive' or 'emotional' would do as well as 'emotive'. Since however we do seem to agree upon that the function transmits the attitude of the speaker, the best term for it must be attitudinal.
9 Longman Dictionary of English Language and Culture, which is the dictionary that, as far as I know, offers the most information about usage and style, lists the following labels for attitude: appreciative, derogatory, euphemistic, humorous (i.e. shows a joking or ironic attitude), pompous (shows a foolishly self-important attitude). This is a good beginning, but those comments are still very rare and they are not sufficient.
10 The reluctance to make non-referential elements of signification the subject of thorough investigation is evident e.g. in John Lyon's major work (1977), a fact which is all the more regrettable since he is considered one of the late twentieth century authorities on the subject. As early as in his thesis (1963: 77) Lyons expresses the hope that, as a consequence of the method of analysis he applies, 'we are no longer tempted to invent rather nebulous "emotive" differences of meaning that can be given no operational significance'. It is therefore not really surprising that, in Semantics (1977), Lyons introduces his discussion of style by referring to Martin Joos; in Joos' terminology, style is reduced to style level, which is defined according to degrees of formality. In this context it is remarkable that, in his renowned closing statement for the 1958 conference on stylistics in Bloomington, Indiana, 'Linguistics and Poetics', Roman Jakobson (1960: 353) characterized Joos' 'emphatic requirement for an 'expulsion' of the emotive elements 'from linguistic science' as a 'radical experiment in reduction--in reductio ad absurdum'.
11 Cf Danesi (1994)
12 No doubt, the metaphorical function is essential here, since if you have only one word for a referent, you cannot distinguish the intensity of the referent from that of the word. In the non-metaphorical use of explosion, it is the explosion that is intense - not the word. But even a linguist might sometimes consider the impact from reality upon words. As an example of 'the contagious reality', one might argue that, even if the word train does not have any feature of rapidity among its semantic components, the train as thing has a great potential of rapidity, and hence there might also be some associative meaning of rapidity in the meaning of the word train. An example of a domain in which reality influences language is the field of social euphemisms. In Swedish there has been a tendency to re-name low-status jobs with fancy terms. Städerska ('cleaning woman') has undergone several upgradings, until lokalhygienist ' ('housing hygienist') became so obviously ironic that it seems to be disappearing! The same goes for friställd (a graceful way of saying laid off implying something akin to becoming liberated) instead of arbetslös, 'unemployed', but these nice words usually either disappear or gradually become 'infected' with the negative value of the referent.
13 Sw. pytte 'minimal' has the same function on the other end of the scale! cf. Sw. ärke-, [ arcos ] 'prince', i.e. first and (fore)most!
14 There are aspects beside evaluation that should also be included in the lexical meaning of a word. If a word that refers to human beings - take lovely as an example - is used more or less only (at any rate prototypically) by and about women, that usage gives a certain feminine 'ring' to the word; this 'ring' will most certainly be considered as highly stylistically relevant, but such facts are very rarely - if at all - mentioned in the dictionary. These facts, called collocational or collocative or distributional meaning, are well known to lexicographers, but they are only indirectly and implicitly communicated to the user of dictionaries, mainly by the sample sentences which sometimes appear to complement the definition. See Sansome (1986: 13-29).
15 Gray (1969) in Style - the Problem and its Solution. Mouton/De Haag. The 'solution' is, that there 'is no such 'thing' as style''! It might be important to point out, that style also can also be defined as rules for writing, for instance punctuation etc, as in Webster's New Collegiate Dictionary. My paper does not deal with this kind of style!
