Collocations across languages: unity in diversity

Dirk Siepmann Collocations across languages: unity in diversity? 1 Introduction 1 Although somewhat theoretical in nature, the present article starts from a phenomenon commonly observed in second language and translation teaching: second-language learners, translation students and even practising translators experience great difficulty in producing idiomatic output in their second language. One central area of difficulty that has been amply documented in the literature is collocation, the significant co-occurrence of a node (or ‘base’) word and one or more other words in a number of discourse events (see for example Sinclair 1991). Learners have been shown to be prone to negative transfer or ‘interference’ in cases where the target-language collocate is different from the source-language collocate (e.g. Hesselhauf 2003), with the node word or ‘base’ being identical (for example, the incorrect collocation *make a photo is often produced by German learners because the German collocation employs the cognate verb machen). Once they are alerted to the problem and introduced to the use of collocational dictionaries, learners tend to make fewer collocational errors. However, a cursory inspection of any kind of translated material will reveal that even very advanced translators often resort to paraphrases of sourcelanguage collocations in cases where collocational dictionaries fail to record an equivalent. This raises a number of questions, one of which is whether translators’ choices and gaps in dictionaries constitute evidence of gaps in the targetlanguage system. More simply put, it is worth enquiring to what extent, if any, a) collocations can be translated by collocations and b) to what extent collocations and the semantic dimensions they encode are unique to one language. To answer these questions, the present article uses a case study on the collocational range of emotion words. Drawing on a large body of work in lexical semantics, it aims to provide a qualitative and quantitative analysis of intralingual and interlingual differences between two emotion nouns and to quantify the extent of collocational difference between these nouns. The analysis is restricted to English, French and German. Section 2 provides a review of the relevant issues in lexical semantics. Section 3 briefly describes the general methodology employed in the EMOLEX project 1 This contribution was supported by the German Research Foundation (DFG) within the EMOLEX project, no. GZ:BL 178/11-1; AOBJ: 571879. 2 Dirk Siepmann (www. emolex.eu) which underlies this study and the additional methodological steps required for the present study. Section 4 presents the results of the empirical analysis, and Section 5 discusses implications for translation, lexicography and language learning. 2 Literature Review There is a time-honoured concern in linguistics with the description of lexical meaning. Broadly speaking, traditional attempts at describing meaning can be ascribed to two schools of thought which, following Keller (1995), may be termed ‘representationist’ and ‘instrumentalist’. The representationist school, which can be traced back to Aristotle (Putnam 1988: 52), starts from the assumption that signs represent concepts which in turn are direct reflections of an outside reality; signs are thus conceived of as autonomous paradigmatic entities which can be described in terms of referential or propositional semantics. The central thrust of research in this area has been towards a search for semantic features (with seminal studies in the generativist tradition [Katz/Fodor 1963] and in the structuralist tradition [Pottier 1963]) or semantic primitives (Wierzbicka 1996). This kind of analysis has been shown to suffer from severe limitations, not the least being the difficulties involved in defining emotion words such as fear (Meyer 2010: 169). The instrumentalist school, taking inspiration from Wittgenstein’s philosophy of language, equates the meaning of a word with its conventional communicative function or ‘use’. One of the most influential theoretical strands has been British contextualism, which posits that words acquire meanings through their syntagmatic co-occurrence (or ‘collocation’) with other words (Firth 1957: 196). A somewhat stronger statement of this position is Sinclair’s idiom principle (1991), whereby meaning is inherent in ‘semi-preconstructed phrases’ (later termed ‘lexical items’) rather than words. Taking this position to its logical end, Hoey (2005) argues that words, collocations and colligations are primed to co-occur with other words, collocations and colligations and that such primings depend on each speaker’s repeated exposure to particular co-texts and contexts. In this view, meaning arises from the two-way priming between lexico-grammatical items and thoughts or feelings. As Blumenthal (2006) has pointed out, representationist and instrumentalist accounts of meaning, far from being incompatible, are mutually illuminating. This is because the ‘identity’ of a word is determined both by the context in which it occurs (i.e. the syntagmatic relations it contracts) and by its content (i.e. the paradigmatic relations it contracts and the distinctive features that distinguish it from other words). Indeed, it has become almost a linguistic commonplace to assert that content and context are interdependent: “the syntagmatic is the linear projec- Feldfunktion geändert Collocations across languages: unity in diversity? 3 tion of the paradigmatic” (Dubois 1969, cited in Blumenthal 2006); “the paradigmatic is the syntagmatic in another guise” (Heringer 1999, cited in Blumenthal 2006). A simple illustration of this is found in the frequent syntagmatic cooccurrence of words that are paradigmatically related, such as the antonyms rich and poor (cf. Jones 2002). However, as Blumenthal (2006) argues, this still begs the question of what meaning is, given that both the essentially Saussurean notion of distinctive values and the contextualist assumption that meaning is context-dependent are ultimately circular. Thus, if the difference between pretty and beautiful can be ascribed to the presence or absence of particular semantic features, such as ‘+/- aesthetic intensity’, the two words are defined only with reference to each other; there is no fixed point and thus no independent basis for comparison (tertium non datur). The same circularity applies to contextualism, which attempts to define a word on the basis of the words found in its environment; the meaning of these words in turn depends on their environment, and so on. Blumenthal intelligently suggests that it would nevertheless be unwise to assume that a word’s semantic properties do not exist independently of its combinatorial properties. If we did, then we could not explain its combinatorial properties on the basis of its content (cf. for example Fr. avoir peur vs. *avoir crainte). In other words, even though the only way to arrive at a description of a word’s semantic properties is to study its combinatorial properties, it is still necessary to assume a specific semantic ‘identity’ to be able to explain differences in combinatorial behaviour. In an intellectual tour de force, Blumenthal proceeds to show that there are ‘third-party’ entities involved in determining meaning, entities which impact both on a word’s content and on its combinatorial properties. Such tertia can be divided into two major categories, viz. a) cognitive and