Fachbereich Sprachwissenschaft Universität Konstanz Arbeitspapier (to appear) Peter Koch Metonymy: unity in diversity In: Regine Eckardt & Klaus von Heusinger (eds.) Meaning Change – Meaning Variation. Workshop held at Konstanz, Vol. II. Peter Koch Metonymy: unity in diversity* Je donne entre les especes de Tropes, la premiere place à la Metonymie, parce que c’est le Trope le plus étendu, & qui comprend soûs luy plusierus autres especes (Bernard Lamy, De l’Art de Parler, Paris 1676).** When dealing with linguistic effects that are usually labelled as ‘metonymy’, we may sometimes doubt that this term really denotes a unitary range of phenomena. Rather, heterogeneity seems to be so typical of ‘metonymy’ that we have to consider repeatedly whether we should exclude certain phenomena from the realm of ‘metonymy’. However, I am convinced that most, though not all, of the multifarious phenomena subsumed under the label ‘metonymy’ form a unity and that many traditional and modern accounts of this unity converge in one point. In the following, I first will try to sketch this unitary understanding of ‘metonymy’ (section 0) and then point out that the indisputable heterogeneity of the corresponding range of phenomena is not chaotic, but can be systematized along different dimensions (sections 1-3). Measuring this pluridimensionality (cf. also Radden/Kövecses 1999, 21-44) will enable us to determine more accurately the true range of metonymy (section 4). 0. Metonymy as a figure/ground effect within frames 0.1. Contiguity, frame, and figure/ground effect The earliest definition of ‘metonymy’ (Rhetorica ad Herennium 4, 32, 43) already makes use of the notion of ‘near and close’. The relation of closeness clearly corresponds to what would later be termed ‘contiguity’ by associationist philosophers and psychologists (a relation identified long before by Aristotle in his De memoria et reminiscentia 451b: 18-22). The rhetorical and the associationist currents of thinking were integrated for the first time by Roudet, who defined ‘metonymy’ as a type of semantic change „résultant d’une association par contiguïté entre les idées“ (1921, 690). The importance of semantic contiguity as underlying metonymy was confirmed by Jakobson (1956) on the synchronic level and by Ullmann (1962, 218-220) on the diachronic level.1 Whereas contiguity is presented from a more or less structuralist perspective as a linguistic relation between signifiés of words (Jakobson, Ullmann; also Dubois et al. 1970, 106 ff.), a cognitive approach should conceive of it rather as a conceptual/perceptual relation.2 In my view, contiguity is the relation existing between elements of a * I express my gratitude to Keith Myrick for the stylistic revision of this paper. ** Special thanks to Brigitte Schlieben-Lange (Tübingen) for this quotation. 1 For a more detailed, critical outline, see Koch 1999a, 144-145; also, cf. Blank 1997, 10-20, 40-43, 230-235; 1999a, 170-172. 2 Even if the notion of ‘contiguity’ sometimes appears in the cognitive literature (cf., e.g., Taylor 1995, 122; Croft 1993, 347; Ungerer/Schmid 1996, 115 f.; Radden/Kövecses 1999, 19), it is generally not systematically exploited. – As for the necessary distinction between linguistic signifiés of words and concepts, cf. Koch 1996a, 2 prototypical conceptual/perceptual frame or between the frame as a whole and each of its elements (cf. Koch 1999a, 145-149). I use the term ‘frame’ here in a very general sense, comprising also ‘scene’, ‘scenario’, ‘script’ etc. 3 Of course, elements of a frame can, in turn, constitute (sub-)frames. In Cognitive Semantics, the notion of frame has turned out to be particularly fertile with respect to processes of ‘perspectivization’ and ‘windowing of attention’ (cf. Fillmore 1977; Dirven et al. 1982; Talmy 1996; Ungerer/Schmid 1996, 205-249). One can speak of ‘processes of perspectivization within frames’ from two totally different semiotic points of view. On the one hand, we can raise the onomasiological question of how different perspectives of a frame are expressed linguistically. For the COMMERICAL EVENT frame, e.g., we have English verbalizations such as the following: (1) a.Engl. Ann bought an old book from Daniel for 90 Cents. b. Engl. Daniel sold an old book to Ann for 90 Cents. c. Engl. Ann paid 90 Cents to Daniel for an old book. d. Engl. The old book cost Ann 90 Cents. On the other hand, we can raise the semasiological problem if different senses of a given linguistic expression correspond to different perspectives within the same frame. It is in this sense that perspectivization has been related to metonymy (cf. Taylor 1995, 90, 107 f., 125 f.).4 Since frames are to be considered conceptual/perceptual gestalts,5 we can easily explicate this semasiological understanding of ‘perspectivization’ in terms of gestalt theory and, more specifically, in terms of ‘figure/ground’ effects. In the visual realm, this can be illustrated by the traditional example of Figure 1, where we perceive either four black squares on a white ground or, alternatively, a white cross on a black ground: v Figure 1 223 f., 226-231; 1996b (unfortunately, this distinction is widely unknown or is questioned by the mainstream of Cognitive Semantics: cf., e.g., Haiman 1980; Taylor 1999, 18-25; cf. also the critical outline in Blank 2000, section 3). 3 Cf. Minsky 1975; Fillmore 1975; 1985; Schank/Abelson 1977; Tannen 1979; Barsalou 1992. – In this context, I avoid the term ‘domain’, because it is often used to denote contiguity-based frames as well as taxonomic hierarchies in an undifferentiated manner (cf. Koch 1999a, 152 f.; cf. also below, Figure 3a/b). I also avoid the ambiguous term ‘idealized cognitive model’ (ICM), proposed by Lakoff (1987) and taken up by Panther/Radden (1999, 9) and Radden/Kövecses (1999); cf. the critical remarks in Koch 1996a, 234 n. 28, concerning its use for metaphor. 4 In a more informal way, the relevance of frames to metonymy has been evoked by Fritz (1998, 45 f.). 5 For the basic notions of gestalt theory, cf., e.g., Wertheimer 1922/23; Köhler 1947; Metzger 1986. 3 Similarly, with respect to a prototypical frame PURSUING AND TRYING TO CATCH AN ANIMAL THAT RUNS AWAY, the French verb chasser can „highlight“ either the aspect of TRYING TO CATCH (2a) or, alternatively, the aspect of MAKING RUN AWAY (2b) (cf. Meillet 1905/6, 259; Jongen 1985, 131 f.; Koch 1991, 296): (2) a.Fr. Nous irons chasser du gibier. ‘We will go to hunt game.’ b. Fr. Nous avons chassé les chiens de notre cuisine. ‘We chased the dogs from our kitchen.’ Thus, the metonymic relation between these two senses of Fr. chasser can be described in terms of a figure/ground effect along the contiguity relations within 6 a given conceptual frame (cf. Koch 1993, 269 ff.; 1995a, 40 f.; 1996a, 236; 1999a, 151-153, 155-157; Blank 1997, 235-243). This figure/ground effect is summarized in the following table: 7 Fr. chasser ... TRY TO CATCH ... MAKE RUN AWAY figure ground sense A (2a) Nous irons chasser du gibier. sense B (2b) Nous avons chassé les > ground chiens de notre cuisine. > figure Table 1 0.2. A basic definition of metonymy Though convinced of the unity of metonymy, one still has to cope with the apparent heterogeneity of so-called ‘metonymies’ or ‘metonymic effects’. Nevertheless, it seems useful to systematize different dimensions of this heterogeneity and then to determine for each dimension separately the applicability of the unitary understanding of metonymy sketched in 0.1. In order to detect several of these dimensions (some of which will be well known, whereas others may be new to the reader), we can begin with the following provisional definition: (3) Metonymy is (i) a linguistic effect upon the content of a given form, (ii) based on a figure/ground effect with respect to cognitive frames and 6 Cognitive semanticists have repeatedly insisted on the fact that metonymies, as opposed to metaphors (s. below n. 13), function within one experiential frame, domain (matrix), etc.: cf. Lakoff/Johnson 1980, 39 f.; Lakoff 1987, 288; Croft 1993, 345-348. 7 Note that the notions ‘figure’ and ‘ground’ have been exploited in Cognitive Linguistics in several, rather different ways. The present application to frames, contiguities, and metonymies should not be confused with other applications, like those delineated, for instance, in Ungerer/Schmid 1996, 156-200. 4 contiguity relations, (iii) and generated by pragmatic processes. This definition truly reflects the interest for metonymy shown by different disciplines as linguistic semantics ((i); s. below section 1), cognitive sciences ((ii); s. below section 2), rhetoric, literary studies, and pragmatics ((iii); s. below section 3). 5 1. Metonymy as a linguistic effect If metonymy is considered to be „a linguistic effect upon the content of a given form” ((3)(i) above), this implies a variable element (the content) and an invariable element (the form). Of course, the invariability of the form can be only of a relative nature. Thus, in example (4), illustrating a metonymic change of content from (a) to (b), the form actually undergoes a dramatic sound change, but due to diachronic continuity, it has to be regarded as an invariant. Consequently, the relevant linguistic effect concerns the content level exclusively. (4) a. b. CLat. tremere ‘to tremble’ OFr. criembre, ModFr. craindre [kRg:dR∂] ‘to fear’ Linguistic effects can be further differentiated according to essentials of human language. In the present context, the following aspects have to be considered: (5) Human language (i) is characterized by the dialectics between conventional rules and individual speech events: innovations always emerge in individual speech events, whereas real language change is a matter of conventionalization;8 (ii) has a paradigmatic and a syntagmatic dimension; (iii) has several subsystems (phonology, grammar, lexicon, etc.). From these points, we can determine three essential problems concerning metonymy as a linguistic effect (1.1.-1.3.). Surely, one could question in any case the statement that metonymy i s a linguistic effect. This will be discussed further in 4.1. 1.1. Innovation vs. conventionality If, according to (5)(i), human language is characterized by the dialectics between individual speech events and conventional rules, we have to distinguish between two types of effects that are both frequently subsumed under ‘metonymy’. To begin with, there are metonymic ad hoc effects that occur spontaneously in discourse. I had the opportunity to personally observe an effect of this kind: As my professional work threatened to cause me to neglect our family life, my wife and I agreed to set aside one day per week for the family (Saturday or Sunday, as a rule). We called it jour fixe (6a). On that day, together with our little boy, we would take a trip, go for a walk, or go sightseeing. Time went by, and our son was getting to the age when boys begin to be bored by promenades or sightseeing with their parents. One day when we announced once again our jour fixe, he protested: „Och nein, bitte keinen jour fixe !“ [‘Oh no, no jour fixe, please !’]