See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/229678358 What Euphemisms Tell Us about the Interpretation of Words Article in Studia Linguistica · November 1992 DOI: 10.1111/j.1467-9582.1992.tb00833.x CITATIONS READS 123 7,779 1 author: Beatrice Cecilia Warren Lund University 44 PUBLICATIONS 1,445 CITATIONS SEE PROFILE Some of the authors of this publication are also working on these related projects: Implicit relations in modifier-head constructions View project All content following this page was uploaded by Beatrice Cecilia Warren on 24 November 2017. The user has requested enhancement of the downloaded file. What Euphemisms Tell Us about the Interpretation of Words.* Abstract: In the last decade or so, linguists of different persuasions have found the assumption that words have fixed meanings unsatisfactory. Word meanings, it is suggested, are dynamic and negotiable. The present article addresses the question of how we go about negotiating word meanings. It is assumed that the meaning of a content word will have some features of meaning which fix reference and connect a word-form with a class of referents. If, in a particular context, we connect a word-form with some referent which we see as a member of a set of referents which does not coincide with the set of referents that we assign the word in question out of context, the class-distinctive features of the members of the contextually induced set automatically become features of meaning and we have created a novel contextual meaning. The question then arises: under what circumstances are we willing to link a word with a novel category of referents. It will be revealed that we accept a word as adequately used provided it implies the new type of referent or constitutes a description of it. I Hypotheses Words uttered in a particular setting may communicate notions they do not mean in the dictionary. Consider: (1) In engineering it is rare to find iron used in its pure form. Generally the metal is alloyed with carbon and other elements to form wrought iron, steels and cast iron. (Quoted from Widdowson1979:96) Metal here must be interpreted "iron"; yet we would not wish to say that "iron" is one of the senses of metal. (2) The man was staggering drunk. He had consumed the whole bottle. The whole bottle here will be taken to refer to "all the contents of the bottle". We would, however, not think of "contents of bottle" as a meaning of bottle. Also, consider and compare the meanings of eye in: (3) My husband´s eyes are brown. (eyes = "irises") (4) My cousin had a black eye. (eye = "area surrounding eye") (5) My aunt had obviously been crying. She had red eyes and a blotchy face. (eyes= "whites of eye") Yet, we would hardly wish to claim that among the meanings of eye there are "iris", "white of eye", "area round the eye". The above examples are offered to illustrate the question which is the basic concern of Fel! Bok märk et är inte defin ierat. this article, namely: if the meaning of a word is the associations which serve to connect it with a category of referents how then can we accept as quite natural that words refer to entities and phenomena that are not members of the categories specified by their meanings. The following hypotheses attempt to answer this question: First hypothesis: There is a difference between the dictionary meaning and the contextual meaning of a word. Dictionary meanings we find in the dictionary and the mental lexicon; they are the meaning or meanings the members of a language community assign a word out of context. The contextual meaning is the meaning the interpreter assigns a word in context. It can, but need not, coincide with the dictionary meaning, which the examples above are meant to illustrate. Therefore, to be able to interpret words, it is not enough to know their dictionary meanings. One must also know how to construct their intended communicative value when they are used. One is guided in this by a number of clues, among which dictionary meaning is one. The important points of this hypothesis are: (i) Dictionary meanings are by definition out-of-context meanings. They can never be part of either sentences or utterances, since, as soon as we are to assign a word a value in any kind of context, we contextualize it. (ii) Interpreters have an open mind as to what words in context communicate. All they expect is that there will be some obvious connection between the notion the word in question is intended to convey and its dictionary meaning. Second hypothesis: In agreement with the traditional view that content words which are not proper names with unique reference represent categories of non-linguistic entities or phenomena, it is assumed that meaning-construction involves retrieval of a set of referents. Moreover, it is assumed that the interpretation of a content word in a context involves finding a referent or some referents for it. Therefore, if we see the particular referent we have retrieved for a word (in a context) as a member of a category which differs from the out-of-context category, we have formed a novel contextual meaning. Third hypothesis: A novel contextual meaning of a word will not be committed to memory as a matter of course. The new meaning must be considered to be of lasting value. This can occur if the novel category with which it is connected is seen as a stable category the members of which we repeatedly wish to refer to and for which there is no other "name". It may, however, also occur even if there is an established "name" for this category, provided the new "name" is seen as a desirable alternative "name". (Consider euphemisms and slang senses.) In other words, it is postulated that we do not only contextualize dictionary meanings but that we also decontexualize contextual meanings. That is to say, although we may consider a meaning of a word as fully warranted by a particular context and as the most plausible in that context, we may nevertheless dismiss it as being purely ad hoc, ie not a proper meaning. Consider again the contextual meanings of metal, bottle and eye in the examples above. The decontextualization process could be seen as a process preventing the mental lexicon from being overloaded. Recapitualing the above, it has so far been hypothesized that established dictionary meanings may be transformed to yield novel dictionary meanings provided the following conditions are fulfilled: (i) A word is matched with a referent (or some referents) deemed Fel! Bok märk et är inte defin ierat. Fel! Bok most plausible in the context at hand and this referent (or these referents) are thought of märk et är as a member (or members) of a category which is not the category of referents conventionally associated with the word in question. (See, however, footnote 1.) (ii) The inte defin interpreter will have accepted this new contextual meaning because there is some ierat. connection between the established and novel meaning; or, more precisely, because there is some connection between the referents in the two categories. (iii) For some reason, the novel contextualized meaning has become memorized and gained general acceptance. The above brings us to the fourth hypothesis, which is: It is possible to distinguish between different types of connection which will yield different types of sense change. In fact, it is postulated that there are the following types of connection: I Particularization: The new contextual referent is a member of a set which is a subcategory of the conventional category of referents of the word in question. Examples include yellow card "warning card in soccer", irradiate "to treat food with gamma radiation in order to preserve", growth "tumour". In other words, the contextual referent is named by its correct name. A yellow card is a card which is yellow. However, it is not any yellow card, since yellowness is not its only defining feature. A more salient feature would be its function, viz that of being used by a referee to warn football players that they have violated a rule. This functional feature need not be explicitly stated. It is sufficient that the expression yellow card enables the interpreter to retrieve the intended referent. Having retrieved it, (s)he will then naturally perceive or infer the existence of this defining feature which will cause yellow card to assume a meaning hyponymic to the unidiomatic meaning of yellow card. II Implication: The contextual and the conventional referent(s) are invariably or frequently concomitant, which causes a more or less probable antecedent-consequent relationship between them, ie if X (old referent), then (probably) also Y (new referent), for example: hang up ("end a telephone conversation"), go to the toilet ("urinate and/or defecate"), sleep with somebody ("have sexual intercourse with somebody"). III Metonymy: There is a co-occurrence relationship between contextual and conventional referents. For instance: body "type of leotard", ie that which (novel referent) goes round (relation) the body (conventional referent); ecstasy "amphetamine", ie that which (novel referent) causes (relation) ecstasy (conventional referent); skirt "woman", ie that which (novel referent) has (relation) a skirt (conventional referent). Metonymy and what I call implications are sometimes confused. There are, however, significant differences between these two types of meaning extension, as will become evident. IV Metaphor: Some property of the conventional referent is also a property of the contextual referent. For instance: mole "secret agent" (both moles and secret agents work under cover); mousse "hair cosmetic" (both the dessert and the cosmetic