b) linguistic. Among cognitive tertia we find, for example, the container schema (a schema consisting of a boundary distinguishing an interior from an exterior), the part-whole schema and the sourcepath-goal schema. Another type of cognitive tertium is the classification of our world according to certain criteria and categories, such as the following: - temporal (e.g. states vs. actions: live [somewhere] vs. settle [somewhere])ontological spheres (first-order entities: humans, animals, objects; second- order entities: states-of-affairs, events, processes; third-order entities outside time and space [idea, fact, hypothesis, etc.]) Linguistic tertia include the following: 4 Dirk Siepmann - onomasiological area (e.g. caviar = culinary context; sturgeon eggs = food science, aquaculture) economy (the ability of a word to describe a complex reality in a highly condensed form; satisfied implies a ‘prehistory’ of desire, content does not) - syntactic and semantic roles (subject vs. object; cause vs. effect) - degree of ‘saturation’ [e.g. ?he has conquered] - contiguity (drunk -> alcoholism, red nose, go to one’s head) Blumenthal’s tertia constitute a significant addition to the armamentarium that semanticists, contrastivists and lexicographers can draw upon. Another major contribution to the description of meaning was made by a number of French researchers working on emotion nouns. The studies in question (e.g. Anscombre 1995; Ruwet 1994; Balibar-Mrabti 1995; Leeman 1995; Mathieu 2000; Plantin et al., 2000; Tutin et al. 2006) offer finely honed corpus-based analyses of the differences between nouns such as amour and affection or peur, angoisse and crainte. This has resulted in detailed typologies of emotion nouns, the most sophisticated of which is probably Tutin et al. (2006), who distinguish six classes of emotion nouns on the basis of their collocational and colligational behaviour. Tutin et al. found that the collocates of emotion nouns can be classified by semantic criteria; thus, many emotion nouns typically combine with verbs expressing the ways in which humans exercise control over their emotions (e.g. calm, conquer, dispel, ease, alleviate, overcome), leading Tutin et al. to posit a semantic ‘dimension’ termed ‘affect control’ which helps to distinguish between ‘interpersonal emotion nouns’ such as love or affection and ‘reactive interpersonal emotion nouns’ such as anger or shame. The original list of five dimensions (aspect, control, manifestation, causation, verbalization) identified by Tutin et al. (2006) has been refined and tidied up in the EMOLEX project to yield an eight-way classification, each with several sub-dimensions (see Table 1). Dimension Value Manifestation physical:active physical:passive verbal external emotion manifestation High Low neutral Control Intensity Causation aspect:phasal:inchoative aspect:phasal+intensity:high Collocations across languages: unity in diversity? Dimension 5 Value aspect:phasal+intensity:low aspect:phasal:terminative Verbalization emotive communicative Polarity internal:positive internal:negative external:positive external:negative Aspect punctual:non-iterative punctual:iterative non-punctual :non-iterative non-punctual :iterative phasal:inchoative phasal:continuative phasal:intensity+high phasal+intensity:low phasal:terminative Modes of Experiencing presence:neutral presence:non-neutral absence:neutral absence:non-neutral Table 1: Semantic dimensions underlying emotion words This classification has proved to be an intersubjectively valid and reliable framework for analyzing emotion words in five languages. It allows both intralingual and interlingual comparisons between individual words and between semantic classes of words (see for example Kern [forthcoming] on an interlingual comparison of French and Spanish words denoting rage and anger). One of the age-old questions that this approach promises to answer is that of the translatability of meaning. There have been many claims in the semantics literature of the existence of asymmetries between the conceptual and the lexical levels, or ‘lexical gaps’ (Strohmeyer 1910: 181-190, Leisi 1953), which are often thought to give rise to interlingual divergences. An oft-cited example is the absence of an adjective for the concept ‘no longer thirsty’ in both English and German, while both languages have an expression for ‘no longer hungry’ (full and satt respectively, with a difference in level of formality); other examples include the lack of a term for the space between the upper lip and the nose in several languages (Blank 2001) or the lack of a separate name for the essential part of some artifacts (knife – handle/blade = spoon – handle/?) (Cruse 1986: 171). From a translatological perspective, however, such lexical gaps are of a purely academic nature, since there is usually no actual interlingual asymmetry. Similar considerations apply to gaps such as German Geschwister, which has no everyday one- 6 Dirk Siepmann word English or French equivalent (use of English sibling is heavily contextdependent, though gaining currency); however, it is important to realize that, from a contextualist perspective, the equivalence relation holds between words and word strings of various shapes and sizes, all of which are lexicalized to the same extent; hence, ich habe drei Geschwister is equivalent to I have three brothers and sisters. In other words, what may be an extended lexical unit (in the sense of Sinclair 2004) in one language may correspond to other types of lexicogrammatical configuration in another (see Siepmann 2005 for further examples). With this in mind, talking of a ‘lexical gap’ would seem to make sense only in the case of cultural peculiarities (cf. Siepmann 2003). It is well recognized, for example, that such culture-specific terms as Trauspruch or Berufskolleg do not translate well into French because they designate concepts unfamiliar to French society. It is easy to see that the same goes for the collocations these terms enter into. Since there is no French equivalent for Trauspruch, for example, the same must be true of the collocation Trauspruch + beten. Lexical gaps arising from cultural peculiarities thus regularly give rise to collocational gaps. Such gaps can only be filled by means of complex circumlocutions, as shown in the following example: (1) Der Beter eures Trauspruches ist glücklich, und wir alle wünschen euch von ganzem Herzen, dass ihr beide heute glücklich seid am Tag eurer Trauung. Der Beter ist glücklich darüber, dass er in und von der Gnade Gottes leben darf. In the present case the noun Trauspruch has to be rendered by means of a lengthy paraphrase, and the noun Beter, derived from the verb beten, has to be translated by the more neutral verb prononcer because its dictionary equivalent prier is intransitive. (2) Vous avez choisi ces paroles de l'Ecriture pour témoigner de votre foi et des valeurs que vous souhaitez placer au cœur de votre couple. Celui qui prononce ces paroles est heureux, et de tout cœur nous vous souhaitons d'être heureux en ce jour où s'accomplit votre mariage. Celui qui prononce ces paroles est heureux qu'il lui soit donné de vivre dans la grâce de Dieu. (my translation) What this example shows, however, is that lexical or, more specifically, collocational gaps involving second-order or third-order entities (in the sense of Lyons 1977) can often be filled at the textual level, although making considerable demands on the translator’s linguistic proficiency and ability to think laterally. Similarly, collocational gaps amenable to language-internal constraints (cf. Siepmann 2003: “expressions in which neither the node nor the collocate can serve as a Collocations across languages: unity in diversity? 