. He clearly had effected a totally individual metonymic innovation (6b) with respect to our (personal) adults’ sense of jour fixe: 8 For the distinction between (ad hoc) ‘innovation’ and ‘change’ (by adoption), cf. Coseriu 1958, 44-46. 6 (6) a. b. jour fixe ‘day reserved for the family’ jour fixe ‘day of boring promenades and sightseeing’ Ad hoc metonymies like this are opposed to metonymic effects that occur between different senses of certain polysemous lexemes: (7) a. b. Engl. child ‘offspring, descendant’ Engl. child ‘very young person’ While in cases like (6) the metonymic sense (b) is an innovation generated ad hoc in discourse, in cases like (7) the metonymic sense (b) has already been lexicalized, i.e. conventionalized (or habitualized), ready for being activated in discourse at any time. Ad hoc metonymy, as a „trope“, has always been a topic of rhetoric. Metonymic change (and polysemy), on the other hand, is investigated by historical semantics, even in those cases where the original non-metonymic sense has eventually disappeared and remains a purely „archeological“ background (cf. (4), where the sense ‘tremble’ is no more present in French). Despite all these differences, the underlying cognitive relation and the corresponding figure/ground effect is in both cases fundamentally the same: in (6) it is a contiguity FAMILY DAY – BORING PROMENADE DAY based on the prototypical experiential frame WEEKEND DAY of a young boy; in (7) it is a contiguity based on the fact that in certain experiential frames we prototypically conceive of a VERY YOUNG PERSON in terms of DESCENDANCE (and also, to some degree, vice versa). From a diachronic point of view, the two phenomena are intimately related. Ad hoc metonymies – if they are „successful“, which is not necessarily the case (cf. (6)) – ultimately c a n induce metonymic changes that lead to metonymic polysemies like (7). Metonymic polysemies, in turn, are generally the result of conventionalized ad hoc metonymies in discourse.9 (This distinction does not apply, however, to metonymies of type C in 2.3.) Terminologically speaking, we should insist on the fundamental cognitive unity of all metonymic effects, without neglecting the crucial difference between innovative and conventionalized effects. Accordingly, we would have to distinguish ‘(ad hoc) metonymy’, involving an ad hoc figure/ground effect, from ‘metonymic change’ and ‘metonymic polysemy’, involving a conventionalized figure/ground effect.10 In the following, for the sake of terminological convenience, I will nevertheless subsume, ‘(ad hoc) metonymy’, ‘metonymic change’, and ‘metonymic polysemy’ under the general label of ‘metonymy’ tout court. 9 For the processes of lexicalization of ad hoc lexical innovations, including, of course, metonymy, cf. Koch 1994, 203-209; 1999a, 139 f.; Blank 1997, 116-130. As for the intimate link between semantic change and polysemy cf. Bréal 1921, 143 f.; Koch 1991, 293; 1994, 203-209; Wilkins 1996, 267-270; Blank 1997, 114-130, and this volume. 10 Speaking of „conventionalized figure/ground effects“ makes sense, because metonymic polysemy is a kind of motivation. Once an existing metonymic figure/ground effect fades away, we get pure homonymy: e.g. Germ. Flegel ‘flail’, hence ‘peasant’ (first metonymic step), hence ‘boor, lout’ (second metonymic step; cf. below ex. (33)); due to the disapperance of the intermediate sense, this is today perceived to be homonymy. 7 1.2. Syntagmatic explicitness According to (5)(ii), human language has a paradigmatic and a syntagmatic dimension. The former concerns the essence of metonymy itself, since the metonymic use of a linguistic element totally changes its position within the paradigm of which it is a part (normally cutting across lexical fields, linguistic subsystems in the sense of 1.3., etc.). In contrast to this, the syntagmatic dimension allows interesting modulations of the metonymic effect – an insight that, to my knowledge, most scholars have not yet accounted for. At this point, we discover an interesting analogy with metaphor, the functioning of which also varies largely depending on the syntagmatic environment. We can distinguish at least the following „syntagmatic“ types of metaphors:11 (8) Explicit metaphor in praesentia: Fr. Partir, c’est mourir un peu. ‘To depart is to die a little.’ (9) Semi-explicit metaphor in absentia: Engl. The rose melted. ‘ The rose lost it petals.’ (Levin 1977, 24) (10) Implicit metaphor in absentia: Engl. I could not get rid of this gorilla. (metaphorical sense ‘ruffian’ only in case gorilla refers to a man, and not to an ape) In the same way, we can distinguish at least three „syntagmatic“ types of metonymy: (11) Explicit metonymy in praesentia:12 Engl. Knowledge is power. (12) Semi-explicit metonymy in absentia: Engl. The buses are on strike. ‘ The bus drivers are on strike.’ 11 Cf. especially Pirazzini 1997, 34-40; cf. also Brooke-Rose 1958, 26-67, 149-152; Searle 1979, 118; Prandi 1992, 127-134; Koch 1994, 210 f. A more fine-grained classification of metaphors is possible, for example with respect to the Aristotelian analogical equation A : B = C : D, whose elements can be realized syntagmatically in varying number (in (8), we have only A and C). 12 This very radical type of metonymy seems to be rather rare, and, due to its predicative form, it is easily confused with the corresponding type of metaphor. However, (11) does not mean that KNOWLEDGE is CONCEIVED OF AS POWER, but that KNOWLEDGE PROVIDES POWER, which is a contiguity relation within a frame, say, SOCIETY (knowledge enables people to occupy central positions in a society, to influence political decisions, etc.). Other metonymies in praesentia are utterances like Palestine is Arafat (frame PALESTINE: Arafat is the most salient person with respect to this frame) or Time is money (Lakoff/Johnson (1980, 7) present TIME IS MONEY as a conceptual metaphor (citing examples like You’re wasting my time, This gadget will save you hours). I think that this interpretation is quite possible (TIME IS CONCEIVED OF AS MONEY), but clearly different from the frame-based figure/ground effect underlying the sense of the everyday dictum Time is money that explicitly presents TIME and MONEY as belonging to the same frame BUSINESS (working hours have to be paid; wasting time costs money; money can buy manpower in order to save time; etc.): TIME „COSTS “ MONEY. 8 (13) Implicit metonymy in absentia: Engl. Where can I wash my hands ? ‘Where can I urinate/defecate ?’ (for example in a restaurant) (cf. Radden/Kövecses 1999, 53) In (8) and (11), the two relata of the underlying cognitive relation (similarity for metaphor, contiguity for metonymy) are both realized syntagmatically ((8) partir — mourir, (11) knowledge — power). In (9) and (12), only the metaphorical/metonymic element (melt/buses) is present, but it stands in a syntagmatic relation to another element (rose/on strike) that reveals its metaphorical/metonymic character. In (10) and (13), the possible metaphorical/metonymic character of gorilla/to wash my hands can be deduced only from the situation or from a larger context. In another context it could be understood „literally“. Beyond any doubt, the three examples of metonymy cited display equally well figure/ground effects within experiential frames: (11) frame SOCIETY (cf. also n. 12); (12) frame BUS; (13) frame TOILET, and this is exactly what distinguishes them from the metaphors of the corresponding types (8), (9), and (10), that all involve mapping across frames and domains. 13 The only difference between the types (11), (12), and (13) concerns the way in which the figure/ground effect is expressed syntagmatically: (11) by a radical pseudo-equation (is) of the two relata of the contiguity relation (KNOWLEDGE – POWER ); (12) by explicitly adding elements of another frame (STRIKE) the highlighted concept ( BUS DRIVER ) also belongs to; (13) by committing to a contextually and only implicitly given frame ( RESTAURANT ) the task to highlight the intended concept (URINATE/DEFECATE). 1.3. Linguistic subsystems concerned According to (5)(iii), human language has several subsystems that can be involved in metonymies. Since metonymy is an effect acting on whole linguistic signs that encompass an expression face (called here ‘form’) and a content face, the phonological subsystem is never of immediate concern to metonymies.14 The traditional examples of metonymy are lexical, but this does not mean that metonymic effects are alien to grammar or even to pragmatics or the discourse level. Rather, we have to distinguish: 1.3.1. 1.3.2. 1.3.3. 1.3.4. 1.3.5. purely lexical metonymies, lexical metonymies with grammatical implications, metonymies as transitions between lexicon and grammar, metonymies within grammar, pragmatic metonymies, 13 As for metaphor, cf., e.g., Bühler 1965, 342-350; Lakoff/Johnson 1980; Liebert 1992, 28-82; Croft 1993, 345- 348; Koch 1994, esp. 213. 14 Effects of phonetic erosion and shrinking, as they occur in processes of grammaticalization (s. 1.3.3., below), are only a by-product of semantic processes, including metonymy (cf. Detges/Waltereit 1999). 9 1.3.6. metonymies as transitions between pragmatics and the lexicon, 1.3.7. metonymies at the level of discourse semantics. 1.3.1. The purely lexical metonymies are exemplified by (2), (4), (6), (7), (11), and (13) and need no further comment. 1.3.2. The verb is a part of speech geared to the linguistic representation of frames that correspond to our conception/perception of states of affairs; on the formal level, this is reflected in valency, a phenomenon that intrinsically characterizes the verb and interweaves it with the grammatical structure of the sentence (cf. Seyfert 1976, 122, 356; Fillmore 1977; Koch 1981, 80 f.; Heringer 1984; Waltereit 1998, 53 f.). Recent studies in historical semantics have shown that the realm of the verb is rich in metonymic processes that also affect its morphosyntactic environment, i.e. its valency. Recessive/causative alternations, also called inchoative/causative alternations, are a current example (cf., e.g., Tesnière 1959, 271; Koch 1991, 294 f.; Haspelmath 1993, 92, 101, 104; Waltereit 1998, 84-91): (14) a. Germ. Die Kritiker verstummen. ‘The critics grow silent.’ b. Germ. Die Leistung der Mannschaft verstummt ihre Kritiker. ‘The team’s performance silences its critics.’ (ad hoc metonymy by a sports reporter reprimanded in a squib in Der Tagesspiegel, June 9, 1995, 23) (15) a. Engl. Suddenly, the door closed. b. Engl. Suddenly, John closed the door. These are clear cases of metonymy, because they display a figure/ground effect between part of an EVENT frame (a) and the whole EVENT frame, including an INSTIGATOR (b).15 Due to the verb-inherent phenomenon of valency, this kind of metonymy is reflected in a reorganization of grammatical valency. The only participant, expressed as a subject in (a), moves to the direct object position in (b), where the subject expresses the INSTIGATOR. From the diachronic point of view, the metonymy goes from PART to WHOLE in (14) and from WHOLE to PART in (15): whereas the causative sense of (14b) does not yet belong to Standard German usage, the causative sense of (15b) is primary (close < OFr. clore ‘to make close’). Other kinds of lexical metonymies with grammatical implications are auto-conversions (s. below 2.2.) and locative alternations or so-called swarm alternations (cf., e.g., Fillmore 1968, 391; Anderson 1971; Koch/Rosengren 1996; Waltereit 1998, 67-74). 