have foamy consistencies); egg "head" (both heads and eggs have oval shapes). V Reversal: The conventional meaning of a word fits our favoured contextual referent provided we reverse it, so that huge means (contextually) "unusually small" or early "late". VI Understatement (alias litotes): The conventional meaning of a word fits our favoured contextual referent provided the degree to which some feature of meaning applies is boosted. For example: drug habit "drug addiction", plump "fat". VII Overstatement (alias hyperbole): The conventional meaning of a word fits our favoured contextual referent provided the degree to which some feature of meaning applies is attenuated. Consider for example: very, which used to have the stronger emphatic sense of "in truth" (cf. verily). It was further hypothesized that reversals and under- and overstatements would typically be purely rhetorical devices and so only rarely result in novel dictionary meanings. Since a certain type of euphemism involves the creation of a novel contextual sense, analyses of euphemisms of this kind should throw light on how the interpretation of words in context can bring about novel senses. Therefore the above hypotheses were tested on a corpus of 500 euphemisms. The results of this investigation will be discussed in fair detail in section III.2 II Euphemisms There are four main ways in which euphemisms may be constructed: (i) The word formation devices of the language in question are made use of. Examples include comfort station -compounding, sanguinary "bloody" -derivation, SAPFU "military blunder" -acronym of surpassing all previous fuck-ups, etc. (ii) Foreign words are imported: lingerie from French, calaboose "jail" from Spanish calabozo, sativa "marijuana" from cannabis sativa, the Latin name. Classical loans are particularly favoured, since they imply learnedness and matter-of-factness and so elevate "the tone" of the word. Consider: anus, rectum, fæces, urine. (iii) The form of the offensive word is modified, or altered according to certain rules: divil,divel (instead of devil), Gad , Gosh, Golly (instead of God), fug (instead of fuck), epar (back slang for rape), elephant and castle (rhyming slang for "arsehole"). (iv) A novel sense for some established word or word combination is created: growth "tumour", go to the toilet "defecate and/or urinate", extra-curricular activities "adultery", ecstasy "amphetamine"; infelicities "blunders", sanitary engineer "garbage man", blessed "damned". Fel! Bok märk et är inte defin ierat. Fel! Bok märk et är Combinations of these devices may occur. Consider, for example, effing "having sex", inte which is formed from F or eff, an abbreviation of fuck, and -ing, a suffix, ie we have a case of abbreviation + suffixation; or J/jay "marijuana cigarette", which is probably short defin for joint, which I have taken to be a particularization of the sense "that which is joined ierat. to something", ie we have a semantic disguise combined with abbreviation; or nards "testicles", which is a phonological modification of nuts/nerts, which in turn is a metaphor, ie "these things which have shapes reminiscent of nuts". Classical morphemes are often borrowed and then composed into words: sanguinary "bloody", perspiration , for instance. Besides these four main devices, there are others, which, however, are very much less common. The offensive word can simply be omitted3 or replaced by some unarticulated noise in speech (Where is the ehum?) We sometimes come across descriptions of the form of the word that we wish to avoid (as when monosyllable stands for cunt) or of the function of the offending word (as when qualified replaces damned, fucking, etc in, eg, He is a qualified fool. According to Partridge (1984: 945), this usage possibly originated with Kipling). There are probably also disguises that have been chosen for no real reason. I suspect that little Willie for "penis" may be an example of this. Finally, there are occasional examples of where the connection between the novel and conventional referent is none of the ones enumerated above. As an example, I may adduce key when it is used to denote "penis" since, allegedly, "it lets the man in and the maid out". Note that, unlike the regular types of relations, such types are not inferrable but need to be explicitly explained. The four main devices may be said to yield either a new form, ie a sequence of phonemes or morphemes not previously used in the language in question4, or a new sense for an established form. Thus we would have the following classification of the main devices for constructing euphemisms (see next page): This classification is more detailed and exhaustive than the previous classifications of euphemisms that I am aware of6, but agrees well with these in that it omits nothing included in these. Of the four main devices, only euphemisms representing semantic innovations are of interest to the issue at hand. Consequently, only these kinds of euphemism were included in the corpus of examples of the present investigation. The examples were excerpted from two dictionaries, viz Spears (1981): A Dictionary of Slang and Euphemisms (400 examples) and Neaman and Silver (1983): A Dictionary of Euphemisms (100 examples). The appendix contains an excerpt of 100 examples of the corpus. This excerpt was originally the basis of a pilot investigation. All the examples in it derive from Spears. Fel! Bok märk et är inte defin ierat. compounding (i)word-formation devices derivation blends acronyms, etc onomatopoeia back slang Formal innovation (ii)phonemic modification rhyming slang phoneme replac. abbreviation (iii)loan words particularizations euphemism s implications metaphors (iv)semantic innovation metonyms reversals understatements overstatements Fel! Bok märk In order to collect examples of euphemisms, some idea as to what words qualify as such et är inte is naturally necessary. The definition I have adhered to is the following: we have a euphemism if the interpreter perceives the use of some word or expression as evidence defin of a wish on the part of the speaker to denote some sensitive phenomenon in a tactful ierat. and/or veiled manner. There are three parts to this definition: (i) The referent is considered a sensitive phenomenon. This explains why it is possible to discern favoured topic areas, such as death, crime, unpopular political and military facts, drug abuse, physical and mental defects, unpleasant bodily waste products and sex (and sex and sex). (ii) The referring expression is thought of as less harsh or coarse and/or less direct than some alternative. This in turn implies that euphemisms tend to name phenomena which are already named (This is confirmed by the present investigation. Of the 500 examples, there are only two that were judged to have otherwise unnamed referents, viz customer´s man, "man who gathers bets for the bookmaker" and bustle pinching "practice of a man rubbing his penis against the buttocks of a woman"). (iii) It is the interpreter's perception that the speaker's choice of expression is dictated by considerations of tact or embarrassment with the referent that determines whether an expression is a euphemism or not. Granted this point, it follows that what is a euphemism "is in the eye/ear of the beholder" and cannot strictly speaking be objectively verified, although normally of course there is consensus among language users as to what words are euphemistic. In other words, in my view, it is not possible to present a definition of euphemism that will act as a shibboleth. Therefore I do not expect 100% consensus that the examples in my corpus are in fact euphemisms. Note also that my "definition" allows as euphemisms constructions which are not necessarily delicate in tone, but which nevertheless can be taken to be evasive expressions constructed out of awareness that the referent in question is taboo. These, then, would be what Neaman and Silver (1983:IX) call "vulgar euphemisms". Euphemisms may be purely contextual, conventional or dead. Purely contextual euphemisms are nonce-formations; conventional euphemisms are euphemisms which originally were contextual but which have established themselves and become dictionary meanings; dead euphemisms are words still in use in the euphemistic sense, but which have lost the euphemistic force they once had. Examples of the latter include disease, idiot, silly, accident. Dead euphemisms should not be confused with obsolete euphemisms, which are euphemistic words or euphemistic meanings no longer in use. Needless to say, transitions from one type of euphemism to another are gradual and there are examples of euphemisms which are on their way to becoming conventionalized, dead or obsolete. The great majority of examples in the present investigation are naturally conventional. There are, however, a fair number of obsolete euphemisms and three that I consider to be dead, viz insane, gay "homosexual", and casualty. It is often suggested that loss of euphemistic force is induced by frequent use. That is to say, it is thought that if a euphemism is used so frequently as a designator of a taboo phenomenon that strong associations are formed between this word and the taboo referent, its euphemisitic force will disapppear. This explanation implies that euphemistic force is simply a matter of the strength of the associative link between a Fel! Bok märk word-form and its taboo referent, ie the stronger the link, the weaker the euphemistic force and vice versa. However, euphemistic force depends primarily on contrast. In the et är inte case of semantic innovation, the contrast is between the basic and