7 basis for translation”) are not necessarily evidence of the absence of a lexicalized meaning; thus, the French noun dossier ‘file’, when used figuratively in its judicial sense, collocates significantly with the adjective vide ‘empty’2, and there are no English or German equivalents based on either file or Akte. A string like son dossier est vide, at least in a judicial setting, must be rendered by they have (got) nothing against/on him. This kind of equivalence mirrors the operations required to translate word-level lexical gaps by means of collocations; again, the same meaning can be expressed adequately in the two languages under survey but the syntagms involved differ in length and nature. The situation is rather more complex with first-order entities, i.e. cases where gaps arise out of real-world constraints, as in the case of the aforementioned Berufskolleg. To give another example, until a few years ago, windows pivoting either horizontally or vertically were virtually unknown in France, while they have been in common use in Germany for decades. It is not surprising, then, that French still has no standard equivalent for the strong German collocation Fenster + kippen / auf Kipp stellen (window + tilt). Only in the specialist language of the building trade do we find the complex French collocation fenêtre + ouvrir + en soufflet; everyday language has to make do with the somewhat imprecise fenêtre + entrouvrir, which does not allow a distinction between sash windows, casement windows and pivoting windows (cf. Siepmann 2003). Despite the large body of more or less anecdotal evidence on various types of lexical gaps, few systematic comparisons have been made to assess the extent to which languages differ in specific semantic areas. Most of the relevant studies are qualitative, asking for example how a specific semantic area is configured in terms of the number and type of lexical items but relying largely on intuitive criteria or dictionaries for identifying such items (see, for example, Proost 2001 on communication verbs). The question has remained unanswered whether lexical (or collocational) gaps are just a semanticist artifact or whether there are interlingual differences in meaning potential or in the realization/actualization of specific potentials (spread love – ?répandre de l’amour). To answer this question, a sophisticated methodology is needed, which will be presented in the next section. 2 This example was provided by John D. Gallagher (personal communication). I have recently noticed that the collocation in question may also be used in other situations involving the consultation of a ‘dossier’. For example, an orphan carrying out research on his or her parents may find that ‘son dossier (à la DASS) est vide’. 8 Dirk Siepmann 3 Methodology Analysis proceeded in four steps. The first step involved generating the combinatory profile (i.e. the collocational and colligational features) of the emotion words being investigated (see Blumenthal 2006 for a detailed discussion of this concept). In the second step, the resulting lists of collocates were classified according to the eight-dimensional semantic grid described above. Three investigators reviewed the category assignment of collocates independently, with one team member formally coding the results for each lemma. In the third step, an attempt was made to calculate the relative weight of particular semantic categories, with a view to creating contrastive combinatory profiles. In the fourth and final step, interlingual equivalences were established, and the collocates of equivalent nodes were compared with the aim of ferreting out collocational gaps. This final step involved the additional use of larger corpora to correct for potential bias in the 100-million-word corpora used in the project and to determine whether any gaps which emerged were genuinely indicative of interlingual differences in meaning potential. One potential limitation of this study must be acknowledged. This concerns the problems attendant upon the retrieval of English phrasal verbs as collocates of emotion nouns. Thus, for example, the common collocation disappointment + get over was not automatically identified by our corpus enquiry software. One reason for this is that common function words like over are not included in searches; another is that current software does not allow us to detect three-word combinations. It is reasonable to assume, however, that most of these oversights were corrected in the final step, which involved detailed study of actual concordances. One important spin-off resulting from this kind of methodology is the compilation of detailed bilingual lists which are apt to make a substantial contribution to both monolingual and bilingual lexicography. As Table 1 shows, Step 4 results in a systematic survey of lexical items along two dimensions (synonymy and equivalence), allowing the lexicographer to semi-automatize the process of dictionary compilation. It also becomes apparent that such an onomasiological approach to dictionary making has clear advantages over a semasiological approach. Briefly put, the onomasiological approach does not require any lateral thinking on the part of the lexicographer, who normally faces three types of difficulties when working on a small section of an alphabetical list: a) retrieving potential equivalents (or synonyms, as the case may be) from memory on the basis of mere intuition, b) deciding whether these equivalents are largely identical and c) knowing whether his or her colleague who is working on another stretch of the alphabet or in the other translation direction has recorded the same equivalents (for more detailed treatment of this question, see Siepmann 2005). For evidence of the difficulties involved in finding equivalents, it is sufficient to take a cursory look at one of the Collocations across languages: unity in diversity? 9 huge translation corpora (e.g. linguee.de) now available on the Web. A search for Enttäuschung + wegstecken will yield a number of interesting translation proposals but none of them are completely satisfactory. A variety of data sources were laid under contribution. For steps 1 to 3, parallel newspaper and fiction corpora were used, which were specifically designed for the EMOLEX project. The newspaper corpora represent a broad sample of ordinary usage in the British, American, French and German press. Around 80 per cent of the texts were culled from ‘highbrow’ broadsheets, while 20 per cent was collected from regional papers. Texts were selected from across the various daily, weekly or monthly sections appearing in the news media under consideration. Equivalence synonymy semantic cate- German gory control: E. wegstecken emotive English control: emotive E. überwinden / verarbeiten / verkraften get over / overcome / recover from / bounce back from control: manifestation E. herunterschlucken swallow control: manifestation E. (nicht) verbergen / verhehlen shrug off / put behind one (not) hide / disguise / conceal Table 2: An excerpt from the bilingual list compiled for disappointment For step 4, account was taken of the largest available text archives for each language, i.e. the large corpora compiled for various languages by the Sketchengine team, the Google Books Corpora for English, the Kookkurrenzdatenbank of the Institut für Deutsche Sprache for German, and the French corpora compiled by the author. 