1.3.3. On the semantic level, grammaticalization processes have been explained in terms of ‘semantic bleaching’ or ‘generalization’ (e.g., Lehmann 1995, 126-129; Hopper/Traugott 1993, 93; Bybee et al. 1994, 289-293) and of metaphor (e.g., Heine et al. 1991, 45 ff.; Stolz 1994; Keller 1995, 230-239). Much more rarely, metonymy has 15 As for part/whole metonymies s. below 2.2. 10 been taken into account (Heine et al. 1991, 61-64; Traugott/König 1991, 210-213 16). As Detges could demonstrate on the basis of different grammaticalization processes (leading to futures, perfects, adverbs, and negations), metonymy is one, or perhaps t h e central, process in grammaticalization (cf. Detges 1998; 1999; in press; Detges/Waltereit 1999). In frames of human action, for instance, there is a strong contiguity between INTENTION and FUTURE. By a figure/ground effect between these two elements of the respective frames, a modal lexical item, as MEngl. wilen in (16a), can shift to an expression form for the future and become totally grammaticalized, as in (16b) (where Peter surely did not w a n t to deny Jesus!). (16) a. MEngl. [...] Deeth [...] ne wol nat han my lyf. ‘Death does not want to have my life.’ b. MEngl. you wolt forsake me zrien. (Jesus to Peter) ‘Thou shalt deny me thrice/You shall deny me thrice.’ (cit. Mossé 1969, 136 and 134) Metonymic processes of this and other kinds, then, accompany diachronic transitions from lexicon to grammar. 1.3.4. Fascinated by grammaticalization, recent linguistic research has in general neglected semantic change within grammar (or has even incorporated it into a particular conception of ‘grammaticalization’). Upon closer inspection, it seems that intragrammatical semantic change is subject largely to the same regularities as lexical semantic change. Just as there are lexical metonymies (s. above 1.3.1.), there are also intragrammatical metonymies. Take, for instance, the relationship between the temporal function of the French imparfait (17a) and its modal, counterfactual function (17b): (17) a. Fr. Jean avait assez d’argent. ‘Michael had enough money.’ b. Fr. Si Michel avait assez d’argent, il achèterait une voiture. ‘If Michael had enough money, he would buy a car.’ From a form expressing PAST like (17a) and implying present-time non-factuality, a figure/ground effect may lead to a form expressing present-time COUNTERFACTUALITY within the same frame. Since the French imparfait is already totally grammaticalized as a tense, its modal function in (17b) has to be viewed as the result of an intragrammatical metonymy (with parallels in several other languages: cf. Taylor 1995, 149-152). Other examples of the (rather common) phenomenon of grammatical metonymy are the emergence of epistemic futures (cf. Bybee et al. 1991, 22-32, without a metonymic interpretation, however) and reanalyses in already grammaticalized reflexive constructions (cf. Detges/Waltereit 1999). 1.3.5. For about three decades, indirect speech acts have been a controversial issue (cf., e.g., Searle 1975; Levinson 1983; Sperber/Wilson 1986, 243 ff.). Be this as it may, it is striking that (pragmatic) frames seem to play an important role in this realm. According 16 ‘Pragmatic strengthening’ and ‘conventionalization of conversational implicatures’, as described especially by Traugott and König, are, in fact, metonymies: cf. Taylor 1995, 152; Koch 1999, 155 f. (s. also below 3.1.). 11 to Searle, one element of a directive speech act ‘frame’ or ‘scenario’ (as we could express it in cognitive terms) is the preparatory condition that the hearer H is able to perform the action A requested by the speaker S. Consequently, by asking for the validity of this preparatory condition, S can perform, through a figure/ground effect, the intended directive speech act: (18) Engl. Can you help me ? ‘Help me !’ So, these and similar cases, concerning various speech act types and their respective conditions, can be considered as ‘speech act metonymies’ (cf. Taylor 1995, 157; Thornburg/Panther 1997; Panther/Thornburg 1999; Gibbs 1999, 72 f.; cf. also the considerations in Schifko 1979, 259 ff.). 1.3.6. A totally different kind of ‘speech act metonymies’ underlies the emergence of performative verbs. This can be exemplified by the verb Lat. mandare: (19) a. Lat. Tibi mando ut hunc nuntium statim Marco amico deferas. ‘I entrust you to deliver this news at once to my friend Marcus.’ b. Lat. mandare ‘to order’ [ORDER ] In the propositional content of a speech act of order, Latin speakers became accustomend to frequently choosing the non-performative verb mandare ‘entrust’. Due to a figure/ground effect within the frame of the speech act of order, mandare eventually came to express the concept ORDER itself (cf. Anscombre et al. 1987; Koch 1991, 295; 1993, 268-271; Blank 1997, 256-258). In this way, metonymic processes accompany so-called ‘delocutive’ diachronic transitions from pragmatics to the lexicon. 1.3.7. As for the level of discourse semantics, s. below 2.3., type C. 2. Metonymy as based on a cognitive relation We have considered metonymy to be „based on a figure/ground effect with respect to cognitive frames and contiguity relations” ((3)(ii)). This implies a further series of differentiations: (20) The cognitive relation of contiguity (i) can be grasped on different levels of abstraction; (ii) has to be defined with respect to entities involved in the process of semiosis (especially concept and referent); (iii) has to be allocated with respect to frames and thereby distinguished from other cognitive relations. From these points, we can determine three essential problems concerning metonymy as based on a cognitive relation (2.1.-2.3.). 12 2.1. Level of abstraction Literally speaking, ‘contiguity’ means ‘spatial closeness’. In this strict sense, ‘contiguity’ would fit only cases of metonymy like (21): (21) a. OFr. riv(i)ere ‘bank, shore’ b. ModFr. rivière ‘river’ ( comp. Engl. river) In fact, the terms ‘metonymy’ and/or ‘contiguity’ have been used in rhetorical, psychological, and linguistic tradition (cf. 0.1.) in a much broader sense. Thus, it seems sound to strip the term ‘contiguity’ of its literal (spatial) limitation and to apply it to all kinds of experiential links within and with respect to frames (cf. Koch 1999a, 145 f.). Nevertheless, traditional and even modern treatments of metonymy continue to display (partly varying) inventories of „real“, „logical“, etc. relations involved in metonymies (cf., e.g., Fontanier 1977, 79-86; Lausberg 1973, § 568; Bredin 1984, 48; Bonhomme 1987, 60-70; Blank 1997, 230-235; 1999a, 176-178; Waltereit 1998, 19-22; Radden/Kövecses 1999, 29-44; Nerlich/Clarke/Todd 1999, 363 f.; Seto 1999, 98-113): – OBJECT/CONTAINER – CONTENT; PLACE – PLACE (cf. (21)); TIME – OBJECT/STATE OF (cf. (6)); PRECEDING – ENSUING (cf. (13) and (16)); CAUSE – EFFECT (cf. (4) and (11)); CAUSE – RESULT (cf. (2)); CONDITION – CONDITIONED (cf. (18)); AGENT – ACT; ACT – OBJECT; AGENT – OBJECT; AGENT – INSTRUMENT/CONTROLLER – CONTROLLED (cf. (12)); PRODUCER – PRODUCT; POSSESSOR – POSSESSED; OBJECT – PROPERTY; SYMBOL – SYMBOLIZED (19); PART – WHOLE (cf. (14) and (15)); and many others. PLACE AFFAIRS There are three problems with these kinds of typologies. First, they are never exhaustive (to which types of relation are we to assign, for instance, the clearly metonymic cases (7) and (17) ?). Second, the types of underlying relations are not really discrete (does (12), for instance, correspond to AGENT – INSTRUMENT/CONTROLLER – CONTROLLED or rather to AGENT – OBJECT or even to POSSESSOR – POSSESSED ?). Third, the relations can be defined at different levels of abstraction. Radden/Kövecses (1999, 38 f.), e.g., subsume under the type EFFECT – CAUSE relations as STATE/EVENT – THING/PERSON/STATE CAUSING IT , EMOTION – CAUSE OF EMOTION, MENTAL/PHYSICAL STATE – OBJECT/PERSON CAUSING IT , PHYSICAL/BEHAVIORAL EFFECT – EMOTION CAUSING IT, etc. On the other hand, Blank (1997, 249-253; 1999a, 178-184), focussing on the fundamental conceptual distinction between static ‘frames’ (in a narrow sense) and dynamic ‘scenarios’, puts forward the two very general relations of CO -PRESENCE and SUCCESSION that are suitable for integrating all possible types of contiguity. All in all, the relations mentioned above (and other similar ones) are convenient formulas that help us to spell out, at a higher or lower level of abstraction according to (20)(i), the very general relation of ‘contiguity’ that encompasses them all. They correspond to different possible types of relations within frames and, therefore, are all susceptible to the figure/ground effect constituting metonymy (we shall have to return, however, to the type PART – WHOLE in 2.2.). 13 2.2. Contiguity and frames In 0.1., I defined metonymy as „the relation holding between elements of a ... frame or between the frame as a whole and each of its elements“. This definition already suggests that the connections between contiguities and frames is subject to variation. Therefore, according to (20)(ii), the relation of contiguity has to be allocated somewhat more precisely with respect to frames. Let us start with the simplest constellation: contiguity as the relation between two referent points X and Y, that are elements of a given frame F (Figure 2b). This has been the basis of the figure/ground effect for many of the examples given so far: (2), (4), (11), (12), (13), (16), (17), (18), (19), and (21). F=X F X Figure 2a Figure 2b F Y X F=X Y Figure 2c Y Figure 2d We can modify the constellation of Figure 2b in two directions. Either we reduce the autonomy of X and Y with respect to F so that the reference point becomes the frame itself (F = X) and the figure/ground effect concerns different internal aspects of the frame (Figure 2a; we will come back to this issue in 2.3.), or we can modify Figure 2b by imagining that X and Y are not totally distinct but are overlapping or interwoven in some way (Figure 2c and 2d). As long as X and Y are more or less on a par, they correspond to divergent perspectives on what is partly the same cognitive „material“ within the frame F (Figure 2c). A particularly striking example are the so-called ‘auto-conversions’ changing a given verb to its own converse (cf. Koch 1981, 358; 1991, 296-299; 1995b, 130 f.; Blank 1997, 269-278; Waltereit 1998, 75-83; Fritz 1998, 124 f.): (22) a. Fr. La société immobilière a loué cet appartement à un étudiant. ‘The realty company let this apartment to a student.’ b. Fr. L’étudiant a loué cet appartement à une société immobilière. ‘The student rented this apartment from a realty company.’ Exactly as in the examples (1a/b), we are faced here with the phenomenon of perspectivization within a given frame, but whereas in (1a/b) this is a synchronic, merely onomasiological problem, (22a/b) also point out a diachronic, semasiological problem. Indeed, Fr. louer, whose primary meaning was the one exemplified in (22a) (< Lat. locare ‘to set, to place; to let’), developed the new sense ‘to rent’ of (22b) that involves a change of perspective within the same frame. Certainly, auto-conversion is a figure/ground effect and, thus, a kind of metonymy, but a rather complex one: first, 14 because of the overlapping of figure and ground;17 second, because it is a lexical metonymy with grammatical implications, namely valency alternation (1.3.2.).18 As reanalysis frequently implies a metonymic process (cf. Waltereit 1999; Detges/Waltereit 1999; s. also below 3.1.), it is not surprising that auto-conversions of type (22a/b) are probably triggered by reanalyses of structures focussing on a sort of conceptual and valencial pivot, which is the RENTED OBJECT (cf. Waltereit 1998, 77-79): (22) c. Fr. Appartement à louer. ‘Apartment to let/to rent.’ The constellation of Figure 2c is not entirely restricted to verbal auto-conversions. It also occurs, for instance, in the two senses of the French noun hôte, expressing two complementary roles in the frame of HOSPITALITY (23); and it may perhaps solve the mystery of the two apparently contrary senses of the Latin adjective altus (24), that in reality correspond to two different perspectives on potentially the same GREAT SPATIAL EXTENSION (cf. Blank 1997, 279): (23) a. Fr. hôte ‘host’ b. Fr. hôte ‘guest’ (24) a. Lat. altus ‘high’ b. Lat. altus ‘deep’ If we radicalize the constellation of Figure 2c, we can imagine that one of the two reference points/perspectives (X) becomes coextensive with the whole frame (F = X), whereas Y remains an element of F, as represented in Figure 2d. This corresponds, for instance, to the PART – WHOLE and the WHOLE – PART metonymies already listed in 2.1. These are illustrated by verbs in (14) and (15). While it seems difficult to find adjectival examples, there is no lack of nouns, as, e.g.: (25) a. AncGr. pus, podós > ModGr. pó*i ‘foot’ b. ModGr. pó*i also ‘leg’ [with many parallels in other languages] In rhetorical tradition, tropes involving PART – WHOLE relations have been also subsumed under ‘synecdoche’. But this latter descriptive category has proved to be extremely heterogeneous, comprising also semantic-cognitive effects based on relations like SPECIES – GENUS, SINGULAR – PLURAL, and possibly others (cf. Lausberg 1973, §§ 572-577). Yet, there is overwhelming evidence that ‘synecdoche’ as a cover term for SPECIES – GENUS as well as PART – WHOLE effects is untenable. Actually, it is necessary to 17 Though verbalizing basically the same number of participants, (22a) and (22b) already display a different internal figure/ground gradation: in (22a) it is the LANDLORD (la société immobilière) that is highlighted as an AGENT, in (22b), it is the TENANT (l’étudiant); in (22a) it is the TENANT that is facultative (cf. La société immobilière a loué cet appartement), in (22d) it is the LANDLORD (cf. L’étudiant a loué cet appartement); etc. Therefore, the metonymical figure/ground effect of auto-conversion consists in the reversal of a complex figure/ground gradation. Blank (1997, 278) speaks of ‘internal metonymy’. 18 Due to the verb-inherent phenomenon of valency, the complexity of auto-conversion is reflected in a reversal of grammatical valency: the LANDLORD is expressed as a subject in (22a) and as an indirect object in (22b) – and vice versa for the TENANT. 15 make a strict distinction between the two cognitive modules represented in Figure 3a and 3b: GENUS SPECIES Figure 3a FRAME SPECIES ELEMENT ELEMENT Figure 3b Module 3a represents the two possible taxonomic relations, SPECIES – GENUS and SPECIES – SPECIES; module 3b represents the two possible (contiguity) relations within frames, ELEMENT – FRAME and ELEMENT – ELEMENT, that we can call ‘engynomic’ (cf. Koch 1998, 121, and Ms.). Taxonomy is concerned with the conceptualization of reality in categories, on different levels of abstraction, while engynomy has to do with the conceptualization of reality in experiential wholes, independently of levels of abstraction (cf. also Seto’s (1999) distinction between C-relations and E-relations19). Taxonomic processes are due to effects of abstraction (extensions SPECIES → GENUS), of concretizing (restrictions GENUS → SPECIES), or of similarity (SPECIES → SPECIES20), whereas engynomic processes (metonymies) are due to figure/ground effects (ELEMENT → ELEMENT, according to Figure 2b, or ELEMENT → FRAME and FRAME → ELEMENT, according to Figure 2d). Partonomies (or meronomies: cf. Cruse 1986) are clearly o n e type of ELEMENT – FRAME relations according to the engynomic module 3b and therefore do not have anything to do with the taxonomic module 3a. The relation of ‘inclusion’ evoked by Meyer (1993, 166-172) in order to save the unity of traditional synecdoche (PART – WHOLE effects being a zone of overlapping between synecdochical inclusion and metonymical contiguity) is misleading, because the conceptualization of taxonomic relations in terms of ‘inclusion’ (and their representation by Euler diagrams) is merely a scholars’ partonomic metaphor that has no relevance to the everday consciousness of speakers (cf. also Seto 1999, 94). Consequently, PART – WHOLE „synecdoches“ ought to be sharply separated from taxonomic processes and definitely assigned to the category of ‘metonymy’, where they belong to the subtype represented in Figure 2d (cf. Le Guern 1973: 29-38; Sato 1979; Henry 1984, 19 f.; Kleiber/Tamba 1990, 10 f.; Croft 1993, 350; Blank 1997, 253-255; Koch 1999a, 153 f.; Nerlich/Clarke 1999 197-203; Radden/Kövecses 1999, 30-36; Seto 1999).21 19 However, note that Seto’s C-relations correspond only to GENUS – SPECIES, whereas E-relations cover FRAME – as well as ELEMENT – ELEMENT. The necessity of integrating, in a symmetrical way, SPECIES – into the taxonomic module is also supported by the existence of co-hyponymous transfers (s. n. 20). ELEMENT SPECIES 20 This corresponds to the diachronic process of co-hyponymous transfer (cf. Blank 1997, 207-216). 21 In rhetorical tradition, the PART – WHOLE process seems to have been just the most constant component in varying characterizations of ‘synecdoche’ (cf. Meyer 1993/95, I, 74 f.). So, if one wants to maintain the label 16 This choice is not only theoretically satisfying, but also more practicable, because the line between PART – WHOLE relations and other contiguity relations cannot always be easily drawn (cf. Le Guern 1973, 29; Meyer 1993/95, II, 171 f.; Koch 1999a, 154; Seto 1999, 96): is the bank a place near the river or is it part of a frame RIVER (cf. (21)) ? is a garment part of a frame PERSON or not ? is the material wood part of a stick or not ? is the counter part of the public house or is it only located therein ? etc. Regardless of our answer to these questions, all these relations are potential bases for metonymic figure/ground effects. 2.3. Concept and referent We defined ‘contiguity’ as a „cognitive relation“. However, since metonymy is an effect concerning linguistic signs, we have to specify its semiotic impact ((20)(ii)). As content counterparts of a linguistic form, there are three entities that ought to be distinguished on the „semantic“ side of semiosis (cf. Raible 1983, 5, and s. above n. 2): 1° the signifié (a virtual linguistic entity that together with the form, i.e., the signifiant, forms the linguistic sign); 2° the conceptual designatum expressed (a virtual extralinguistic entity); 3° the referent the linguistic sign refers to (an actual, individual extralinguistic entity). Even if one considers the distinction between the linguistic signifié and the extralinguistic concept to be indispensable, metonymy can still not be accounted for in terms of relations between signifiés (cf. above n. 2 and Koch 1996a; 1996b; 1998, 115125; 1999a, 144 f.). As metonymy activates extralinguistic knowledge of the speakers, it is clearly based on contiguity relations between conceptual designata. The third entity involved in semiosis, the actual, individual referent, seems to be included in metonymic processes only insofar as it is subsumed under a concept in whose contiguity and frame properties it participates. Nevertheless, despite the conceptual nature of contiguities, the entity ‘referent’ is not totally indifferent to metonymic processes. In (12), for instance, the conceptual figure/ground effect with respect to buses is necessarily accompanied by a shift from one (class of) referent(s) to a totally distinct (class of) referent(s). In contrast to this, the examples presented under (26) and (27), that are sometimes discussed under the heading ‘metonymy’, do not necessarily imply any shift of reference with respect to the „literal“ (holistic) reading of Paul and dog, respectively: (26) Engl. Paul is tanned. [sc. Paul’s skin] (27) Engl. The dog bit the cat. [sc. the dog’s teeth] According to Langacker (1987, 271-274; 1993, 29-35), predications like these do not involve anything other than a restricted ‘active zone’ of the global referent of Paul, the dog etc. Now, as Kleiber (1991; 1999, 99 f., 124, 142-146) could demonstrate, certain characteristics of certain parts are able to characterize the whole. Accordingly, the ‘synecdoche’ for the PART – WHOLE subtype of metonymy, a new term for SPECIES – GENUS processes has to be found. In contrast to this, Nerlich/Clarke (1999) restrict the term ‘synecdoche’ just to SPECIES – GENUS processes. 17 normal way to refer to these (salient) parts/active zones is to choose a form referring literally to the whole, without any real shift of reference. Kleiber calls this principle métonymie intégrée. Thus, we have to posit a first important distinction between ‘referent-sensitive’ processes as in (12) and ‘non-referent-sensitive’ processes as in (26) and (27). A second important distinction can be gathered from the two following well-known examples: (28) Engl. His native tongue is German. [sc. language] (29) Engl. The ham sandwich is getting restless at table 20. [sc. the customer who ordered a ham sandwich] Both tongue (28) and ham sandwich (29) undergo a shift of reference in comparison to their respective „literal“ meaning and are therefore ‘referent-sensitive’. In (28), however, the shift of reference is only a by-product of a more important fact: the figure/ground effect opened up by tongue constitutes a particular lexical-conceptual solution to express the concept (FORM OF) LANGUAGE. It once started as an ad hoc metonymy that was then lexicalized in a metonymic change, leading to today’s metonymic polysemy of tongue (s. above 1.1.). Even if it never had been lexicalized, it always would be potentially relevant for the lexicon of English. In this sense, we can denominate metonymies like (28) ‘concept-oriented’. In contrast to this, the conceptual figure/ground effect achieved in (29) serves only the purpose of finding an expedient form that guarantees accessibility of the intended referent in a given discourse. Therefore, metonymies of the ham sandwich type are not just ‘referent-sensitive’, but also ‘referent-oriented’ (buses in (12) belongs to the same type). In these cases, the distinction between ad hoc metonymy and metonymic polysemy does not apply. Within a given context and on the basis of relevant frame knowledge, a referent-oriented metonymy is always possible, and the problem of lexicalizing, for instance, ham sandwich as a conceptual „solution“ for CUSTOMER WHO