16 For instance Cassirer (1975).
17 'The so-called lexical meanings in certain signs are nothing but artificially isolated contextual meanings, or artificial paraphrases of them. In absolute isolation no sign has any meaning' says Louis Hjelmslev in Prolegomena to a Theory of Language p. 45, quoted from Geckeler (1971: 131). In principle, today's semanticists seem to follow this reasoning: Jean Aitchison gives an insightful characterisation of the situation in Words in the Mind (1987). Nevertheless, I have my doubts about the consequences of Hjelmslev's point of view. While I thoroughly agree with the hermeneutic paradox which states that the whole is composed of parts which obtain their meaning from the whole, language still has definable units in the form of words which can be organized in dictionaries -- in all their incompleteness. It is important, however, to recognize that even in dictionaries the line of demarkation between system and application/manifestation is transgressed: the examples in dictionaries are in fact text-nuclei, which means that even in a dictionary you have no completely context-free lexicography!
18 The term style level (or register) is used to convey information about the context in which a particular word may be found - or rather: usually is found.The hierarchical notion of 'level' is a tradition from classical rhetoric, as in Rhetorica ad Herennium, and was formalized in the 'rota Verigilii' in the Middle Ages. The 'high level' encompasses language spoken from the throne and the pulpit, and other communication in formal situations and in communication between people who do not know each other intimately or at all. 'High level' also means a language of the 'upper classes', containing old and sometimes even obsolete morphological, lexical and syntactic structures (e.g. in the legal tradition) as opposed to 'informal' language, which is more often spoken than written, using more modern words and expressions and even slang and is used in communication between people who do know each other - perhaps even intimately. Information about style level is usually the only information about usage and style provided in dictionaries.
19 There is a connection between style level and style value in that, on the whole, we do not find strong negative (derogatory or contemptuous) words on high levels. It is our aim in the project to investigate further the links between style level and value.
20 A definition in these terms is propagated in my thesis Deskriptiv stilistik from 1970. In Sweden 'style' was often defined as a linguistic concept at that time, and the term 'linguistic stylistics'(e.g. Enkvist, Nils Erik, Linguistic stylistics. The Hague etc 1973; Sanders, Willy, Linguistische Stiltheorie. Göttingen 1973) has been used to stress the fact that one does not want to exceed the linguistic level in one's investigations. In the sixties and seventies there was a rather intense discussion about the definition of the concept of 'style'. There has been some controversy between such scholars and others, who, like myself, think that the linguistic levels do not suffice for a total stylistic analysis.
You can nowadays find definitions of 'style' in these terms in most English and American publications, for example in Geoffrey Leech (1987: 76): stylistics studies 'the relation between the form of the text and its potential for interpretation'.
21 Relevancy is a crucial concept in stylistics, but it seems to be very difficult to establish general rules for it. For socalled impressionistic stylistics the problem is not finding relevant traits but arguing for their relevancy. For formalistic and quantitative stylistics the problem lies in establishing the relevancy of its results. Cf Grey (1969: 21): 'The important matters, the matters that will lead to objective proof or disproof, are the matters of deciding what is to be counted, what is the nature of that which is counted, and why it should be counted' and Enkvist (1964: 25): '... there is, of course, no risk of the statistician's putting the linguist and the literary critic out of business: in practice only the linguist and the critic can tell the statistician what features are worth counting in the first place'. Apart from creating a vicious circle Enkvist could be asked, what findings of interest the statistician could ever make then, that the 'linguist' or the 'literary critic' has not already made before him and without the statistic apparatus - considering that we do not just want to make a contribution to distinctive stylistics as for instance in authorial attribution.
22 This fact, so sadly disregarded or repressed in most so-called linguistic stylistics, nevertheless pops up in most definitions of the concept. See for instance Lyons (1977 : 107).
23 Meaning is scalar: meanings can be prototypical, inherent or only potential.
24 See footnote 7 above!
25 These evaluative meanings are often labeled 'co-meaning' (bibetydelse) in Svensk ordbok.
26 There is good reason to be aware of the ambiguity of this term: 'The reader should be on his guard whenever he meets the term 'connotation' in semantics' says John Lyons, (1977: 176). It was John Stuart Mill who introduced the term in semantics (Lyons 1977 : 175f.). For two modern overviews cf Catherine Kerbrat-Orecchioni, La connotation (1977), and Beatriz Garza-Cuarón, Connotation and meaning. (1991).