4 Results The main objective of this study was to determine to what extent, if any, emotion words differ from one language to another. The overall finding is somewhat disappointing from the contrastivist’s perspective and can be summarized as follows: first, there are few significant differences in the distribution of categories across 10 Dirk Siepmann languages; second, there is comparatively sparse evidence of collocational gaps or interlingual difference. Let us now look at the results in more detail. As previously noted, a semantic grid was constructed for each emotion word, which made it possible to compare the relative weight of particular semantic categories in the three languages under survey. Disappointment and its equivalents will be discussed in full detail but less space will be devoted to pride and contempt so as not to overwhelm the reader with excessive information. 4.1 Disappointment/déception/Enttäuschung The pie charts below illustrate the relative weights taken up by each of the semantic categories for the English noun disappointment and the German noun Enttäuschung. The percentages shown are arrived at by adding up the log likelihood values for each semantic dimension. Fig. 1: Distribution of semantic dimensions for disappointment and Enttäuschung The graphs suggest that there are differences between German and English with respect to the relative weight taken up by some dimensions. The table below lists the four most significant dimensions in English and German in descending order of importance: Enttäuschung 1. Intensity (41.9 %) 2. Polarity (23.9 %) 3. Aspect (10.9 %) 4. Manifestation (6.3 %) Disappointment 1. Intensity (44.5 %) 2. Verbalization (16.6 %) 3. Polarity (15.6 %) 4. Aspect (9.6 %) Collocations across languages: unity in diversity? 11 Table 3: Enttäuschung/disappointment: Distribution of semantic dimensions A first thing to note at this point is that typical dictionary definitions tend to focus on just one or two of these dimensions. Thus, both the Oxford Advanced Learner’s Dictionary (OALD) and Duden Universalwörterbuch place particular emphasis on the dimension of polarity, probably because the other dimensions were deemed not to contribute anything of substance to the meaning of the word. Disappointment 1 [uncountable] sadness because something has not happened or been as good, successful, etc. as you expected or hoped 2 [countable] a person or thing that is disappointing (OALD, s.v. disappointment) Enttäuschung a) Nichterfüllung einer Hoffnung od. Erwartung, die jmdn. unzufrieden o.ä. stimmt b) (o. Pl.) das Enttäuschtsein (Duden Universalwörterbuch, s.v. Enttäuschung) This raises the question of whether the quantitative differences in dimensional weighting shown above provide pointers to actual qualitative interlingual contrasts. One might ask, for instance, why the dimension of manifestation is almost completely absent from English disappointment, whilst it is the fourth most important dimension in German. The data show that there are at least three strong German collocations belonging to the category of manifestation whose direct equivalents were not found to be significant in the 100 million word English corpora: die E. war ihm anzusehen / anzumerken, Tränen der Enttäuschung, vor Enttäuschung weinen. This is not to say, however, that these meanings cannot be expressed in English; a cursory examination of other corpora reveals the common occurrence of you could see the disappointment on his face/in his eyes and also throws up some evidence of tears of disappointment and cry in/with disappointment. On the other hand, the significant English collocation disappointment in one’s voice is far more common than its direct German equivalent (Enttäuschung in der/seiner Stimme?), although it may be argued that the meaning in question is covered by die E. war ihm anzumerken. It is of course tempting to assume that such interlingual differences in the relative weighting of particular dimensions are indicative of fundamental cultural divergences. As sociologist Kate Fox (2004: 141) notes for England at least, England is possibly the only culture in the world where no tears at all is entirely normal and acceptable … The English do not measure grief in tears. Too many tears are regarded as somewhat self-indulgent, even a bit selfish and unfair. 12 Dirk Siepmann One might even go so far as to say that, although disappointment, in an English setting at least, may not be manifested through tears, it is still permissible to show it through one’s voice. It would be a reasonable expectation, then, to find the same pattern for other emotion nouns that combine with Tränen in German but, as we have seen, it would certainly be misleading to suggest that collocations expressing the dimension of manifestation do not exist in English, especially since native Anglophone cultures differ in the degree to which emotions are openly expressed. Turning now to the dimension of verbalisation, which appears to be far more prevalent in English, it can be observed that express disappointment in particular is a highly significant collocation in English; the same applies, to a lesser extent, to voice d. and speak of (one’s) d. Here too, however, closer inspection suggests that other factors are at play, namely that the verb express has a general tendency, especially in newspaper language, to team up with emotion words (e.g. admiration, affection, antagonism, concern, dissatisfaction, etc.) and that the literal noun + verb equivalents of these English collocations (emotion noun + äußern/kundtun/zum Ausdruck bringen) are dispreferred in German in favour of verb + adjective/adjectival participle collocations of the type sich enttäuscht zeigen/äußern. What this also shows is that, when comparison is made between two words of the same word class, the apparent interlingual differences found can usually be resolved by an examination of potential class-shifts in translation. The translation procedures underlying such class-shifts (often referred to as ‘transpositions’ in the literature), which have traditionally been based on the translator’s intuition, thus become predictable to a certain extent: a difference in the weighting of dimensions, while clearly not indicative of lexical gaps, often points to the need for transposition. Also noteworthy in the above dictionary definitions is the inclusion, in OALD, of a second sense division (‘a person or thing that disappoints’) which has not been identified as such by German or French lexicographers (cf. the definition of déception in Le Petit Robert: fait d’être déçu; sentiment pénible causé par un désappointement, une frustration). It is, however, fairly easy to find French and German specimens of use that exemplify exactly this kind of meaning: (3) Vor allem das vorher so gehypte Duell zwischen Freiburgs Torjäger Papiss Demba Cissé (15 Tore) und seinem Frankfurter Kollegen Theofanis Gekas (14) war eine einzige Enttäuschung. (www.zeit.de) Although English exhibits a wider variety of punctual-iterative uses illustrating this second sense (e.g. many / myriad / various / repeated / serial / subsequent / another / further/…) than either German or French, English aspectual uses were not found to be more significant in terms of log likelihood. It would therefore seem necessary to introduce a second (or third, as the case may be) sense division in Formatiert: Englisch (USA) Collocations across languages: unity in diversity? 