ORDERED A HAM SANDWICH or bus for BUS DRIVER does not even arise. To put it more precisely, referent-oriented metonymies do not pertain at all to the realm of the lexicon, but to discourse semantics (s. above 1.3.7.). They simply represent one type of referring expressions that assure textual coherence (cf. Brown/Yule 1983, 212 f.). Referent-oriented metonymies only occur in nouns (or noun phrases), whereas conceptoriented metonymies can be found in different parts of speech and even in more complex constellations: noun (11), (21), (23), (25), (28); verb (2), (4), (14), (15), (19), (22), (30); adjective (24); verb + noun (13); grammaticalization (16); intragrammatical semantic change (17); speech act (18); etc. (for the special case (6)/(7) s. below). Thus, in the valency environment of verbs, we absolutely have to distinguish between referentoriented metonymies like (12) and (29) that only concern the nouns inserted on the discourse level, and concept-oriented metonymies like (30a/b) that concern the 18 polysemous verb itself and its participant roles (cf. the distinction between ‘insertional level’ and ‘role level’ in Waltereit 1998, 55 f.; 1999, 234-236):22 (30) a. Engl. The waiter served the ham sandwich. [role: SERVED DISH] b. Engl. The waiter served the customer. [role: SERVED PERSON] In 0.1. we started from the conceptual figure/ground effect as an essential for metonymy. Now, we have seen that the relevance of concepts, but also of referents to metonymy has to be specified and differentiated. Since all these criteria do not totally coincide, it seems more reasonable to define, within a realm of contiguity-based effects on invariant23 linguistic forms (Figure 4), a central standard type of metonymy (area A), fulfilling the three criteria 1° figure/ground effect, 2° ‘referent-sensitivity’ and 3° ‘concept-orientation’. The ham sandwich type (area C) is less central, because it displays not only just referent-sensitivity (2°), but also referent-orientation that is opposed to concept-orientation (3°). Due to the figure/ground effect (1°), it still can be regarded as a case of metonymy. In contrast to this, the cases of so-called métonymie intégrée ((26), (27) = area E) definitely have to be excluded from the realm of metonymy (cf. also Waltereit 1998, 31-33). They are not referent-sensitive (2°), as we saw above. They do not raise a truly lexical problem and are therefore not concept-oriented (3°; it would be inconceivable, for instance, that dog could be understood as a new lexical solution for DOG’S TEETH). Last but not least, they are characterized by the phenomenon of ‘active-zones’, which is certainly contiguity-based (frame-based), but only constitutes a „silent“ by-product of reference and thus simply does not involve a real figure/ground effect (1°; in contrast to this, the ham sandwich type C truly depends on a figure/ground effect). 22 It is open to question, however, whether ‘role’ metonymies really arise diachronically from ‘insertional’ metonymies, as Waltereit suggests. 23 The criterion of invariance is important, as we saw at the beginning of section 1. We will have to return to this matter in 4.2. 19 2° referent-sensitivity 1° figure/ground effect (29) the ham sandwich (28) tongue C A (referent-orientation) (27) the dog E (7) child B 3° conceptorientation (31) book D contiguity-based effects on invariant linguistic forms Figure 4 Within the area of contiguity-based effects on invariant forms, the figure/ground effect actually seems to be criterial for metonymy. This is confirmed by a fourth type D of contiguity-based24 effect that has been studied intensively in recent years: (31) a. Engl. This book weighs ten pounds. b. Engl. This book is a history of Great Britain. c. Engl. This book, that is a history of Great Britain, weighs ten pounds. The conceptual effect from book in (31a) to book in (31b), is not referent-sensitive (cf. the referential interlacement of (a) and (b) in (c)). According to Croft (1993, 349 f.), cases like this one should not be included in the realm of metonymy. If this is a case of non-referent-sensitivity (and a fortiori of non-referent-orientation), the shift from (31a) BOOK AS A TOME to (31b) BOOK AS A TEXT is yet clearly concept-oriented. Cruse (1996) denominates conceptual variants of this kind as different semantic ‘facets’ of the same word. Semantic differentiations in terms of facets are „deeper“ than mere contextual variation, but „shallower“ than real (metonymic) polysemy of the type A. The problem with ‘facets’ is that, on the one hand, they tend to „proliferate“ depending on the contexts the word is inserted in, and that, on the other hand, they do not have to be 24 Taylor (1995, 125) expressly relates this type of examples to perspectivization within a frame. 20 differentiated at all25 (which happens only exceptionally with type A polysemies) in many rather common contexts, as, e.g., in (31d): (31) d. Engl. I like books. Kleiber (1999, 87-101) has convincingly demonstrated that type D is open to a treatment somewhat analogous to type E (certain ‘active zones’ of a whole are able to characterize that whole). So, there is neither referent-sensitivity (2°) nor a figure/ground effect (1°). The difference between type D and type E resides in the fact that the former is concept-oriented, as we saw above (3°; cf. Figure 4). Types D and E have in common their indifference to diachronic processes (the distinction between ad hoc metonymy and metonymic polysemy in the sense of 1.1. does not apply). What about our examples (6) jour fixe and (7) child (that are of the same kind as in this context) ? At first, they seem to belong to the ‘facet’ type D. Indeed, child is not referent-sensitive (2°), as is shown by (32c). If it is non-referent-sensitive (and a fortiori non-referent-oriented), it is still clearly concept-oriented (3°): it is susceptible to a conceptual shift from (32a) DESCENDANT to (32b) VERY YOUNG PERSON. (32) a. Engl. Although being well advanced in years, John and Mary had a child after all. b. Engl. John and Mary treated the child like an adult. c. Engl. Although being well advanced in years, John and Mary had a child after all. They treated it like an adult. However, child is a typical example of a metonymic change. Diachronically, its first sense corresponds to (32a); the posterior sense to (32b). This is a nonnegligible difference with respect to type D, that is immune to diachronic processes. So, we have to posit for (6) and (7) a separate type B (cf. Figure 4) characterized by a real figure/ground effect (1°) that potentially brings about metonymic change (in the case of the ad hoc metonymy (6) this has not been accomplished in the final analysis; in the case of (7) it has). We can grasp the peculiarity of type B thanks to the model represented in Figure 2d. The non-referent-sensitivity (2°) implies that the reference point is the frame itself (F = X) and that the figure/ground effect concerns different internal aspects of the frame that are coextensive with the whole frame. An important principle governing metonymic change and polysemy is one called ‘inductive generalization’ by Dik (1977): contiguity relations hold only for salient, prototypical members of the conceptual categories involved, but the metonymic process generalizes them to these categories as wholes (cf. Geeraerts 1997, 68 ff.; Koch 1995a, 40 f.; Koch 1999a, 150 f.). This applies, of course, to type A. The contiguity between LANGUAGE and TONGUE, for instance, does not imply that language is always and only realized by means of the tongue (if we think of written language or of the other organs of articulation). The principle of inductive generalization applies to type B in an even more radical sense. As this type is non-referent-sensitive, the (classes of) referents of, 25 Pustejovsky’s proposal in terms of a Cartesian type (1995, 90-95) represents a purely descriptive device for constructing meta-entries. The role of frames, contiguities, and figure/ground effects is not really elucidated. S. also below n. 28. 21 say, child (7) in the two metonymically related senses are potentially identical. But in reality, the extensions of the categories corresponding to the concepts DESCENDANT and VERY YOUNG PERSON overlap only in the most salient and prototypical case: notwithstanding that we are used to conceiving of a very young person as the descendant of somebody, we can also view an older person as someone’s descendant and we can consider a young person independently of his/her lineage. Therefore, the lexicalization of the metonymic sense of child presupposes inductive generalization. These extensional prototypicality effects are unknown to type D, although it is equally nonreferent-sensitive. Whether we have in mind BOOK AS A TOME or BOOK AS A TEXT, the extension of the category is always identical. 26 Another feature that clearly distinguishes type B from types C and E (and probably D, as well) and that associates it with type A is the fact that it is not confined to nouns (as in (6), (7), and (33)), but can be realized also in other parts of speech (as in (34) and (35)):27 (33) a. MEngl. boor ‘peasant’ b. ModEngl. boor ‘coarse or awkward person’ (34) a. Germ. billig ‘appropriate’ b. Germ billig ‘cheap’ c. Germ. billig ‘worthless’ [two metonymic steps (a) > (b) > (c)] (35) a. Engl. There is a book on the table. b. Engl. There is no bread. [LOCATION] [EXISTENCE] (cf. Koch 1999b, 293 f.) All in all, metonymy is to be considered a prototypical notion, as represented in Figure 4. Type A, combining all three criteria 1°, 2°, and 3°, is the central instance of metonymy, whereas type B, lacking 2°, and type C, lacking 3°, are less central. Since the figure/ground effect (1°) is absolutely compulsory, types D and E have to be excluded from the realm of metonymy (crosshatched area in Figure 4).28 3. Metonymy as generated by pragmatic processes 26 In summary, we can say, from the referential point of view, that, firstly, with type A, the extensions of the two concepts involved are, of course, totally disjunct, secondly, with type B, the respective extensions overlap and, thirdly, with type D, they are totally congruent. 27 Note that in a case like (34) or (35), the ‘non-referent-sensitivity’ typical of type B does not concern the reference qualities of the adjective or verb, but those of the accompanying head noun or verbal participants. 28 The systematics represented in Figure 4 also reveals the insufficiency of Pustejovsky’s explanation in terms of ‘qualia structure’ (1995, 31 f., 85-104; s. also above n. 25), that describes in the same way, for instance, cases of type A (e.g. Engl. lamb, fig) and cases of type D (e.g. Engl. door, window). – A further type of contiguity-based effect (e.g. My ex-husband is parked on the upper deck. He is taking the bus today.) has been analyzed as metonymy in noun phrases (Nunberg 1978) or, alternatively, as metonymy in the predicate (!) (Nunberg 1995). The latter solution has been criticized, e.g., by Kleiber (1999, 121-148), who treats this type according to the principles of métonymie intégrée and of ‘active zones’ (which would correspond to our n o n - metonymic type E in Figure 4), and by Panther/Radden (1999, 10 f.), who insist on the metonymic character of the noun phrase (which would correspond to our type C), explaining its anaphoric behavior by the principle HUMAN OVER NONHUMAN. Cf. also Waltereit 1998, 56-58. 