27 In Ullman (1962: 128) the part dealing with this problem is titled 'Emotive overtones'.
28 The subtle distinction between you and thou/thy in the paragraph quoted shows that there was an even greater range of variety in the 17th century than now. See (1992 : 5-37).
29 Even if vulgar might have been less derogatory in Shakespeare's day than it is today, the negative attitude in all three words is quite obvious. (Boorish is etymologically related to the germanic boer, 'peasant')
30 The Introduction to the Longman Synonym Dictionary (1986) begins with 'Those who work with language know that there is no such thing as a true "synonym".' The results of identity word tests, according to Miller & Johnsson, show that even words as close to each other as midday/noon and asylum/madhouse get no more than 3.94 and 3.04 (of a maximum of 5), respectively, in a similarity rating test (Miller & Johnson-Laird (1976: 254).
The classical question in discussions of synonymy is whether a difference in form always results in a difference in meaning. Many semanticists argue that the difference between two words is never entirely and exclusively formal or conceptual; with a definition of style as a function between form and content style cannot be considered to be an isolated variable, since it is always dependent on both form and content.
31 'Style is detail of meaning or small-scale meaning. [. . .] Where there is either no difference in meaning at all, or else a gross difference, we do not say there is a difference in style; where the difference in meaning is relatively subtle and is present along with some basic similarity on the primary level, we call the difference in meaning a difference in style.' Beardsley (1958: 223).
32 There is often the alternative NOT to mention something, which is a rather pertinent objection to a definition of style as 'choice between items that mean roughly the same': '... stylisticians seek out accounts of linguistic form, of the lexicogrammar of texts in general, which may serve as useful diagnostic instruments in uncovering patterns, tendencies, textual omissions, or silences or foregroundings (in short, style). Toolan (1992: 162).
33 Novák, P. & Sgall, P. (1968: 291-297). 'On the Prague functional approach'. Travaux Linguistiques de Prague 3. (Prague: Editions de l'Academie Tchecoslovaque des Sciences).
34 '... functionalist explanations look for relations beteen language and what is not language, wheras formalist explanations look for relations between the elements of linguistic text itself' Leech (1987: 76). (I would like to change Leech's word formalist into systemic, though.)
35 The chapter following the one about the Organon model, fully developed in Bühler (1978 [1934]), bears the title 'Die Zeichennatur der Sprache' -- 'the nature of sign of language' -- which makes it rather obscure whether Bühler meant 'language' or 'utterance' and if he actually made any difference between the two.
36 Halliday (1976: 26-31). It might be of some relevance that this same argument was used by Halliday in a lecture in Göteborg, October 13th 1994, which indicates that he has not abandoned this view since 1976.
37 It is characteristic for John Lyons that he considers the 'descriptive function of language' to be 'more distinctive of natural languages than is the interpersonal function, which is shared by other human and non-human signalling systems. But this does not mean that the descriptive function is more basic than the other functions.' (1997 p. 56).
38 Cf Osgood (1970).
39 Jakobson (1960: 354).
40 '... the conative function finds its purest grammatical expression in the vocative and imperative' (Jakobson 1960: 355).
41 In linguistics with German connections, Bühler is more acknowledged: cf for instance Baldinger (1980: 212ff.).
42 For example 'Die psychophysischen Systeme sind Selektoren als Empfänger und arbeiten nach dem Prinzip der abstraktiven Relevanz --- und die psychophysischen Systeme sind Formungsstationen als Sender. Beides gehört zur Einrichtung des Signalverkehrs.' (1978: 28).
43 The late Swedish linguist Bertil Malmberg is quite categorical as to the value of Bühler's model:'För varje form av textanalys är Bühlers schema grundläggande' (For any form of textual analysis, Bühler's schema is fundamental') (1959: 237).
44 Cf Sperber and Wilson (1986).
45 The question marks in this table only underscore my point: it has not been possible to really determine these facts from the dictionaries I have consulted!
46 Thanks to Jens Allwood and Jerker Järborg for valuable criticism of an earlier version of this paper and in particular to M.A.K. Halliday for his very special kind of encouragement!