the novel sense. defin Consider, for instance, disease, which originally meant "discomfort". Once this sense ierat. disappeared, we can hardly think of disease as a litotes and a mitigating way of talking about illness. In the case of formal innovations, the contrast is between the common, vulgar term and the learned, elevated term (consider piss and urinate, cock and penis) or between the explicitly said and the implied (consider God and Gosh, fuck and eff.). As long as the two alternatives exist, euphemistic force can survive even great frequency of use. In fact, in my corpus there are some very old and well-established euphemisms. For instance, according to Spears, loose was first used in the euphemistic sense "sexually loose" in the late 1400s, and void has occurred in the senses "vomit", "urinate/ defecate", "blow one's nose" since the 1400s. Nevertheless, frequency is most probably a factor involved in the weakening of euphemistic force. Once a particular euphemistic sense of a word has established itself, it can cause avoidance of this word in any other sense so that the euphemistic sense ousts all other senses and in this way causes its own death as a euphemism. Insane probably lost its sense "not healthy" in this way and so also its status as a euphemism. According to Spears and the OED, gay derived its sense "homosexual" from "sexually frivolous", a dictionary meaning I do not associate with gay - hence in my mental lexicon gay is neither a euphemism, nor is it one of the senses of gay "happy", but rather a homonym of gay "happy". Casualty defined by Neaman and Silver as "dead or wounded in a war" has, I have suggested, lost much of its euphemistic force. My explanation for this would be that casualty no longer is used in any other sense, at least not if I may modify the definition to be "dead or wounded in some calamity, esp war". However, because of its vagueness ("dead or wounded"), it is still more euphemistic than dead or killed. In this connection, let me mention two examples in my corpus from Spears which demonstrate the point just made, ie once a euphemistic sense has established itself, this may render the word itself taboo. The two words are tap for "nut that secures a bolt" and weathervane for "weathercock". Neither of the referents of these words is taboo, but their original designators (ie nut and cock) are because of their senses "mad" and "penis", respectively. These examples, then, represent a special and rare "metalinguistic" type of euphemism, ie constructions which replace taboo words, not real-world taboo phenomena. III Results and Discussion The distribution of my examples among the different groups was the following: Particularizations Metaphors Implications Metonyms Understatement Overstatements 208 146 51 28 10 5 Reversals Cases of analogy* Cases of double classification* Redefinitions * Questionmarks* Metalinguistic euphemisms 2 4 32 3 9 2 * These groups have not yet been introduced to the reader but will be discussed below. Particularizations As we see from the statistical results, a frequently used device for creating euphemisms is to use a general term which in its context has to be particularized to make sense. The relation between the out-of-context set of referents and the set of referents connected with the euphemistic sense is that of proper class inclusion, ie the referents of the euphemistic sense must be a subcategory of the referents of the dictionary sense from which it derives. It is possible to discern four main types of particularization among my examples. In the case of verbs, certain adjectives and deverbal and relational nouns, specification of the sense typically involves retrieving some particular argument(s): Verbs dump heave raise scale weaken without emission Adjectives innocent Deverbal nouns automanipulation kiester stash "defecate" (feces) or "vomit" (contents of stomach) "vomit" (contents of stomach) "get an erection" (penis) "mount a woman" (ex the sense "climb up") "lose an erection while copulating" (penis, semen) "being a virgin" (sexual experience) receiver "playing with one's genitals" "stash of illegal things in the rectum" (kiester ="rectum") "partner in fellatio" (penis) Relational nouns end parts below "foreskin" (of penis) "genitals" (of people) In the case of verbs, adjectives and deverbal nouns, specification of meaning may also involve retrieving the manner in which something occurs or is enacted: Fel! Bok märk et är inte defin ierat. Deverbal nouns adjustment to the front assault cross-cultural communications satisfaction "retreat" "rape" "persuasion of others to think favourably of one's own country" "orgasm" Verbs enjoy a woman lose a meal offend "have sex with a woman" "vomit" "smell from perspiration or bad breath" Adjectives nasty odd "smutty" "homosexual" In the case of nouns -in particular in the case of single nouns- meaning-specification frequently simply involves retrieving a relevant subcategory: Nouns action business butch chum curlies drug engine gum house partridge red-light house sand-box tampon virtue (of women) "sexual intercourse" "sexual intercourse" or "defecation/urination" "virile homosexual" (ex "tough man") "prison inmate"(ex "roommate") "pubic hairs" "dope" "equipment used by opium smokers" "opium gum" "brothel" "prostitute" (ex "woman, girl") "brothel" "cat toilet" "vaginal tampon" "chastity" Sometimes it is the subcategory of an argument, whether it is implicit or explicit, that has to be specified. Consider: anti-personnel weapon beverage room: business girl curiosa sanitary paper "weapon against enemy soldiers and populace" "beer parlour" "prostitute" (prostitution) "pornography" (ex "books dealing with unusual matters") "toilet paper" Fel! Bok märk et är inte defin ierat. Needless to say, there are examples of particularization which involve combinations of these manœuvres: lose parts behind occupy Behavioral Skills "menstruate" -In this case specification involves retrieving the argument "blood"+ the manner in which this loss of blood occurs. "posteriors" -Specification involves retrieving the argument "of people/animals"+ relevant subcategory. "take sexual -Specification involves retrieving the argument possession "woman"+ manner of taking possession. of a woman" "remedial - Specification involves Training (BEST) training in retrieval of argument"sailors in US Navy" US Navy" +manner of training. The interpreter will have to retrieve information of the kind exemplified above from his general knowledge of the world and/or the context at hand. In choosing missing bits of information, (s)he will be guided by the requirement that the end result must be a referent or some referents which fit the context. If this referent or these referents are perceived as members of a category, the word serves as a category label and a meaning is formed. If this meaning is conventionalized, we have a new dictionary meaning of the word in question. The particularization process is of paramount importance in semantic theory. It does not only explain how the meaning of established word-forms may become specialized. It can also serve to explain how novel coinages (compounds, deverbal nouns, denominal verbs, phrases of different kinds, etc) acquire their idiomatic meanings, ie meanings which are not equivalent to the sum of the parts and the relations between them. Let us consider but a few examples: blackbird: As we all know, not all birds that are black are blackbirds; nor are all blackbirds necessarily black. This is because, although the compound is normally an accurate description of the set of entities which forms its referents, it is not a definition. When we coin words or phrases which we wish to function as designators for classes of entities or phenomena, all we have to produce is some expression which enables the interpreter to associate to the intended set. Once the interpreter feels confident that (s)he has retrieved the class aimed at, (s)he will turn the knowledge that enables her/him to refer entities or phenomena to this set into the meaning of the expression. The meaning is formed the moment the expression is matched with a particular set of referents. porch: This is a denominal verb, a frequent collocate of which is newspaper. (The example derives from Clark & Clark (1979).) Filling in the implicit relations between porch and newspaper and adding the semantic element which makes porch into a verb, we get the paraphrase "cause newspaper to be on porch". However, the meaning of porch is more specific and is better given as: "deliver newspapers at subscribers' Fel! Bok märk et är inte defin ierat. Fel! Bok porches". This meaning derives from the fact that (i) delivering newspapers by throwing märk et är or otherwise causing them to be on the porches of the subscribers is a recurrent and inte nameworthy event in American life and (ii) "causing newspaper to be on porch" defin sufficently clearly suggests just this category of events. ierat. print-out. This is a deverbal noun. Adding the semantic element which renders print-out a noun, we get the paraphrase: "that which is printed out". However, according to the dictionary, the meaning of this derivative is: "the printed output of a computer". Again. we find a meaning which has been formed by some interpreter succeeding in matching a descriptive expression with a particular category of entities. co-pilot. The meaning of co- is given by Quirk et al (1987:I.24) as "jointly", "on equal footing". Hence the meaning of co-pilot ought to be "sb who is a pilot jointly and on equal footing (with some other pilot)". The dictionary, however, defines co-pilot as "the secondary or