22 We have considered metonymy to be „generated by pragmatic processes” ((3)(iii)). Actually, the respective figure/ground effects are exploited or simply occur depending on different speakers’ or hearers’ intentions, needs, or reactions. Here, we have to take into account at least the following aspects: (36) Metonymies (i) can be induced by speakers vs. by hearers; (ii) vary in their relation to Grice’s conversional maxim of relevance. From these points, we can determine two essential problems with regard to metonymy as generated by pragmatic processes (3.1.-3.2.). 3.1. Metonymies: speaker-induced vs. hearer-induced The millenary interest of rhetoricians in metonymy as a ‘trope’ suggests that it is always a device of linguistic expression chosen by speakers. In fact, this view is valid for many metonymies. Let us take just three of the examples already cited (that in other respects are very different from each other). The speaker who created the metonymy (14b) Germ. verstummen ‘to silence’ was searching for an expedient expression and conceptualization for the action he referred to; he could just as well have chosen the already existing, but awkward construction zum Verstummen bringen. The speaker who uses the metonymy (29) Engl. ham sandwich avails himself of a means of reference that works very well – but only – in a strong institutional context. The speaker who once invented the speech act metonymy of the type (18) Engl. Can you help me ? had particular pragmatic motives (to which we will come back in 3.2.) for verbalizing the speech act in an indirect manner instead of choosing the direct form Help me ! Thus, in all these and in many other cases, the figure/ground effect is a speaker’s choice. Yet, metonymies are not neccessarily induced by speakers. In the case of (6) jour fixe ‘day of boring promenades and sightseeing’, it is the son who has reinterpreted the form heard in the context of the frame WEEKEND DAY (and who uses it accordingly in Och nein, bitte keinen jour fixe !). This type of metonymy is obviously induced by the hearer, who is confronted with an utterance of his interlocutor in a given setting. He apprehends the over-all pragmatic sense of the utterance exactly as it is meant by the speaker, but reconstructs the conceptual meaning of one of its elements via a figure/ground effect with respect to the speaker’s view, without, however, affecting the general pragmatic sense (cf. Koch 1999a, 155 f.). He then in turn, as a speaker, applies the metonymy in one of his own utterances. That this is not mere fancy, can be illustrated by a passage from an authentic historical text: in (37b), OSpan. pregón can still express the concept HERALD, but notwithstanding the over-all sense of the utterance, it also could have highlighted the contiguous concept ANNOUNCEMENT within the same frame (cf. also Blank 1997, 246, 252). 23 (37) a. OSpan. pregón ‘herald’ (< Lat. praeco ‘herald’) b. OSpan. Por Castiella oyendo van los pregones ... (Poema de Mío Cid, 287, cit. DCECH, s.v. pregón) ‘All over Castile you are hearing the heralds/the announcements.’ c. OSpan. and ModSpan. pregón ‘announcement’ In the context of grammaticalization theory, this hearer-based procedure has also been termed ‘pragmatic strengthening’ or ‘conventionalization of conversational implicatures/inferences’ (Traugott/König 1991, 193 ff.). More precisely, we are dealing here with processes of reanalysis. Waltereit (1999) and Detges/Waltereit (1999) point out that reanalysis, first of all, is a hearer-induced, primarily semantic process on the basis of an invariant sound chain referring to an invariant state of affairs and, secondly, that it frequently implies a metonymic figure/ground effect. Conversely, we could claim that all hearer-induced metonymies are a sort of reanalysis. Since Detges and Waltereit could establish that rebracketing and category relabeling in the sense of Langacker (1977, 58) are not criterial for reanalysis, this notion can be applied not only to grammar, but also to lexical units, as in (6) and (37b). Our example (22c) shows, by the way, a reanalysis having the form of a hearer-induced lexical metonymy with grammatical implications (according to 1.3.2.).29 3.2. Metonymies and the maxim of relevance It is elucidating to study the different ways metonymies function with respect to Grice’s (1975) well-known conversational maxims, especially the maxim of relevance (cf. esp. Sperber/Wilson 1986). 3.2.1. Hearer-induced metonymies (3.1.) are simply compatible with the maxim of relevance. The hearer imputes relevance to the utterance of the speaker and chooses for the critical element a conceptual interpretation that beyond the figure/ground effect is consistent with the pragmatic over-all sense. 3.2.2. For speaker-induced metonymies, the pragmatic problem of relevance, discussed in the following, arises only at the moment they are introduced ad hoc (1.1.). (Once they are conventionalized, there may still be a need for disambiguation, but not necessarily a problem of relevance.) 126.96.36.199. Certain speaker-induced metonymies need repair and, indeed, are reparable, thanks to the maxim of relevance that enables the hearer to counterbalance the figure/ground effect. 188.8.131.52.1. A first type is represented by referent-oriented metonymies, as (12) the buses or (29) the ham sandwich (type C in Figure 4, that is necessarily speaker-induced). 29 Further examples of hearer-induced metonymies are: (2), (7), (19), (21), and (34b/c). The status of ex. (17) in this respect is still in need of investigation. – The phenomenon of ‘implication’ that Warren (1992, 51-72, 101; 1998) explicitly distinguishes from ‘metonymy’ probably coincides with hearer-induced metonymies of type B in Figure 4. In my opinion, these ‘implications’ can nevertheless be regarded as metonymies, because the criterion of the figure/ground effect (that they fulfill) is more decisive for metonymy than those of speaker-induction and of referent-sensitivity (s. above 2.3.). 24 Here, the speaker resolves, above all, a referential task relying on the maxim of relevance. 184.108.40.206.2. A second type recovers part of the concept-oriented metonymies (types A and B in Figure 4), where the speaker resolves, above all, a lexical conceptualization task relying on the maxim of relevance. Depending on the intervention – or not – of further pragmatic problems on the part of the speaker, we can distinguish two subtypes. First, the conceptualization task can be confined to a merely lexical problem: there may be imprecise conceptualizations ((23) hôte, (25) pó*i), preferences for „expedient“ solutions ((14) verstummen, (15) close, (24) altus, (30) serve), salience effects ((28) tongue), lexical gaps, etc. Second, the conceptualization task is accomplished under the pressure of additional pragmatic constraints as taboos ((13) wash my hands), politeness ((18) Can you help me ?), etc. (For the bases of all these and other motivations, cf. Blank 1997, 345-405; 1999b; Brown/Levinson 1987; Allan/Burridge 1991). 220.127.116.11. But there still remains a separate group of speaker-induced, concept-oriented metonymies (types A and B) that are n o t i n t e n d e d to be repaired in the name of the maxim of relevance (and possibly other maxims like that of quantity): ((4) craindre, (11) power, (16) wilen, (33) boor, (34a/b) billig). In these cases, the figure/ground effect produces expressiveness (cf. also Mair 1992; Koch/Oesterreicher 1996; Geeraerts 1997, 104-106; Blank 1999b, 63-66, 80-82). It becomes a device of everyday rhetoric (cf. Stempel 1983) that simply i s i n t e n d e d to flout the conversational maxim(s) in order to provide scope for enhancement of the speaker’s own image, for persuasion, for social role assignment, etc. – as long as expressiveness is recognized as such.30 4. The realm of metonymy We have circumscribed the realm of metonymy, defining it as a linguistic effect upon the content of a given form, based on a figure/ground effect within conceptual frames and generated by pragmatic processes. This definition allowed us to observe the linguistic, conceptual, referential, and pragmatic modulations of the figure/ground effect and to demonstrate, thereby, the wide range and the unity in diversity of metonymy. On the other hand, the figure/ground effect allowed us, within the realm of contiguity-based effects on invariant linguistic forms, to definitively exclude from metonymy certain ‘active zone’ phenomena (types D and E in Figure 4). From the point of view of Figure 4, metonymy seems to constitute a (central) prototypically structured subtype of contiguity-based effects on invariant linguistic forms. But this circumscription raises two further questions: 1° Must there always be a linguistic form of expression in metonymies ? (4.1.) 2° Does this expression always have to be invariant in all respects ? (4.2.) 30 As is well known, linguistic expressiveness tends to erode over the course of time. Interestingly, Detges/Waltereit (1999) point out that the inevitable loss of expressiveness is due to a kind of reanalysis on the part of the hearer. 25 4.1. Metonymy as a linguistic phenomenon It has been claimed that metonymies occur in and between different „ontological“, i.e., semiotic, realms (Radden/Kövecses 1999, 23-29). Example (38), for instance, would represent a ‘metonymic’ relation between a linguistic form and a concept that is at the base of any linguistic sign. (39) would represent a ‘metonymic’ relation between a concept and a referent that underlies the referential function of any linguistic sign. (38) form Fr. arbre — concept TREE (39) concept TREE — referent that is a tree According to Gibbs (1999, 66), „people think in metonymy“, and he posits „metonymic models of thinking“, which would presuppose ‘metonymic relations’ between concepts, as for example: (40) concept HAIL A TAXI — concept DRIVE IN A TAXI I think that all these considerations overextend the notion of ‘metonymy’. When we ascertained in (3)(ii) that metonymy is b a s e d on contiguity relations, this implies that metonymy is not identical with contiguity. The latter is a static relation, whereas the former is a dynamic linguistic process. Without any doubt, (38)-(40) exemplify relations of contiguity. Linguistic signs develop from a habitualized contiguity between a form and a concept (38). 31 Yet, according to Keller (1995, 160-180) the emergence of signs (symbolification) pertains to a very deep semiotic layer that is logically and genetically prior to metonymization. We should not call both processes ‘metonymy’, even if we assign contiguity an important role in symbolification, and this all the more as only one relatum of the contiguity relation in symbolification is conceptual in nature. Obviously, similar problems arise for (39), even if we are willing to admit that a concept (expressed by a linguistic form) stands in a relation of contiguity to a referent. As for (40), concepts (or sub-frames) belonging to the same frame are undeniably contiguous. But, as Radden/Kövecses (1999, 27) put it cautiously, „it is not clear at the moment if the form of a category [i.e., the signifiant ] is required in order to perform the metonymic shift or if this metonymy can also operate at the purely conceptual level.