13 German and French dictionaries. Further support for the particular relevance of this sense in English comes from two collocational gaps: first, sad disappointment, whose literal equivalents are almost non-existent in French and German quality texts, and, second, avenge + disappointment, a collocation from the sporting world in which disappointment is almost synonymous with defeat. The closest German and French come to this is using the literal equivalents of defeat and transposing English disappointment to an adjectival modifier: enttäuschende Niederlage + wettmachen and défaite décevante + effacer: Le MHR a ainsi effacé la décevante défaite concédée au Stade Français (32-16) en ouverture … (www.midilibre.fr) Overall, then, there appear to be slight interlingual differences in the relative weighting of semantic dimensions but none of them seriously affect the translatability of items. This brings us to the question of collocational gaps. For each of the languages under survey, potential collocational gaps were identified by comparing the collocates of semantically equivalent nodes; as an illustration, the table below shows the collocational gaps found between English and German: German mit Bier/Wein herunterspülen jdn vor einer ~ bewahren (schnell) verfliegen sich entladen (auf jdn / bei jdm in etw.) der Stachel der ~ (tief sitzen) bevorstehen ? / sich ankündigen ? / warten Sport: enttäuschende Saison; Zeit der ~ Moment schmerzlicher ~ ? verharren / schwelgen in ~ ? schmerzlich/herb/arg Niederlage wettmachen (or: sich für enttäuschende Leistung revanchieren) zum Scheitern verurteilt English drown one’s ~ with spare sb a ~ quickly disappear / fade / wear off / … (vent, take out, bear the brunt) (pain of ~) loom season of ~ pang/twinge of ~ wallow in/dwell on (a) ~ sad ~ avenge doomed to ~ Table 4: Equivalences between German Enttäuschung and English disappointment The following four cases may be distinguished. 1. The collocation can be rendered in the L2 but the rendition involves a different node, suggesting that the L1 node is polysemous. This is the case 14 Dirk Siepmann with doomed to disappointment and avenge one’s disappointment. This may be described as ‘full equivalence’. 2. The collocation can be rendered in the L2 but it is far less significant, as is the case with mit Bier/Wein herunterspülen, which translates as drown one’s disappointment (with a drink/in rum). This collocation does not occur in the English 100-million-word corpus, and there are only around 40 occurrences in the 150-billion-word Google Books corpus. We might refer to this type of equivalence as ‘partial equivalence’. 3. The collocation can be rendered in the L2 but it corresponds to several different word combinations which have traditionally been classified as ‘free’ combinations consisting of two semantically autonomous words (cf. for example Hausmann 1999). An example of this is die Enttäuschung ist (schnell) verflogen, where English has to offer a wide variety of possible renditions: disappear (quickly), fade, wear off, etc. We here have to do with ‘multiple equivalence’. 4. The collocation can be rendered in the L2 but the rendition involves a semantic shift (known as ‘modulation’ in translation studies) which reflects the subtly different perspectives from which speakers of different languages perceive the phenomenon in question. In terms of Mel’čuk’s lexicological theory (see, for example, Mel’čuk/Clas/Polguère 1995), the collocations in question may be described as lacking a functional equivalent in the L2. This is often the case with metaphorical expressions such as Stachel der Enttäuschung, which usually occurs in the multi-word expression der Stachel der Enttäuschung sitzt tief. Although a near-literal, if uncommon rendering is possible (the sting of disappointment was acute / rankled), sting would here be interpreted as referring to the pain caused by disappointment rather than the sting of an insect, which is more prevalent in the German metaphor. Another fairly good example is provided by spread love, whose literal French equivalent répandre de l’amour tends to occur mostly in Canadian French, probably under the influence of English, whereas the mainland French tend to prefer donner or partager de l’amour. Another clear example which will be presented in the next section is provided by German bescheidener Stolz and its exact French equivalent fierté modeste, which need to be rendered by quiet pride in presentday English (as a search on Google Books reveals, modest/humble pride was common from the 15th to the 19th centuries). Turning briefly now to the language pair English-French, we find that the picture is very similar (fig. 2). The rank order of dimensions exhibits some minor differences, with ‘control’ being significantly more common in French than in English. However, these differences do not translate into any tangible qualitative divergences. There are only Collocations across languages: unity in diversity? 15 a few collocational gaps in evidence, some of which parallel the divergences between English and German. Thus, be doomed to disappointment translates as être voué à l’échec into French, and several other collocations are rare but not inexistent in French such as déception persistante (lingering disappointment) or triste déception (more usually déception amère/douloureuse). Conversely, the common French collocations alimenter/nourrir + déception have no clear equivalent in English (?increase / reinforce disappointment, ?make [sb] even more disappointed). Overall, the proportion of collocations which do not map directly onto each other appears to be somewhat higher than in the language pair EnglishGerman, at around 15 per cent Fig. 2: Distribution of semantic dimensions for disappointment and déception Another pertinent example is afforded by the high-frequency collocation la déception était à la hauteur de l’attente, which can only be rendered obliquely by something like sb’s hopes were shattered. Closer inspection of the wider discoursal context in which specific collocations occur throws up a small number of intriguing interlingual differences. Thus, for example, the English collocation avoid + disappointment has literal equivalents in French éviter + deception and German vermeiden + Enttäuschung but is notably common in quite specific stereotypical contexts. Examples: (4) Book early to avoid disappointment! (5) Avoid disappointment and confirm your reservation today. These uses are easily identifiable as directive speech acts commonly found in advertising; within them, the collocation avoid disappointment is a standard lexical 16 Dirk Siepmann item intended to arouse consumers’ desire for instant gratification and their fear of missing out on some great opportunity. The same use appears to be gaining ground in Canadian French (probably under the influence of English) but is still fairly rare in the French of France; the French collocation also has a second sense taking the pattern qqn évite une déception à qqn (sb spares sb a disappointment). The German equivalent is considerably less frequent in newspaper corpora but does occur on the Internet, probably due to interference from English. Here are some examples which may count as more natural German equivalents: (6) Buchen Sie jetzt, damit Sie auf jeden Fall dabei sein können! (7) Lassen Sie sich diese Gelegenheit nicht entgehen und reservieren Sie jetzt! In the case of such co-occurrences as avoid + disappointment + book (v) we are dealing with preferred ways of putting things that, in Sinclair’s idiom, ‘constitute single choices’ rather than word combinations based on linguistic creativity. 