relief pilot of an aircraft". Again, we witness that the meaning of an expression ultimately depends on the features of the referents that determine category membership and not on the dictionary meanings of the morphemes which form the expression. The particularization process also accounts for what Stern considers to be fluctuations, ie, for instance, the different "values" we give red in red sky, red hair, red car, red cheeks. I see all these "values" as particularized contextual meanings of the dictionary meaning of red. As always, interpretation means matching an expression with a referent, which in turn may involve particularization. Note that in each case the decontextualization process restores the original, "wide" dictionary meaning of red. One and the same expression may of course conceivably serve as a label for more than one category. Consider print-out, eg. However, once an expression has acquired a particular established meaning, using that expression as a label for another category which is in no way connected with this is normally avoided. If this convention is broken, the effect may be something of a pun. The surprise is produced by the fact that one has to match an expression in a particular context with a class of referents quite unrelated with the set one is used to associating with the expression in question, but which it equally well picks out. Examples from my corpus include: Established meaningparticularization cut out to be a gentleman "destined to be a gentleman" Novel meaningparticularization "circumcised" make a deposit "put money in a bank" "defecate" one-finger exercise "piano excercise involving one finger" "digital stimulation of the genitals of a woman" take one's drops "take one's medicine" "have a drink of alcohol" uplifting "elevating" "sexually arousing" In studies of euphemisms, the device of using a cover term for the offensive one is frequently mentioned. (See Greenough & Kittredge (1902), Wellander (1923), Stern (1931), Williams (1975)). In studies of semantic change, however, particularization is not given much attention, which seems inconsistent. True, Paul (1920) includes "specialization of meaning" in his classification of semantic change, but fails to make a clear distinction between effect and meaning-changing process. Stern does discuss particularizations, but devotes only a couple of pages out of his 450-page treatise to this phenomenon. This cavalier treatment of particularizations contrasts with my insistence that it is the most basic meaning-creating mechanism involved in the interpretative process. Implications As already stipulated, if a novel contextual sense is an implication, the connection between the conventional and novel sets of referents is that of an antecedent to a consequent (if X is valid, then Y is (probably) valid too). The conventional sense most frequently represents the antecedent and the novel sense the consequent as in the following examples: Established sense if New euphemistic sense then auntie 2 (slang –> slang) "elderly homosexual male" "elderly homosexual male past his prime period of desirability" black out (NS)* (theatre & film) "remove light from" "make disappear" go home in a box (NS) "be transported home in a coffin" "be dead" loose 1 (of person) "unattached" "sexually loose" inner-city "the inner part of a city" "ghetto, slum" *NS signifies that the example is excerpted from Neaman & Silver. In some few cases, the conventional sense appears to represent the consequent and the novel sense the antecedent: bend an elbow (NS) Established sense then New euphemistic sense if "curve one's arm" "drink" groaning "moaning" have a good stomach (NS) "have a stomach that causes no trouble" "giving birth" "eat voraciously" Fel! Bok märk et är inte defin ierat. Fel! Bok märk Fairly frequently, the conventional sense could equally well be consequent and antecedent et är (and the novel sense antecedent and consequent): inte defin ierat. Established sense if/then New euphemistic sense then/if do one's bit (NS) "do one's duty" "die in service" bush ranger (NS) "sb who ranges the bush" (Australian English) checked out (NS) "discharged from hospital" crack the claret jug (NS) "crack the nose" (claret jug ="nose") crepe hanger1 (NS) "person on which crepe hangs" you are showing your form (NS) wear the apron (NS) high "you're showing what the shape of your body is like" "to have one's apron high" "escaped prisoner" "be dead" "have a nose bleed" "person in mourning" "you're revealing your naked flesh" "be pregnant" As in the case of euphemisms, in discussing implications it is important to distinguish between purely contextual, conventionalized and dead implications. Again, purely contextual implications are nonce senses. The difference between conventionalized and dead implications is that in the case of conventionalized implications, the "truth conditions" of the conventional (implying) as well as the novel (implied) senses must apply, whereas in the case of dead implications, it is only the truth conditions of the implied sense that need to be met. Inner city with the implied sense "ghetto, slum" can serve as an example of a conventionalized implication, at least if I am correct in assuming that it can only be used appropriately about slums in the centres of towns and not about slums in suburban areas. Hang up "end a telephone conversation" can serve as an example of a dead implication in that it can be used to convey this sense even if a telephone conversation is ended by somebody putting down the receiver. Intuitively we think of both implications and particularizations as literal senses. There is nothing figurative about using sand-box in the particularized sense "particular type of box containing sand, serving as "a cat-toilet""; nor about using inner city in the sense "slum". I account for this by the fact that in neither case there is any violation of defining features. Particularizations are different from implications in that when we particularize, we replace the general dictionary sense with the specialized contextual sense. We cannot by sand-box in an utterance mean both "any sand-box" and "certain type of sand-box serving as a cat toilet". We can by inner city mean both "centre of city" and "ghetto, slum", since, in the case of purely contextual and conventionalized implications, we have the interesting phenomenon of two senses being carried simultaneously by one linguistic unit in an Fel! Bok märk utterance and yet there is no pun intended. If it is general knowledge that a likely reason for not being at one's desk in an office is that one has to go the toilet, then the message that et är somebody is away from her/his desk will not only carry the information that the person in inte question is away, but also where (s)he is. Implications are vague since the interpreter can defin only conclude from circumstantial evidence whether they are intended or not. The cunningierat. speaker can therefore use words in such a way that they invite the interpreter to believe that they carry intended implications and then later insist that they did not. This is possible since in the case of implications it is the conventional meaning that is considered the proper contextual meaning. In this respect implications are different from other types of non-coded meanings, which invalidate the coded meanings they are derived from and so are considered the only proper contextual meanings. Particularizations and implications do not only differ in that implications, unlike particularizations, can be secondary senses. Another difference is that in the case of particularizations, the novel referents must be properly included in the conventional set of referents. This need not be so in the case of implications. Slums are not necessarily city centres. These two differences can be used to distinguish between particularizations and implications. However, since they are not necessarily mutually exclusive, they do not constitute infallible criteria and so there are examples in my corpus which could equally well be classed particularizations as they could be classed implications, eg odd "homosexual". If one is homosexual, one is odd - hence, there is a plausible antecedentconsequent relation. However, homosexuals would by many be seen as a proper subcategory of odd people. Further, it is not inconceivable that in a context one could by odd mean both "peculiar" and "homosexual". Since implications are literal senses, it should be unproblematic to distinguish between metaphors and implications. The existence of dead implications, however, complicates matters. If one blacks out something by removing light from it on stage or in shooting a film, one makes it invisible. Therefore black out can have the implied sense "make disappear". It is of course also possible to use black out as a figure of speech - "to do that which is like blacking out in the film studio or on stage in that one makes something disappear". If we find black out used in the sense "make disappear" without there being any removal of light involved, there are then two possibilities: either we are dealing with a dead implication or with a figure of speech. I have not been able to solve this dilemma. In choosing between implication or metaphor I have mainly trusted intuition, but I have tended to class those cases as implications which allow the non-coded sense to be valid, irrespective of whether the coded sense is valid or not. Unlike particularizations, implications have been observed and fairly extensively discussed by linguists concerned with semantic change (and also grammaticalizations). Sometimes, in particular