“ The same problem has arisen for metaphor (cf. Lakoff/Johnson 1980, 4, and the critical remarks in Koch 1994, 213). Though underlining the conceptual basis of metonymy, we should acknowledge the linguistic nature of the metonymical process as a whole. This is supported especially by the considerations in section 1 and, furthermore, by the fact that contiguities do not automatically lead to metonymies (cf. types D and E in Figure 4). In this way, our cognitive conception of metonymy remains, in spite of all, compatible with the traditional conception implying the existence of a linguistic form undergoing metonymy. 31 However, as underlined in 2.3., the linguistic sign itself consists of a form (signifiant) and a signifié. 26 4.2. Metonymy and the invariant linguistic form If we insist on the existence of a linguistic form in metonymy, this presupposes invariance of the respective form. At the beginning of section 1, we already resolved the (purely external) problem of sound change. But there are morphological aspects to be considered, too. What about a case like (41), for example ? (41) a. Engl. They ate fish. b. Engl. They will fish salmon. They are repeatedly categorized as metonymies (cf., e.g., Dirven 1999; Gibb 1999, 65; Nerlich et al. 1999, 272 f.; Radden/Kövecses 1999, 36). Surely, there is a clear contiguity relation between the concepts FISH and TO FISH, but strictly speaking, the two forms fish and (to) fish are not identical on the morphological level, because they belong to two different word classes. Normally, we would speak of ‘conversion’ in such a case. A somewhat similar problem is found in the following example in a slightly more complicated form (cf. Koch, in press): (42) a. Ital. pera ‘pear’ b. Ital. pero ‘pear tree’ There is an obvious contiguity relation between PEAR and PEAR TREE . Due to a morphological process, namely gender change, the phonological and morphological shape of the two Italian words is not totally identical, but the lexeme is the same. Shall we call this metonymy ? Fundamentally, the same problem arises, in an even more radical form, in the following examples of suffixation and composition: (43) a. Fr. ferme ‘farm’ b. Fr. fermier ‘farmer’ (44) a. Germ. Brief ‘letter’ b. Germ. Briefmarke ‘stamp’ We could cite analogous examples for number change, prefixation, idioms etc. Beyond any doubt, contiguity relations are the cognitive basis not only of rhetorical effects, referential effects and semantic change, as demonstrated throughout this paper, but also of many lexical processes like (41)-(44) involving a morpho-lexical change of form (cf. Koch 1999a, 157-159). Either we have to subsume all this under the label of ‘metonymy’ (a solution perceptible in Schifko 1979), or we have to restrict ‘metonymy’ to contiguity-based effects on a linguistic form that is really invariant, i.e., we have to exclude phenomena like those in (41)-(44). In view of the figure/ground effect that has turned out to be central for metonymy, this is not a purely terminological issue: ceteris paribus, it is only the conceptual perspectivization that changes in metonymy. Therefore, we should stick to the criterion of total morpho-lexical invariance of linguistic form and choose the more restrictive solution that agrees with the traditional conception of ‘metonymy’. The examples in (41)-(44) are contiguity-based, but not ‘metonymies’. This decision certainly does not 27 prevent us from exploring the relations between metonymy and other contiguity-based phenomena, but it puts metonymy in its right place. REFERENCES Allan, Keith/Burridge, Kate (1991): Euphemism and Dysphemism. Language Use as Shield and Weapon, New York/Oxford: Oxford University Press. Anderson, Stephen (1971): „On the role of deep structure in semantic interpretation“, in: Foundations of Language 7, 387-396. Anscombre, Jean-Claude/Létoublon, Françoise/Pierrot, Alain (1987): „Speech act verbs, linguistic action verbs, and delocutivity“, in: Jef Verschueren (ed.), Linguitic Action, Norwood (N.J.): Ablex (Advances in Discourse Processes, 23), 45-67. Barsalou, Lawence W. (1992): „Frames, concepts, and conceptual fields", in: Adrienne Lehrer/Eva F. Kittay (eds.), Frames, Fields, and Contrasts. New Essays in Semantic and Lexical Organization, Hillsdale (N.J.)/London: Erbaum, 21-74. Blank, Andreas (1997): Prinzipien des lexikalischen Bedeutungswandels am Beispiel der romanischen Sprachen, Tübingen: Niemeyer (Beihefte zur Zeitschrift für Romanische Philologie, 285). Blank, Andreas (1999a): „Co-presence and succession: A Cognitive Typology of Metonymies“, in: Panther/Radden 1999, 169-191. Blank, Andreas (1999b): „Why do new meanings occur ? A cognitive typology of the motivation for lexical semantic change“, in: Blank/Koch 1999, 61-89. Blank, Andreas (2000): „Neuere Entwicklungen in der lexikalischen Semantik“, in: Günter Holtus/Michael Metzeltin/Christian Schmitt (eds.), Lexikon der Romanistischen Linguistik, Tübingen: Niemeyer, I (forthcoming). Blank, Andreas (this volume): "Polysemy in the lexicon" Blank, Andreas/Koch, Peter (eds.)(1999): Historical Semantics and Cognition , Berlin/New York: Mouton de Gruyter (Cognitive Linguistics Research, 13). Bonhomme, Marc (1987): Linguistique de la métonymie, Bern: Lang (Sciences pour la communication, 16). Bréal, Michel (5 1921): Essai de sémantique (Science des significations), Paris: Hachette. Bredin, Hugh (1984): „Metonymy“, in: Poetics Today 5, 45-58. Brooke-Rose, Christine (1958): A Grammar of Metaphor, London: Secker & Warburg. Brown, Gillian/Yule, George (1983): Discourse Analysis, Cambridge etc.: Cambridge University Press. Brown, Penelope/Levinson, Stephen C. (1987): Politeness. Some Universals in Language Use, Cambrige: Cambridge University Press (Studies in International Sociolinguistics, 4). Bühler, Karl (2 1965): Sprachtheorie. Die Darstellungsfunktion der Sprache, Stuttgart: Fischer. Bybee, Joan/Pagliuca, William/Perkins, Revere D. (1991): „Back to the Future“, in: Traugott/Heine 1991, II, 17-58. Bybee, Joan/Perkins, Revere D./Pagliuca, William (1994): The Evolution of Grammar. Tense, Aspect, and Modality in the Languages of the World, Chicago/London: Chicago University Press. Cole, Peter /Morgan, Jerry L. (eds.)(1975): Speech Acts, New York: Academic Press (Syntax and Semantics, 3), Coseriu, Eugenio (1958): Sincronía, diacronía e historia. El problema del cambio linguistico, Montevideo: Universidad de Montevideo. Croft, William (1993): „The Role of Domains in the Interpretation of Metaphors and Metonymies“, in: Cognitive Linguistics 44, 335-370. Cruse, D. Alan (1986): Lexical Semantics, Cambridge: Cambridge University Press. Cruse, D. Alan (1996): „La signification des noms propres de pays en anglais“, Sylvianne RémiGiraud/Pierre Rétat (eds.), Les mots de la nation, Lyon: Presses Universitaires de Lyon, 93-102. Detges, Ulrich (1998): „Echt die Wahrheit sagen. Überlegungen zur Grammatikalisierung von Adverbmarkern“, in: Philologie im Netz 4, 1-125. (http://www.fu-berlin.de/phin) 28 Detges, Ulrich (1999): „Wie entsteht Grammatik? Kognitive und pragmatische Determinanten der Grammatikalisierung von Tempusmarkern“, in: Lang/Neumann-Holzschuh 1999, 31-52. Detges, Ulrich (in press). „Time and truth: the grammaticalization of resultatives and perfects within a theory of subjectification“, in: Studies in Language. Detges, Ulrich/Waltereit, Richard (1999): „Grammaticaliziation vs. reanalysis: a simplest systematics of functional change in grammar“ (manuscript). DCECH = Corominas, Joan/Pascual, José A. (21980-91): Diccionario crítico etimológico castellano e hispánico. 6 vol., Madrid: Gredos. Dik, Simon C. (1977): „Inductive generalisation in semantic change“, in: Paul J. Hopper (ed.), Studies in Descriptive and Historical Linguistics. Festschrift for Winfried P. Lehmann, Amsterdam/Philadelphia: Benjamins (Current Issues in Linguistic Theory, 4), 283-300. Dirven, René (1999): „Conversion as a conceptual metonymy of event schemata“, in: Panther/Radden 1999, 275-287. Dirven, René/Goossens, Louis/Putseys, Yvan/Vorlat, Emma (1982): The Scene of Linguistic Action its Perspectivization by SPEAK, TALK, SAY and TELL, Amsterdam/Philadelphia: Benjamins Pragmatics and Beyond, III, 6). Dubois, Jacques et al. (1970): Rhétorique générale, Paris: Larousse. Fillmore, Charles J. (1968): „Lexical entries for verbs“, in: Foundations of Language 4, 373-393. Fillmore, Charles, J. (1975): „An alternative to checklist theories of meaning“, in: Proceedings of the Annual Meeting of the Berkeley Linguistics Society 1, 123-131. Fillmore, Charles J. (1977): „Scenes-and-Frames-Semantics“, in: Antonio Zampolli (ed.), Linguistic Structures Processing, Amsterdam: Benjamins (Fundamental Studies in Computer Science, 5), 55-81. Fillmore, Charles J. (1985): „Frames and the semantics of understanding“, in: Quaderni di semantica 4, 217-240. Fontanier, Pierre (1977): Les Figures du Discours, Paris: Flammarion. Fritz, Gerd (1998): Historische Semantik, Stuttgart/Weimar: Metzler (Sammlung Metzler, 313). Geeraerts, Dirk (1997): Diachronic Prototype Semantics. A Contribution to Historical Lexicology, Oxford: Clarendon. Gibbs, Raymond W. (1999): „Speaking and thinking with metonymy“, in: Panther/Radden 1999, 61-76. Grice, H.P. (1975): „Logic and Conversation“, in: Cole/Morgan 1975, 41-58. Haiman, John (1980): „Dictionaries and encyclopedias“, in: Lingua 50, 329-357. Haspelmath, Martin (1993): „More on the typology of inchoative/causative verb alternations“, in: Bernard Comrie/Maria Polinsky (eds.), Causatives and Transitivity, Amsterdam/Philadelphia: Benjamins (Studies in Language Companion Series, 23), 87-120. Heine, Bernd/Claudi, Ulrike/Hünnemeyer, Friederike (1991): Grammaticalization. A Conceptual Framework, Chicago/London: Chicago University Press. Henry, Albert (2 1984): Métonymie et métaphore, Bruxelles: Palais des Académies. Heringer, Hans-Jürgen (1984): „Neues von der Verbszene“, in: Gerhard Stickel (ed.), Pragmatik in der Grammatik. Jahrbuch 1983 des Instituts für deutsche Sprache, Düsseldorf: Schwann (Sprache der Gegenwart, 60), 34-64. Hopper, Paul/Traugott, Elizabeth C. (1993): Grammaticalization, Cambridge: Cambridge University Press. Jakobson, R. (1956): „Two aspects of language and two types of aphasic disturbances“, in: id./M. Halle, Fundamentals of Language, Den Haag/Paris: Mouton 1971, 67-96. Jongen, René (1985): „Polysemy, tropes and cognition or the non-Magrittian art of closing curtains whilst opening them“, in: Wolf Paprotté/René Dirven (eds.), The Ubiquity of Metaphor, Amsterdam/Philadelphia: Benjamins (Amsterdam Studies in the Theory and History of Linguistic Science, IV, 29), 121-139. Keller, Rudi (1995): Zeichentheorie. Zu einer Theorie semiotischen Wissens, Tübingen/Basel: Francke (Uni-Taschenbücher, 1849). Kleiber, Georges/Tamba, Irène (1990): „L’hyponymie revisitée: inclusion et hiérarchie“, in: Langages 98, 7-32. 