4.2 pride/fierté/Stolz For pride/fierté/Stolz, the analysis will be qualitative rather than quantitative and centre on the question of equivalence. The table below shows equivalences between the dimensions found to be relevant to these nouns. E F show, exhibit afficher, montrer, manifester, ne pas cacher, ne pas dissimuler glow (with), puff transparaître (dans sa voix, etc.), rayon(with), swell ner (dans son vi(with), flush sage), émaner, non (with), brim (with), exude, dissimulée ooze, radiate, (-> sentir) palpable, evident, apparent, manifest, visible swallow, ravaler, remballer D DimensionValue demonstrieren, zeigen Manifestation physical:active vor Stolz platzen / strahlen / triefen (-> intensity), mitschwingen, mitklingen, ansehen, anmerken, heraushören, aufblitzen, schwellen; offensichtlich, sichtlich, unverhohlen, unverkennbar, erkennbar, hörbar, sichtbar, spürbar, verhohlen, klammheimlich unterdrücken (mit u.- Control External Emotion Collocations across languages: unity in diversity? E shelve, pocket, set aside F D DimensionValue em Stolz), schlucken ausgeprägt, groß, Intensity grande, immense, profonde, extrême, unbändig, gehörige plein de, exacerbée; Portion déborder de, être empli de; excès de certaine; manquer leise; verhalten de; brin de; pointe de susciter ; (re)tirer; hervorrufen; mit Stolz Causa-tion evoke, instil, foster, promote, donner; faire/être (la erfüllen; fördern cultivate, nur- f. de), procurer (à ture qqn); remplir de ; inculquer; insuffler; sujet; objet; motif Quelle matter (for), source; raviver; envahir qqn cause (for), source (of); surge (of) salvage ; flatter; entretenir, flatter; aufrechterhalten; maintain, pre- garder intacte bewahren serve, retain, sustain, keep alive boost, enhance, build, strengthen, reinforce humble; hurt, blesser, heurter, kränken, verletzen, wound, sting, écorcher, piétiner, ankratzen (angekratzt) dented, hurt, bafouer, titiller; en injured, wound- prendre un coup, être ed; blow (to) malmenée rauben, nehmen crush, smash, enlever shatter, bruise, destroy express exprimer (zum Ausdruck brin- Verbaligen) zation affirmer, clamer, -> mit/voller Stolz revendiquer, procla- verkünden, behaupten, mer, dire hinweisen (etc.) Polarity justifiable, juste, légitime; mo- berechtigt; bescheijustified; real, deste den, maßvoll; freudig genuine fierce, great, immense, considerable, tremendous, terrific quiet 17 Manifestation High Low Neutral aspect:phasal:inchoative aspect:phasal: continuative aspect:phasal+ intensity:high aspect:phasal+ intensity:low aspect:phasal: terminative emotive communicative external:positive Dirk Siepmann 18 E F D false, overweening, arrogant, overbearing, foolish, stubborn, misplaced, childish, juvenile, sinful, perverse arrogante, déplacée, mal placée, naive, stupide, débile, excessive, puérile, insolente, dédaigneuse, maladive, vaniteuse, malsaine falsch, naiv, selbstzerstö-rerisch, kindlich, melancholisch, trotzig, krankhaft, eitel restore, rekindle; regain renaître (Robert); wiedergewinnen (meist recouvrer, retrouver, Partizip) réappropier ; redonner, rendre (a. causative) intact intact, indomptable DimensionValue external:negative Aspect lose; abandon perdre feel, have, take éprouver, ressentir; (p. in) sentiment de civic, local, identitaire, nationale, national, patri- patriotique, materotic; ethnic nelle, paternelle, (AE), gay, virile, masculine, racial; mascu- gaie, etc. ; de mâle; line, macho, se draper dans, family, person- partager al, professional; share non-ponctual:iterative ungebrochen, ungebeugt verlieren empfinden kollektiv, national, patriotisch ; männlich, mütterlich, väterlich, Stolz auf seine Identität, Nationalstolz, etc. Modes of Experiencing phasal:inchoative phasal: continuative phasal: terminative presence:neutral presence:non-neutral The overall impression is one of great conformity. The collocates found fall into the same categories across the three languages under survey; there are few gaps in evidence. One area where German and English exhibit greater flexibility than French is in the external manifestation of pride, with verbs such as radiate (pride)/ooze (pride)/glow (with pride), (vor Stolz) platzen/strahlen/triefen and adjectives of the type evident and (offen-)sichtlich being particularly common. This explains the apparent difficulty experienced by the makers of PONS Großwörterbuch Deutsch-Französisch in translating the following item: (8) der Stolz dieses Vaters auf seine Kinder ist offensichtlich = ?il est évident que la fierté de ce père, ce sont ses enfants (PONS, s.v. Stolz) Collocations across languages: unity in diversity? 19 The lexicographer must have correctly intuited that the literal rendition fierté + évidente, especially in predicative use, is out of step with target-language norms. However, the solution s/he has opted for is problematic, not so much because évident now modifies the entire statement but because the meaning has changed. Whereas the German sentence simply says ‘the father is proud of his children, and this is obvious’, the French sentence implies that the children are the father’s sole or major source of pride. Transposing copular be + adjective to a verb + adverb collocation produces a more accurate translation (note that the German sentence sounds slightly artificial and awkward in the first place): (9) La fierté que ce père éprouve pour ses enfants transparaît clairement. Another example of this kind of intracategorial shift is afforded by humble (v) + pride, a biblical expression (2 Chronicles 32:26; ‘Notwithstanding Hezekiah humbled himself for the pride of his heart’) which still has some currency. Since humble can here be glossed as ‘bring down’, it comes into the category of causation and the subcategory denoting decreasing intensity. Since the metaphor at work here is ‘pride is a person’ (Kövecses 1986: 49), other verbs belonging to the same category will at least render part of the English meaning: German kränken or ankratzen or French écorcher or piétiner. A handful of items might qualify as genuine collocational gaps. One such expression is share + pride/partager + fierté, which has no natural German equivalent. This is confirmed by native speaker’s intuitions about Stolz, which is an inward-directed feeling that cannot be passed on to others within a brief time span. A sentence like the following is almost untranslatable into German, at least under the assumption that we have to do with the literal rather than the figurative sense of share (‘to tell other people about your ideas, experiences and feelings’): (10) They enjoyed sharing their pride in their country with us. (??Sie ließen uns mit Freude an ihrem Nationalstolz teilhaben.) Other contexts are somewhat easier to translate but still ring awkward. An example: (11) Share our pride in the University of Nebraska. (Lassen Sie uns gemeinsam stolz auf die Universität Nebraska sein.) Another potential gap is the French expression se draper dans sa fierté. The expression itself is not listed in Le Petit Robert, but the roughly equivalent item se draper dans sa dignité is glossed as affecter une attitude de dignité offensée, orgueilleuse, for which PONS suggests sich auf seinen Stolz zurückziehen and Langenscheidt proposes sich in beleidigtes Schweigen hüllen. We are here deal- Formatiert: Englisch (USA) 20 Dirk Siepmann ing with a phraseological rather than a collocational gap. Two examples culled from the Internet illustrate current usage: (12) Les réactions vives qui figurent sur cette page montrent bien à quel point les Algériens refusent qu'on leur fasse la moindre critique ou observation. Et encore moins quand cela vient de la part d'un Français. Ils ferment leurs oreilles et se drapent dans leur fierté, 50 ans après c'est inquiétant, il serait temps de passer à autre chose. (http://www.franceculture.fr/emission-le-secret-des-sources-les-mediasalgeriens-face-aux-soulevements-arabes-2012-03-17) (13) Force est de constater, malheureusement, que nos demandes sont restées lettres mortes. Pire, loin d’accepter quelque critique que ce soit, la mairie se drape dans sa fierté et son orgueil et affiche une volonté ostensible de nous ignorer. (http://asbbe.fr/Page_13_Article_T%C3%A9l%C3%A9rama.html) In the case of the first example, both of the above translation proposals may fill the bill. The second example might be translated by ‘die Stadtverwaltung/das Rathaus zeigt sich bürgerfremd (und abgehoben)’. As for English equivalents, the solutions found in dictionaries are unsatisfactory (‘stand on one’s dignity’ which, in modern usage, may be glossed as ‘to demand to be treated with more respect than other people because you think you are more important’, e.g. Despite living in a grand chateau they do not stand on their dignity). By contrast, the literal translation wrap oneself in one’s dignity is attested several thousand times on the Internet and constitutes a perfect equivalent. There is also some evidence of the polysemy of English pride, as shown by the wide variety of English verbs expressing causation and increasing intensity: boost / enhance / strengthen / build / reinforce (etc.) + pride. The literal German equivalent Stolz does not have the meaning of ‘self-esteem’ (?Stolz stärken) which is involved here. Interestingly, unabridged English-German dictionaries fail to record this sense, which may be translated into German as Selbstwertgefühl or Selbstbewusstsein, with typical verbal collocates such as stärken, steigern and aufbauen. The same goes for French, where développer l’estime de soi is a viable equivalent of the aforementioned collocations. The ‘self-esteem’ meaning of pride also impinges on other categories, such as causation:aspect:phasal:terminative (destroy sb’s pride = Selbstwusstsein zerstören). Other apparent collocational gaps may be attributed to differences in frequency and, perhaps, corpus choice. Thus, although pride + real/genuine appears to be significant in English only, literal French and German translations such as vraie fierté and echter Stolz are perfectly acceptable. The same is true of surge of pride, which commonly occurs with the verb feel. An example: Feldfunktion geändert Collocations across languages: unity in diversity? 21 (14) … when abroad I suddenly feel a surge of pride when someone says what a wonderful place England must be to live … The most natural way of rendering this item into German and French is by transposing the noun surge to the verbs aufsteigen and monter, respectively (ich spüre Stolz in mir aufsteigen / wie Stolz in mir aufsteigt; je ressens la fierté monter en moi); here too the translations have the ring of authenticity but are impossible to locate in any of our corpora. 5 Discussion On the basis of the sample examined here, it can be seen that the correspondences between English, French and German emotion words and collocations are more precise than might have been expected. Differences in the weighting of particular semantic dimensions are subtle and may be due to corpus bias. What may appear to be lexical or collocational gaps at first blush can often be filled by resorting to shifts in class or rank; in other words, the overwhelming majority of collocational meanings (or ‘meaning units’) can be translated without any loss of content. The further one moves into traditional phraseological territory (as in the case of ‘se draper dans sa fierté’), the more likely it is that full equivalence cannot be achieved. Further work is required to explore the possibility of defining a threshold value for each semantic dimension which, if exceeded, would indicate the presence of polysemy or the need for oblique translations. In practical terms, this might help the lexicographer decide when and where to establish additional sense divisions. There is also some evidence that in areas where two languages are in close contact, as in Quebec, the collocational profiles are even more congruent; thus, the direct equivalent of inevitable disappointment ‘déception inévitable’ is common in Canadian French, but not significant in the French of France. It would be interesting to carry out interviews to ascertain whether bilingual journalists or highly proficient translators have an intuitive knowledge of such differences but this seems somewhat unlikely. Similarly, we find what one might term ‘intertemporal’ equivalents, i.e. collocations which have passed out of use in one language but still have a wide currency in another; one example of this was modest/humble pride, a usage widely current in the nineteenth century whose German and French equivalents have continued to be used until the present day. The absence of wide differences between the three languages under investigation is not an unexpected finding, given the close proximity of the linguacultures under survey. The findings obtained are consistent with previous research in psy- 22 Dirk Siepmann chology (e.g. Scherer et al. 1986) suggesting that in European countries the commonalities in emotional experience are much larger than the differences (note, however, that the methodology of such studies has been called into question; see Goddard 2007). This is not to say, however, that such differences do not exist between more distant languages. Thus, a study of Malayan and English emotions nouns roughly denoting ‘surprise’ revealed considerable divergences (Goddard 2007). By the same token, experience in translation and lexicography suggests that other vocabulary areas may be characterized by more substantial interlingual difference (e.g. time expressions). More broadly, it is worth remembering that the translatability of collocations does not mean that these are in any sense fully identical across languages. There is no symmetrical and reversible relationship of identity between items in different languages (with few obvious exceptions such as terms [atom-Atom] or place names [Munich-München]). The existence of such relationships would presuppose that speakers of the languages in question draw on the same body of discourses in which literal translations of identical items are ‘primed’ (in the sense of Hoey 2005) in identical ways. Nothing could be further from the reality, of course; Meillet and many Saussurians were right in stating that “une langue est un système où tout se tient”, that is, language is an integrated relational system where each item is defined by the (intralingual) relations it entertains with other items; hence the impossibility, or at least, improbability of two such relational systems being identical. Yet there are clearly relationships of similarity between items in different languages, and an argument can reasonably be made that such similarity increases with the length of the item. The phenomenon of synonymy, by virtue of being the intralingual counterpart of equivalence, may help us understand this idea. At the word level, synonymy is probably inexistent. Even apparent synonyms like German Orange and Apfelsine may differ in terms of particular features such as currency of use (Apfelsine is now on its way out). As this present study has shown, however, synonymy becomes more likely at the collocational level. Unverhohlener Stolz and unverkennbarer Stolz or piétiner + fierté and bafouer + fierté are (almost) perfect synonyms. Once we progress to the sentence level, we may obtain perfectly synonymous utterances based on different items, a feature of language dictionary makers are well acquainted with. Here are two examples, from a dictionary currently under compilation (Siepmann, i.V.): (15) A few weeks ago Koch-Mehrin was stripped of her doctoral title. <-> A few weeks ago Koch-Mehrin had her doctoral title revoked. (16) He spent three months on work placement at Siemens. (BE) <-> He did a three-month internship at Siemens. (AE) Collocations across languages: unity in diversity? 