in earlier literature, implications were thought of as cases of metonymy. Stern was probably the first to use a distinctive term, viz permutation. As is well known, Grice (1974) coined the terms implicatum (for the implied sense) and implicature (for the Fel! Bok meaning-creating process). Since neither Stern's, nor Grice's terms quite cover what I havemärk et är here described, I have not made use of either of these. inte defin Metaphors ierat. If a word is applied to referents which are not in its conventional class of referents, it is felt to have a non-literal meaning. If the justification for this switch of referent set is that the new type of referent has at least one property in common with the old type, we have a metaphor. This represents the traditional Aristotelian view of metaphors, a view which is still accepted by many linguists, even if not by all. (See, eg, Cooper (1986)). These criteria hold true for the 146 metaphors in the present investigation and also for the 450 examples of metaphors in a previous investigation (Warren to appear). I therefore see no reason for abandoning them. We form meanings for metaphors basically in the same way as for particularizations, ie once we have retrieved the novel set of referents, the knowledge which enables us to refer entities or phenomena to this set becomes the meaning of the metaphor. However, the property or properties shared by the conventional and the novel referent will not only serve to justify the switch from one category of referents to another. They also play a part in the meaning-forming process either by highlighting a particular part or particular parts of the definition for the new referents or by expressing the coiner's attitude to the referent, in which case they form evaluative features of meaning. (For an explanation of the term evaluative features of meaning, see footnote 1.) Of the meaning-creating processes described here, none has received as much attention as metaphorization. Indeed, it may even be true that there is no single linguistic phenomenon that has been as extensively discussed as metaphors. There are of course several reasons for this. One may be that metaphorization is more conspicuous than particularization, implication and often also metonymy. Another may be the intriguing character of metaphoric meanings. This in turn is probably caused by the fact that there are often several properties that could connect the two referent sets and so the interpreter cannot be certain that (s)he has retrieved the intended one or ones. This also means that by considering the meaning of a metaphor, one may discover aspects which had not been immediately apparent. All this contributes to the elusively suggestive character of metaphoric meanings. Consider, eg, artillery in the sense "equipment for injecting drugs". Guns and syringes are elongated objects which emit something through a hole with suddenness and great impact. But since the coiner chose artillery rather than gun, (s)he may also wish to express the view that the injection of a drug has as violent and overwhelming an effect as the firing of heavy artillery guns. Or, consider dumplings for "breasts". Is it the shape, the firm but yet soft consistency or the colour of a woman's breast that are highlighted by this metaphor? Does it convey an appreciative or derogative attitude? Fel! Bok Metaphors may also surprise and at the same time delight an interpreter in that they maymärk suggest unexpected parallels which nevertheless strike one as very apt. Consider in this et är inte respect the following examples from my corpus: defin ierat. Established sense Metaphoric sense addict (NS) baby (NS) "sb addicted to alcohol/drugs" "person who allows himself to be repeatedly conned" Metaphor "infant" "marijuana" Metaphor (Same coming-backfor-more habit) (Tenderness- lack of strength) Marijuana is a mild drug. backfire (NS) "the event of a car's incomplete combustion of fuel" "break wind" (Similarity as to sound) big brown eyes (NS) "organs of sight which are brown" "nipples" (buzz) the brillo "female pubic hair" "steel wire" Metaphor Metaphor (Same ring within ring pattern) Metaphor (Both consist of a mass of strands) headlights (NS) magic wand "front lights of cars" "breasts" Metaphor (Same shape and degree of noticability) "wand used for conjuring" "(erect)penis" Metaphor (Shape+the magic of erection) pardon my French "the French language" "swear or taboo words" Metaphor (?Both sound nasty) parenthesis "brackets" "female genitals" (sic!) parsley Metaphor (Similar shape) "herb" "pubic hair" Metaphor (Similar shape) talk German "utter German words" "break wind" Metaphor (Similarity as to sound) waterworks "system of pipes and and tanks of water" "urinary organs" Metaphor (Same function, ie both are systems for transporting liquid) Metaphors may be more or less expressive, more or less imaginative, more or less evaluative. When paradise is used in the sense "female genitals", the coiner has clearly expressed a point of view, but when curl paper occurs in the sense "toilet paper", or envelope in the sense "condom" or reef in the sense "marijuana cigarette" (the shape of the cigarette is presumably compared to the shape of a rolled up reef sail), the tone is more matter-of-fact. As we see, metaphorization is a very flexible instrument for creating meanings. Perhaps the most important aspect of metaphorization, however, is its ability to supply a "name" for the otherwise inexpresssible. Since euphemisms generally are alternative expressions for already existing "names", this particular feature of metaphorization is, however, not apparent in the present investigation. Metonyms Fel! Bok märk et är inte defin ierat. If a word is applied to referents which are not in the conventional set of this word but which are connected with the conventional referents in one of the ways listed below, we have a metonym: Causal relation: The conventional sense represents RESULT The metonymic sense represents SOURCE Paraphrase Word Ashes "cinders" "marijuana" "that which produces ashes" heartburn (NS) "burning sensation in the esophagus" "jealousy" "that which causes heartburn" red-comb "comb of a cock that is red" "sexual arousal" "that which causes a (cock's) comb to go red" scar "wound" "heroin" ?"that which (the use of which ) causes wounds" The conventional sense represents SOURCE The metonymic sense represents RESULT Paraphrase nature 1 "the force that regulates the universe" "the menses" "that which nature enforces" nature 3 "the force that regulates the universe" "the genitals" "that which nature provides" nature 5 "the force that regulates the universe" "libido" "that which nature enforces" red cross "international relief organization" "morphine" "that which the Red Cross supplied" Whole-Part relation: The conventional sense represents PART The metonymic sense represents WHOLE Paraphrase dress for sale "dress that one may buy" "prostitute" "sb who has a dress and is for sale" indoor plumbing "plumbing which is indoor" "that which contains plumbing" Panama red "red colour from Panama" Tan "skin browned by the sun" "mulatto" "toilet" "a potent marijuana" "that which is red and from Panama" "sb who has a tan" Trident * The conventional sense represents WHOLE The metonymic sense represents PART "type of submarine" "missile designed for this submarine" The conventional sense represents PLACE The metonymic sense represents OBJ Locative relation: bathroom (NS) groin "room with a bath" "the fold where abdomen joins either leg" Fel! Bok Paraphrase märk et är inte "that which goes with Trident" defin ierat. Paraphrase "WC" "that which is in bathrooms" "pubic area" "that which is at the groin" Johnny "the penis" "condom" "that which goes on the Johnny" main vein "the most important blood vessel" "the female genitals" ?"that which is where the most tail (NS) red lamp important blood vessel is" "caudal end" "male or female genitals""that which is at the tail" The conventional sense represents OBJ The metonymic sense represents PLACE Paraphrase "brothel" "the place which has a red lamp" "lamp which is red" Equative relation: The conventional sense The metonymic sense Paraphrase represents CONSTITUENT represents WHOLE MATTER/CONSTITUENTS killer-weed 1 silver * "weed which is a killer" "precious metal" "PCP"( a drug) "cutlery" "that which consists of weeds (and kills (pains))" "that which is made of silver" *Trident and silver do not occur in my corpus of euphemisms, but derive from Warren (to appear). They are included to demonstrate that the relations in question do occur. These relations do not only occur as connecters between established and novel sets of referents in the case of metonyms. Together with resemblance relations, they are also regularly employed as covert connecters in noun-noun compounds (bullet hole, nosetip, lighthouse, golf season, silver spoon, sponge cake), in adjective-noun combinations (vocal noise, noisy children, facial skin, furred animals, polar bear, civic group, golden hair), in genitive constructions (Dickens's novels, John's leg, John's place, John's picture, he got his father's eyes), between the noun contained in denominal verbs and its object or subject (calve "(of cows) produce calf", label something "make label be part of something", shelve books "make books be on shelves", father a child :"be the father of a child", mother a person "act like a mother to a person".) There is ample linguistic evidence that equation, composition, possession, location (in time and space), causation and comparison form a set of basic covert default relations. That is to say, they are relations which we choose