29 Kleiber, Georges (1991): „Paul est bronzé versus La peau de Paul est bronzé. Contre une approche référentielle analytique“, in: Harro Stammerjohann (Hg.), Analyse et synthèse dans les langues romanes, Tübingen: Narr (Tübinger Beiträge zur Linguistik, 347), 109-134. Kleiber, Georges (1999): Problèmes de sémantique. La polysémie en questions, Villeneuve d’Ascq: Presses Universitaires du Septentrion. Koch, Peter (1981): Verb · Valenz · Verfügung. Zur Satzsemantik und Valenz französischer Verben am Beispiel der Verfügungs-Verben, Heidelberg: Winter (Reihe Siegen, 32). Koch, Peter (1991): „Semantische Valenz, Polysemie und Bedeutungswandel bei romanischen Verben“, in: id./Th. Krefeld (eds.), Connexiones Romanicae. Dependenz und Valenz in romanischen Sprachen, Tübingen: Niemeyer (Linguistische Arbeiten, 268) , 279-306. Koch, Peter (1993): „Kyenbé – tyonbo. Wurzeln kreolischer Lexik“, in: Christian Foltys/Thomas Kotschi (eds.), Berliner romanistische Studien. Für Horst Ochse, Berlin: Freie Universität (Neue Romania, 14), 259-287. Koch, Peter (1994): „Gedanken zur Metapher – und zu ihrer Alltäglichkeit“, in: Sprachlicher Alltag. Linguistik – Rhetorik – Literaturwissenschaft. Festschrift für W.-D. Stempel, Tübingen: Niemeyer, 201225. Koch, Peter (1995a): „Der Beitrag der Prototypentheorie zur Historischen Semantik: eine kritische Bestandsaufnahme“, in: Romanistisches Jahrbuch 46, 27-46. Koch, Peter (1995b): „Aktantielle ‘Metataxe’ und Informationsstruktur in der romanischen Verblexik (Französisch/Italienisch/Spanisch im Vergleich)“, in: Wolfgang Dahmen et al. (eds.), Konvergenz und Divergenz in den romanischen Sprachen. Romanistisches Kolloquium VIII, Tübingen: Narr (Tübinger Beiträge zur Linguistik, 396), 115-137. Koch, Peter (1996a): „La sémantique du prototype: sémasiologie ou onomasiologie?“, in: Zeitschrift für französische Sprache und Literatur 106, 223-240. Koch, Peter (1996b): „Le prototype entre signifié, désigné et référent“, in: Hiltraud Dupuy-Engelhardt (ed.), Questions de méthode et de délimitation en sémantique lexicale. Actes d’EUROSEM 1994, Reims: Presses Universitaires de Reims, 113-135. Koch, Peter (1998): „Saussures mouton und Hjelmslevs træ: zwei Schulbeispiele zwischen Semstruktur und Polysemie“, in: et multum et multa. Festschrift für Peter Wunderli zum 60. Geburtstag, Tübingen: Narr (Tübinger Beiträge zur Linguistik, 428), 113-136. Koch, Peter (1999a): „Frame and Contiguity: On the Cognitive Basis of Metonymy and Certain Type of Wort Formation“, in: Panther/Radden 1999, 139-167. Koch, Peter (1999b): „Cognitive aspects of semantic change and polysemy: the semantic space HAVE/BE“, in: Blank/Koch 1999, 279-305. Koch, Peter (in press): „TREE and FRUIT. A cognitive onomasiological approach“, in: Studi Italiani di Linguistica Teorica ed Applicata. Koch, Peter (Ms.): „Lexical typology“, to appear in: Martin Haspelmath et al., Typology and Universals , Berlin/New York: de Gruyter. Koch, Peter/Oesterreicher, Wulf (1996): „Sprachwandel und expressive Mündlichkeit“, in: Zeitschrift für Literaturwissenschaft und Linguistik 102, 64-96. Koch, Wolfgang/Rosengren, Inger (1996): „Locative alternations“ in English and German: Different lexicalizations of the same conceptual structure, Lund: Germanistisches Institut (Sprache und Pragmatik. Arbeitsbereichte, 43). Köhler, Wolfgang (1947): Gestalt Psychology. An Introduction to New Concepts in Modern Psychology, New York: Liveright. Lakoff, George/Johnson, Mark (1980): Metphors We Live By, Chicago: Chicago University Press. Lakoff, George (1987): Women, Fire, and Dangerous Things. What Categories Reveal about the Mind, Chicago/London: University of Chicago Press. Lang, Jürgen/Neumann-Holzschuh, Ingrid (eds.)(1999): Reanalyse und Grammatikalisierung in den romanischen Sprachen, Tübingen: Niemeyer (Linguistische Arbeiten, 410). Langacker, Ronald W. (1977): „Syntactic reanalysis“, in: Charles N. Li (ed.), Mechanisms of Syntactic Change, Austin: University of Texas Press, 57-139. Langacker, Ronald W. (1987): Foundations of Cognitive Grammar. Vol. I: Theoretical Prerequisites, Stanford: Stanford University Press. 30 Langacker, Ronald W. (1993): „Reference-point constructions“, in: Cognitive Linguistics 4, 1-38. Lausberg, Heinrich (2 1973): Handbuch der literarischen Rhetorik, München: Hueber. Le Guern, Michel (1973): Sémantique de la métaphore et de la métonymie, Paris: Larousse. Lehmann, Christian (1995): Thoughts on Grammaticalization, München: Lincom (LINCOM Studies in Theoretical Linguistics, 1). Levin, Samuel (1977): The Semantics of Metaphor, Baltimore: John Hopkins. Levinson, Stephen C. (1983): Pragmatics, Cambridge: Cambridge University Press. Liebert, Wolf-Andreas (1992): Metaphernbereiche der deutschen Alltagssprache. Kognitive Linguistik und die Perspektiven einer Kognitiven Lexikographie, Frankfurt/M. etc.: Lang (Europäische Hochschulschriften, 1, 1355). Mair, Walter N. (1992): Expressivität und Sprachwandel. Untersuchungen zur Rolle der Subjektivität in der Entwicklung der romaniischen Sprachen, Frankfurt/M. etc.: Lang (Europäische Hochschulschriften, 178). Meillet, Antoine (1905/6): „Comment les mots changent de sens“, in: id., Linguistique historique et linguistique générale, Paris: Champion 1948 (Collection Linguistique, 8), 230-271. Metzger, Wolfgang (1986): Gestaltpsychologie. Ausgewählte Werke aus den Jahren 1950 bis 1982. M. Stadler/H. Crabus (eds.), Frankfurt/M.: Kramer. Meyer, Bernard (1993/95): Synecdoques. Etude d’une figure de rhétorique, 2 vol., Paris: L’Haramattan. Minsky, Marvin (1975): „A framework for representing knowledge“, in: Patrick H. Winston (ed.), The Psychology of Computer Vision, New York etc.: McGraw-Hill, 211-277. Mossé, Fernand (1969): Handbuch des Mittelenglischen. Übersetzt von Herbert Pilch und Ursula Siewert, München: Hueber. Nerlich, Brigitte/Clarke, David D. (1999): „Synecdoche as a cognitive and communicative strategy“ in: Blank/Koch 1999, 197-213. Nerlich, Brigitte/Clarke, David D./Todd, Zazie (1999): „‘Mummy, I like being a sandwich’. Metonymy in Language Acquisition“, in: Panther/Radden 1999, 361-383. Nunberg, Geoffrey D. (1978): The Pragmatics of Reference, Bloomington: Indiana University Linguistics Club. Nunberg, Geoffrey D. (1995): „Transfers of meaning“, in: Journal of Semantics 17, 109-132. Panther, Klaus-Uwe/Radden. Günter (eds.)(1999): Metonymy in Language and Thought, Amsterdam/Philadelphia: Benjamins (Human Cognitive Processing, 4). Panther, Klaus-Uwe/Thronburg, Linda (1999): „The potentiality for actuality metonymy in English and Hungarian“, in: Panther/Radden 1999, 333-357. Pirazzini, Daniela (1997): Cinque miti della metafora nella Übersetzungswissenschaft. Problemi di traduzione delle immagini figurate nella coppia di lingue: Tedesco (Lingua di partenza) – Italiano (Lingua d’arrivo), Frankfurt/M. etc.: Lang (Europäische Hochschulschriften, IX, 27). Prandi, Michèle (1992): Grammaire philosophique des tropes. Mise en forme linguistique et interprétation discursive des conflits conceptuels, Paris: Minuit. Pustejovsky, James (1995): The Generative Lexicon, Cambridge, MA/London: The MIT Press. Radden, Günter/Kövecses, Zoltán (1999): „Towards a theory of metonymy“, in: Panther/Radden 1999, 17-59 [= elaborated version of „Metonymy: developing a cognitive linguistic view“, in: Cognitive Linguistics 8 (1998), 37-77]. Raible, Wolfgang (1983): „Zur Einleitung“, in: Helmut Stimm/id. (eds.), Zur Semantik des Französischen, Wiesbaden: Steiner (Beihefte zur Zeitschrift für französische Sprache und Literatur, N.F., 9), 1-24. Roudet, Léonce (1921): „Sur la classification psychologique des changements sémantiques“, in: Journal de Psychologie normale et pathologique 18, 676-692. Sato, Nabuo (1979): „Synecdoque, un trope suspect“, in: Revue d’Esthétique 1-2, 116-127. Schank, Roger C./Abelson, Roger P. (1977): Scripts, Plans, Goals, and Understanding, Hillsdale (N.J.)/New York: Erlbaum. Schifko, Peter (1979): „Die Metonymie als universales sprachliches Strukturprinzip“, in: Grazer Linguistische Studien 10, 240-264. 31 Searle, John R. (1975): „Indirect speech acts“, in: Cole/Morgan 1975, 59-82. Searle, John R. (1979): „Metaphor“, in: Andrew Ortony (ed.), Metaphor and Thought, Cambridge: Cambridge University Press, 92-123. Seto, Ken-ichi (1999): „Distinguishing metonymy from synecdoche“, in: Panther/Radden 1999, 91-120. Seyfert, Gernot (1976): Zur Theorie der Verbgrammatik, Tübingen: Narr (Tübinger Beiträge zur Linguistik, 73) [ 21979]. Sperber, Dan/Wilson, Deirdre (1986): Relevance. Communicatoin and Cognition, Oxford/Cambridge (Mass.): Blackwell. Stempel, Wolf-Dieter (1983): „‘Ich vergesse alles’. Bemerkungen zur Hyperbolik in der Alltagsrhetorik“, in: Manfred Faust et al. (eds.), Allgemeine Sprachwissenschaft, Sprachtypologie und Textlinguistik. Festschrift für Peter Hartmann, Tübingen: Narr (Tübinger Beiträge zur Linguistik, 215), 87-98. Stolz, Thomas (1994): Grammatikalisierung und Metaphorisierung, Bochum: Brockmeyer (Bochum-Essener Beiträge zur Sprachwandelforschung, 23). Talmy, Leonard (1996): „The Windowing of Attention in Language“, in: Masayoshi Shibatani/Sandra S. Thompson (eds.), Grammatical Constructions. Their Form and Meaning, Oxford: Clarendon, 235287. Tannen, Deborah (1979): „What’s in a frame“, in: Roy O. Freedle (ed.), New Directions in Discourse Processing, Norwood (N.J.): Ablex (Advances in Discourse Processes, 2), 137-181. Taylor, J.R. (2 1995): Linguistic Categorization. Prototypes in Linguistic Theory, Oxford: Clarendon. Taylor, John R. (1999): „Cognitive semantics and structural semantics“, in: Blank/Koch 1999, 17-48. Tesnière, Lucien (1959): Eléments de syntaxe structurale, Paris: Klincksieck. Thornburg, Linda/Panther, Klaus (1997): „Speech act metonymies”, in: Wolf-Andreas Liebert/Gisela Redeker/Linda Waugh (eds.), Discourse and Perspective in Cognitive Linguistics, Amsterdam/Philadelphia: Benjamins (Current Issues in Linguistic Theory, 151), 205-219. Traugott, Elizabeth/Heine, Bernd (eds.)(1991): Approaches to Grammaticalization. 2 vol., Amsterdam/Philadelphia: Benjamins (Typological Studies in Language, 19). Traugott, Elizabeth C./König, Ekkehard (1991): „The semantics-pragmatics of grammaticalization revisited“, in: Traugott/Heine 1991, I, 189-218. Ullmann, St. (1962): Semantics. An Introduction to the Science of Meaning, Oxford: Blackwell. Ungerer, Friedrich/Schmid, Hans-Jörg (1996): An Introduction to Cognitive Linguistics, London/New York: Longman. Waltereit, R. (1998): Metonymie und Grammatik. Kontiguitätsphänomene in der französischen Satzsemantik, Tübingen: Niemeyer (Linguistische Arbeiten, 385). Waltereit, R. (1999): „Reanalyse als metonymischer Prozeß“, in: Lang/Neumann-Holzschuh 1999, 19-29. Warren, Beatrice (1992): Sense Developments. A Contrastive Study of the Development of Slang Senses and Novel Standard Senses in English, Stockholm: Almqvist & Wiksell. Warren, Beatrice (1998): „What is metonymy?“, in: Richard M. Hogg/Linda van Bergen (eds.), Historical Linguistics 1995, Amsterdam/Philadelphia: Benjamins (Current Issues in Linguistic Theory, 162), II, 301-310. Wertheimer, Max (1922/23): „Untersuchungen zur Lehre von der Gestalt“, in: Psychologische Forschungen 1, 47-58. Wilkins, David P. (1996): „Natural tendencies of semantic change and the search for cognates“, in: Mark Durie/Malcolm Ross (eds.), The Comparative Method Reviewed, Oxford: Oxford University Press, 264-304.