among to form a meaning for linguistic units consisting of two nouns, unless context explicitly prepares the ground for some other relation. In a neutral Fel! Bok context, for example, we would interpret Mary's ring as the ring that Mary owns. We märk would, however, be willing to replace this possessive relation by "given by", if context et är inte makes it evident that we are referring to a ring that Mary gave to somebody. defin ierat. Creating metaphors and metonyms involves violating at least one of the defining features forming part of the established meaning, which causes the interpreter to think of the new referent as an unorthodox referent and new meaning as transferred or figurative. The difference between metaphor and metonymy is that the connecter in the case of metaphors is a resemblance-type relation, which may involve more than one shared property, whereas in the case of metonyms there is one and only one connecter. This difference is illustrated by the following examples: My cousin, who is an actor, is a Hamlet. —> "like Hamlet in some respect, ie, possibly, a brooding and indecisive person who feels he has to set things right, or, person who admires his father but has mixed feelings towards his mother, etc" My cousin, who is an actor, is Hamlet. —> "person who represents Hamlet" (equative relation) Metonyms and implications are alike in that the relation between the conventional and novel set of referents is one of concomitance. Metonyms are different from implications in at least the following two respects: (i) whereas implications start off as literal, secondary senses, the metonymic senses are non-literal senses which do not coexist with the senses from which they are derived, (ii) whereas implications may be nouns, verbs or adjectives, metonyms are almost exclusively nouns. The two examples below illustrate these differences: What this organization needs is some muscle. Muscle is a metonym. The interpretation can be paraphrased"that which muscles produce, ie power". This restaurant needs muscular doormen. Muscular in muscular doormen implies "strong". Minor Devices Particularizations, implications, metaphors and metonyms represent the four major devices for creating novel dictionary senses. Understatements, overstatements and irony (here reversals) are other well-known meaning-changing devices, but since they do not often produce senses that are conventionalized, they have been referred to the category of minor devices. However, since in the case of understatement some undesirable feature Fel! Bok märk may be downgraded and in the case of overstatement some desirable feature may be upgraded, these two devices would seem to fit the purpose of euphemisms particularly et är inte well and so the number of cases of under- and overstatement could be expected to be greater than normal. Comparing their number in the present study with their occurrences defin in Warren (to appear), I have indeed found a slight increase. The lists below include all ierat. the examples classed as under- and overstatements in this investigation. Established sense Litotes confusion (NS) (West Indian) "bewilderment" "street fight" cuddle "fondle" "copulate" drug habit "habit (concerning drugs)" "(drug) addiction" erotica "works of an erotic nature" "pornography" mother-loving "loving one's mother" "mother-fucking" indisposed "mildly ill" "ill" insobriety "not in the state of sobriety" "drunkeness" jolly "happy" "intoxicated" unsober "not sober" "intoxicated" Established sense Hyperbole courtesan "kept mistress" "prostitute" epicure "one devoted to the pleasures of the senses" "glutton" gourmand (NS) "person who only eats fine food" "glutton" live upon nothing (NS) "live without using up any resources" "be stringily frugal" maiden lady "young, unmarried lady" "unmarried female, ie spinster" sanitary engineer "person trained in engineering" "garbage man" Admittedly, many of these examples could be classed in other ways. Confusion, jolly, insobriety, unsober, for instance, could be considered to be implications; erotica and cuddle particularizations; engineer a metaphor, etc. What the examples have in common is that in the case of litotes, the conventional referents of the expressions are "less bad" in some respect than the actual contextual referents; in the case of hyperbole the contextual referents are judged to be not quite as good as the conventional referent. Since so many of my examples of litotes and hyperbole are equivocal, it must be conceded that the present investigation does not confirm the existence of under- and overstatements as autonomous meaning-changing processes. Nevertheless, there is little doubt that up- and downgrading of this type exists. We have all experienced people Fel! Bok professing that they are a bit tired or somewhat perturbed when in fact it is evident that märk et är they are exhausted and seriously worried. Similarly we frequently witness exaggerated inte expressions of gratitude, for example. On the other hand, it is natural for over- and defin understatements to be combined with other meaning-changing devices. Metaphors are ierat. fairly frequently hyperbolic, for example. The most interesting aspect of under- and overstatement is that they represent attempted prototypical extensions. That is to say, the coiner expresses a willingness sincere or insincere- to include some contextual referent or referents in a referent class of higher status than may be strictly justified. The interpreter, however, does not accept this. If (s)he had, (s)he would not have thought of the expression as an under- or overstatement. In this under- and overstatements differ from other meaning-creating processes. Reversals could -in the same way as litotes and hyperbole- be taken to suit the purpose of euphemisms, since they enable the coiner to refer to or express "something bad" with its opposite. Blessed "damned" is a good example of this. Reversals are, however, difficult to conventionalize and the present study yielded only two possible examples, namely: Established sense Reversal ?enviable disease "disease to be envied" "syphilis" virtue "goodness" "drinking and womanizing" Analogy is another well-known meaning-changing mechanism. It is different from the mechanisms discussed so far in that the connection between the novel and the established sense does not depend on a connection between two classes of referents but on a connection between two words. If uncle means "gentlemen's toilet", then aunt can be made to mean "ladies toilet" and if girl has acquired the meaning "cocaine", then scag (="ugly young woman") could be used in the sense "heroin"7 . Analogy may form the basis of puns. Ass and bum are synonyms, Johnny and Jack are variants of John. Therefore Johnny bum can be made to mean "jackass". The pun effect derives from the fact that ass and bum are synonyms only in the sense "posterior", not in the "donkey sense". Just as in the case of hyperbole and litotes, I have found that analogy tends to combine with other meaning-changing mechanisms. If pansy is used to refer to effeminate males, this paves the way for cohyponyms such as daisy and buttercup to be used in this sense, too. Consider also the 19th century euphemisms for "trousers", ie inexpressibles, inexplicables, unspeakables, unutterables and unwhisperables, all of which occur in my corpus and which have been classed as (analogical) particularizations. 7 Fel! Bok märk In the above discussion of the results, two particular problems involved in a et är classification of this type have become evident, namely (i) in some cases alternative classifications are possible (either a particularization or an implicature, for example), (ii) inte defin in some cases devices have combined to produce a sense change, notably hyperbolic ierat. metaphors and analogical metaphors or particularizations. In the former case, I chose what to me seemed the most plausible alternative. In the latter case I formed a class of double classifications, but restricted this to include combinations of the major devices only. Combinations of devices are not very surprising in the case of polymorphemic units such as, eg, with a bay window "pregnant", which involves a metaphor (bay window= "protruding belly") and an implication, or drop one's load "defecate", which involves a particularization (load = "feces") and an implication. A more interesting finding of this study as well as of Warren (to appear) is the ease with which metonymic and metaphorical processes apparently combine to form a novel meaning for one and the same lexical unit. Consider some of the examples of this investigation: big brown eyes (NS) "organs of sight which are brown" "breasts" Meto+Meta ("that which has that which is like big brown eyes, ie nipples") bush "thicket" "female genitals" Meto+Meta ("that which has sth which is like a bush") bust "sculpture of head and upper part of body" "woman's bosom" Meta+Meto ("that which is like that which is part of a bust") cyclone "windstorm" "a dose of PCP" Meto+Meta ("that which causes sth which is like a cyclone in degree and kind of effect") grove of egaltine ?mahogany "thicket of the egaltine" "type of dark wood" "female genitals with pubic hair" "a mulatto" Meto+Meta ("that which has that which is like thicket of egaltine") Meto+Meta ("sb with a colour like that of mahogany") paradise 1 "heaven" "female genitals" Meto+Meta ("that which causes one to be in a state as wondeful as paradise") paradise 2 red-cap "heaven" "cap that is red" "cocaine" Meto+Meta As above "penis" Meto+Meta ("that which has sth which is like a red cap") red-jacket "a jacket coloured red" "capsule of Seconal" sauce1 "liquid dressing" "veneral disease" Fel! Bok Meto+Meta märk ("that which has sth et är which is like a red jacket") inte Meto+Meta defin ("that which causes ierat. secretions like sauce") Redefinitions There have been some attempts to explain semantic change in terms of proto-type semantics, notably Aijmer (1985), Geerærts (1985) and Lakoff (1987). The argument has been that since the members of a referent class may vary as to the degree to which they adhere to the criteria which justify membership (some are better, ie more prototypical members than others), the referent set connected with a word can easily be made to encompass new peripheral members and in this way meanings of words change character. Whereas I accept this explanation for one particular type of semantic change, which is fundamentally different from those so far described, I do not think it can be used to account for all types of semantic change. The fact that we can derive the sense "leotardlike blouse" from the sense "human frame" of body does not automatically make this article of clothing even an untypical member of the category of which body ("human frame") is a prototype. The fact that we see a link between the senses "leotardlike blouse" and "human frame" is merely evidence that we believe we know how one sense gave rise to another, not that we include a new type of member in the referent class of body. As I see it, it is because we cannot include blouses in the same category as human bodies that we create a new referent class for the word body. Nor do I think that it is possible to argue that, because we see "human frame" as the primary nuclear sense and "leotardlike blouse" as a secondary, peripheral sense, "human frame" is the prototype sense of body. This would mean not only that we divorce sense from reference, but also that we make prototype and original synonymous, which would rob the notion of prototype of those very features that make it interesting for semanticists. However, prototype semantics can account for redefinitions. Consider for example the definitions of death and dead before and after the introduction of life-supporting apparatus. Before the advent of life-supporting apparatus people agreed that the criterion for death should be cessation of heart beating. After the advent of this apparatus, some people changed their criteria for death and consider a person dead if (s)he has irretrievably lost perception and cognition, irrespective of whether the heart is kept beating or not. These people have then changed not only their criteria for death, but also the meanings of dead/death. Compare also the present-day definition of nurse to that of the 19th century. When it became possible for men to become nurses, the referent class of the word nurse accepted a new type of member which, although distinct from the conventional members, nevertheless was included in the set and caused the meaning of the word to be modified. It is important to make a clear distinction between these sense modifications and the sense developments described above. A sense development involves creating a novel, additional referent class for a word and produces polysemy. A sense modification (alias prototypical extension) involves altered membership criteria and does not produce polysemy but a redefinition. Sense modifications occur when there is some change in character of the conventional referents which is perceptible but insufficient to affect their identity.(consider death, dead) or when a new type of member is accepted in addition to the conventional ones (consider nurse). It is then possible to class as sense modifications (or redefinitions) words used in an inexact manner as in the following examples from my corpus: drug "a narcotic" "all stimulating substances including dope, alcohol and marijuana" killer-weed 2 "PCP" "mixture of PCP and marijuana" old Joe "syphilis" "any veneral disease" Some Final Comments on Results Figures imply preciseness. However, in view of the equivocal and uncertain nature of some of my examples and of the cases of double classsifications that I have ignored, the figures given at the beginning of section III are to be taken as more-or-less values. They serve two main purposes. The more examples one finds of a particular pattern, the better it is supported and the firmer our knowledge of it is. The figures indicate which devices belong to well-established categories. They also serve to give an idea of the relative frequency of the different devices. In Warren (to appear) the order of the frequency among the major devices was: metaphors-particularizations-implicationsmetonyms; in the present investigation the order is: particularizations-metaphorsimplications-metonyms. The great frequency of particularizations in the present study is most probably due to the fact that this device suits the purpose of euphemistic expressions particularly well. As has alredy been mentioned, cases of at least litotes are more numerous among euphemistic than non-euphemistic senses for similar reasons. Finally, the reader should be made aware that in a study of semantics, we are dealing with notions which are not open to observation. A dictionary meaning of a word is an "abstract" of notions the nature of which we infer from how people use words. In a study of this particular kind, we have a postulated original meaning, a postulated novel meaning and a postulated manner in which the novel meaning has been derived from the original one. These postulations can hardly be verified in any other way than by being in accordance with the intuitions of language-users, and so the classification of my examples does not aspire to represent actual etymologies but to be plausible explanations for how the words under study may have acquired their novel meanings. The question of interest here is not primarily how the particular words in the corpus have actually acquired their new meanings but how words in general may acquire new meanings. IV Summing Up Above the following theoretical suggestions have been made: Fel! Bok märk et är inte defin ierat. Fel! Bok märk (i) The meanings we assign a word in our mental lexicon and in a context do not amount et är inte to the same thing. The interpretation of a word may involve changes of its memorized, defin coded meaning. These modifications occur regularly, as a matter of course. They are ierat. interpretive mechanisms which the coiner can trust that his interlocutors will abide by in deciphering the communicative value of the words (s)he uses. That there are such tacitly agreed upon norms of interpretation appears to be confirmed by the existence of the type of euphemism investigated here, since originally these could only have been associated with their intended referents as a result of the application of "rules" of interpretation taken for granted. (ii) At present the most common view among theorists appears to be that only coded meanings are proper meanings and that we will abandon these proper meanings only if they fail to yield a sensible meaning of the whole and that all non-coded meanings are in a sense pragmatic meanings. In contrast to this view, it has been suggested that all word meanings in utterances are automatically contextualized and that to the language-user the proper meaning of a word in a context can be a non-coded meaning produced by one of the regularly employed interpretive mechanisms. (iii) In trying to understand how meanings are created, I have seized upon the following: to have a meaning a word must have a class of referents. Whenever we match a word with a category of entities or phenomena which we have not previously connected with this word, we have created a novel meaning for it (at least a novel contextual meaning) in that we automatically turn the category-specific features of the novel referents into features of meaning. (iv) The creation of a novel contextual meaning is the work of a moment and need not lead to the creation of a novel dictionary sense. Probably it often does not. The creation of a novel dictionary meaning involves the general acceptance of some usage and is necessarily gradual. Some of the meaning-changing mechanisms produce senses which are difficult to conventionalize and so serve mainly as rhetorical devices, whereas others (notably metonymy and metaphor) serve rhetorical and naming purposes equally well. (v) One should clearly distinguish between those changes of meaning which result in polysemy and those which result in a redefinition since they are fundamentally different. (vi) The existence of the meaning-creating interpretive processes suggested here can hardly be considered controversial since they have all been suggested before. However, my definitions of the devices and explanations of how they work differ in some respects from traditional views. I have, for instance, been able to demonstrate that different interpretative mechanisms may combine to produce one new meaning. If we look upon the interpretive process from the decoder's point of view, (s)he seems to abide by the following main maxims in interpreting a word: (i) find a fitting contextual Fel! Bok referent (or some fitting contextual referents) for the word in question (ii) find a reason märk for the speaker to use this word. If these two stipulations are satisfied, the decoder will et är inte feel confident that (s)he has understood the word. defin ierat. If we look upon the process from the encoder's point of view, (s)he seems to reason in the following manner: I need a word for this referent (or these referents). Since there is no word for it/them - or since the word that there is is a word that I wish to avoid - I must form some novel combination of morphemes or use some alternative word or words in such a way that the decoder can match word or word combination and referent. An obvious way to achieve this is to make the word or words describe the referent(s). As we have seen, particularizations describe their referents (although insufficiently) and so do metaphors and metonyms, which should be evident from the paraphrases of combined metonyms and metaphors above. However, whereas particularizations are explicit descriptions, metaphors and metonyms are covert descriptions which have to be inferred. Consider wagtail: "that which has a tail that wags (ie a bird of the Motacillidæ species)" and curl paper "that which is like curl paper in that it is thin and tissuelike (ie toilet paper)". The underlined sections of these paraphrases represent parts of the description left to the interpreter to fill in; the sections in parentheses represent the result of the retrieval of the contextual referent. Another way would be to use a word or word combination which would not only be appropriate to the referent in question but which predictably would cause the interpreter to associate to another intended referent. (If, for example, we tell the driver of a car which has stopped in front of red traffic lights that the lights have turned green, we have -economically and tactfully- conveyed our advice that (s)he should set the car in motion.) A third way would be to use a word the meaning specifications of which fit the intended referent to all intents and purposes, although it has features which make it different from the normal referents of this word. Evidently, we can predict with fair accuracy what goes without saying and by using this intuitive knowledge we can successfully put words to indefinitely many uses. NOTES *I would like to thank Östen Dahl, David Minugh, Magnus Ljung, Anna-Brita Stenström and Eleanor Wikborg, who have read a draft of this article and given valuable advice as to its final shape. 1 In the interest of brevity, the above account is a simplified version of my ideas concerning word meaning. In the full version (Warren, forthcoming), certain distinctions are made which will here be pointed out only in passing. For instance, as the reader will have noticed, I accept the idea that the meaning of a word is made up of components of meaning, which i refer to as features of meaning, sometimes simply features. I have, however, found it necessary to include among the features of meaning of a word not only features which specify what kind of category its referents form, but also (inter alia) features which indicate the speaker's evaluation of the referents in question. Such features do not fix reference. Hence different speakers may agree as to the meaning of, eg, goody and baddy and yet choose different referents for these words. These types of Fel! Bok feature account for the fact that there is a difference in meaning between, say, whore and märk et är prostitute in spite of the fact that these words have the same referents. This in turn inte means that provided there is a difference in evaluative meaning, it is possible for a contextual meaning to be different from a dictionary meaning even if it is connected with defin a category of referents which is the same as the out-of-context category of referents for ierat. this word. It is, however, comparatively rare that novel contextual meanings of this kind develop into novel dictionary meanings. For this reason I feel justified in deemphasizing the importance of these aspects of word meaning in the present study. Further, a distinction is made between not only between dictionary meaning and contextual meaning but between dictionary meaning, contextual meaning and memorized meaning. Dictionary meaning is part of langue (ie that which is generally accepted by the language community in question although not necessarily in the mental lexicons of all the members of the community), contextual meaning is part of parole, memorized meaning is meaning stored in the mental lexicon of individual speakers and therefore not part of parole; it is, however, not part of langue either, being meaning which is not yet accepted or current in too small a circle of people. (This is in accordance with my contention that the creation of a novel contextual meaning is normally the work of a moment, whereas the creation of a novel dictionary meaning is a gradual process which involves memorization.) It should also be made clear that by referent I do not mean only those entities or phenomena that nouns refer to but any non-linguistic entity or phenomenon represented by a word. In other words, in my parlance the verb rain has a referent, viz the event of precipitation in the form of water drops. Whether this event is ongoing or habitual, in the past, future or present will be indicated by tense-& aspect-denoting morphemes. Finally, it should be pointed out that the distinction between contextual and dictionary meaning is not equivalent the distinction between reference and meaning . Whereas reference involves the denoted entity or phenomenon per se, contextual meaning involves that which is communicated about the referent. 2. These hypotheses have also been tested on 500 novel dictionary senses of standard English words and 500 English slang senses derived from standard senses, such as, for example, acid meaning "LSD". The results of that particular project are described in Warren (to appear) 3 Johnson and Murray (1985:152) give the following amusing instance of ellipsis in combination with implication: Note from Willie's mother to his teacher:"Willie can't come to school because he hasn't been. I've given him something to make him go, and when he's been, he'll come." 4 There may of course be homonymic clashes. Consider, eg, jay above. 5 Abbreviations could also be considered a word formation device. 6 That is to say: Cedersciöld (1900: 59-85); Greenough & Kittredge (1902: 300-309); Jespersen (1905: 241-246), Wellander (1923: 19-21); Stern (1931:330-332); Williams (1975: 198-203), repeated in Neaman & Silver (1983:9-12). These are classifications pertaining to ways of forming euphemisms. It is also quite common that linguists classify according to their rationale. Ullmann (1963: 193-195), eg, divides euphemisms into three groups: (i) those inspired by fear, as manifested in the avoidance of using God's name (ii) those dictated by a sense of delicacy (euphemisms of death, defects and crimes) and (iii) those resulting from a sense of decency (euphemisms of sex and private parts and swear words inter alia). 7 Uncle, in turn, may possibly have assumed the meaning "gentlemen's toilet" on the analogy of gentlemen (in the sense "gentlemen's toilet"). Neaman and Silver (1983:197) say about girl "cocaine" that "because cocaine is purported to be an aphrodisiac for women, it is frequently identified by such female names as Her, girl." That would make girl "cocaine" into a metonym, ie "that which is for a girl" - a purpose-relation. Purposerelations are common covert connecters in noun-noun compounds and adjective-noun combinations, but, it seems, less common as metonymic connecters. References Aijmer, Karin. (1985). "The Semantic Development of will ", pp 11-21 in Fisiak, Jacek (ed). Historical Semantics, Historical Word-formation, Berlin-New YorkAmsterdam:Mouton. Cederschiöld, Gustaf. (1900). Om kvinnospråk och andra ämnen. (The article on euphemisms was first published in1896.) Clark, Eve and Herbert Clark. (1979). "When nouns surface as verbs", Language 55:4, pp 467-481. Cooper, David. (1986). Metaphor, Oxford: Blackwell. Geerærts, Dirk. (1985). "Prototype Theory and diachroninc Semantics", Indogermanische Forschungen 88: 1-32. Greenough, James Bradsheet and George Kittredge (1902). Words and Their Ways in English. London: Macmillan & Co. Grice, Herbert. (1974). "Logic and Conversation", pp 41-58 in Cole, P and J Morgan (eds). Syntax and Semantics 3: Speech Acts, New York & London: Academic Press. Jespersen, Otto. (1902). Growth and Structure of the English Language. London: Macmillan & Co. Fel! Bok märk et är inte defin ierat. Fel! Bok Johnson, Diane and John F Murray "Do Doctors Mean What They Say?" in Enright, D. märk et är J. (ed). Fair of Speech, Oxford University Press. inte Lakoff, George. (1987). Women, Fire and Dangerous Things. Chicago & London: The defin ierat. University of Chicago Press. Levinson, Stephen. (1983). Pragmatics, CUP. Neaman, Judith and Carole Silver. (1983). A Dictionary of Euphemisms. London: Hamish Hamilton. Partridge, Eric. (1984). A Dictionary of Slang and Unconventional English, London, Melbourne and Harley: Routledge and Kegan Paul. Paul, Hermann. (1920, first published in 1880). Prinzipien der Sprachgeschichte, Hall a S: Max Niemeyer. Quirk, R., Greenbaum, S., Leech, G., Svartvik, J. (1987). A Comprehensive Grammar of the English. Language . London:Longman. Ullmann, Stephen. (1963). "Semantic Universals", pp 172-207 in Greenberg, Joseph (ed.) Universals of Language, The MIT Press, Cambridge, Massachusetts. Spears, Richard. (1980). Slang and Euphemisms. New York: Jonathan David Publishers. Stern, Gustaf. (1931). Meaning and Chage of Meaning, Göteborgs Högskolas Årsskrift XXXVIII. 1932:1. Warren, Beatrice. (to appear) Sense Developments. Wellander, Erik. (1923). Studien zum Bedeutungswandel im Deutschen II:9-21. Uppsala Universitets Årsskrift. Widdowson, Henry G. (1979). Explorations in Applied Linguistics, Oxford University Press. Williams, Joseph. (1975). Origins of the English language., New York: The Free Press. View publication stats