23 Taking this to its logical conclusion, we would expect the highest degree of synonymy to be found at the text level, but this assumption is almost impossible to handle empirically. What should be evident, however, is the close analogy with “equivalence in difference” (Jakobson 1959/1966: 233), i.e. interlingual equivalence. While this cannot be achieved at the word level, the evidence presented in this article and the argument from synonymy just made would appear to suggest that interlingual relationships of similarity obtain between extended items such as collocations. The implications of this study for translation and lexicography are such that they call for a reappraisal of widely held beliefs and procedures. One such belief concerns the use of primary interlingual associations, as reflected in the following statement by two translation scholars: Sind eventuelle Unklarheiten geklärt, so gelangt der erfahrene Übersetzer durch interlinguale Primärassoziation schnell zu dem englischen Substantiv anger als Übersetzung des deutschen Begriffs Wut. Um jedoch sicherzugehen, dass […] übereinstimmen, bietet es sich ggf. an, die Definitionen in einsprachigen Qualitätswörterbüchern zu vergleichen […] Nach Überprüfung der Annahme, dass anger in der konkreten Übersetzungssituation die bestmögliche Wahl auf der paradigmatischen Ebene ist, muss der Übersetzer nun eine Wahl auf der syntagmatischen Ebene treffen. Dieser Selektionsprozess erweist sich jedoch als weitaus komplexer als der erste. (Kimmes/Kornelius 2010: 9) This view of translation overlooks the close interdependence between the paradigmatic and syntagmatic axes of language patterning (see Section 2 above). In the case of a (partial) collocational gap like they avenged last year’s disappointment, neither an interlingual primary association nor a dictionary look-up would prevent the translator from committing a serious error. Just like the theoretical linguist, the translator has to work “on the assumption that meaning is created on both axes” (Sinclair 2004: 170); in other words, the two axes need to be kept in view for as long as possible in the selection process. This process involves three steps: Phase 1: syntagmatic analysis (collocational profiling) of several potential equivalents Phase 2: meaning analysis of potential equivalents; choice of closest equivalent (if there is no direct equivalent, search for equivalents at the phrase level, redistribution of information units, etc.) Phase 3: choice of the best possible collocations and colligations A second implication concerns the challenge to established ways of thinking about translation procedures, especially modulation, transposition and amplification (French ‘étoffement’). While these have traditionally been conceptualized as being governed by the translator’s linguistic intuition, the present study has shown that 24 Dirk Siepmann collocational gaps may, as a rule, be filled by means of modulation, and that transposition is the method of choice when there is a wide difference in frequency between source-language and target-language items (e.g. amour fraternel – Bruderliebe (rather than brüderliche Liebe) or freudiger Stolz – pride and joy – joie et fierté (rather than joyous pride or fierté joyeuse). With a view to translator training, it must be stressed, however, that most of the time systematic searches will enable the translator to spot direct collocational equivalents between languages. Other, mainly lexicographic implications follow from the existence of stereotypical expressions based on collocations, such as ‘book early to avoid disappointment’. It should be noted that such word combinations as avoid disappointment or grand débat (cf. Blumenthal 2008) would not even have been considered as collocations under traditional phraseological approaches and would thus have been discarded by the lexicographer; their significance and specific meanings only become apparent when the wider context is fully taken into account. Here one implication is the need for collocational and general-language dictionaries to extend their coverage beyond two-word collocations and traditional phraseology (cf. also Blumenthal 2008); another is the importance of incorporating sufficient contextual and stylistic information into collocational dictionaries. Thus, a collocation such as express + disappointment is typical of journalistic usage but rare in speech. From a language teaching perspective, it is important to determine which collocates allow for the widest possible range of combinations with emotion nouns and are thus fairly ‘safe to use’ for L2 learners. Among the potential verbal candidates for English are express, show and feel; the only adjectival collocate that can be combined with a large number of nouns appears to be great. These four words are probably those that should be taught in the beginning stages (i.e. the A1/A2 levels of the Common European Framework), with the caveat that, as just noted, attention needs to be drawn to the preferred use in the written registers of noun + verb collocations like express disappointment. In everyday speech the preference is for verbal (e.g. for French: je suis déçu) and adjectival expressions (X est décevant) as well as exclamations (Dommage!, Dommage que + clause, c’est pas vrai, quand je pense que …; such expressions are listed under the heading ‘déception’ for the B2 level of the CEFR in Beacco et al. 2004). Students who wish to progress beyond the A2 level may be sensitized to using collocational dictionaries to express finer shades of emotional meaning in their writing. Collocations across languages: unity in diversity? 25 References (a) Dictionaries DUDEN = DUDEN DEUTSCHES UNIVERSALWÖRTERBUCH. 4., neu bearb. u. erw. Aufl. Mannheim u. a.: Dudenverlag 2001. ALD = OXFORD ADVANCED LEARNER’S DICTIONARY OF CURRENT ENGLISH. Hg. Jonathan Crowther. Oxford: Oxford University Press 82008. [11948 A.S. Hornby]. PONS = PONS GROßWÖRTERBUCH FRANZÖSISCH-DEUTSCH. Hgg. Jean Chassard, Marianne Frischmann-Chautard. Stuttgart: Klett 1996. (b) Other references Anscombre, Jean-Claude (1995): “Morphologie et représentation événementielle: le cas des noms de sentiment et d'attitude.” - Langue Française 105, 40-54. Anscombre, Jean-Claude (1996): “Noms de sentiment, noms d'attitude et noms d'abstraits.” In: N. Flaux, M. Glatigny et D. Samain (eds.) : Les noms abstraits, Actes du Colloque de Dunkerque (1992), 257-273. Paris: Presses universitaires du Septentrion. Balibar Mrabti, Antoinette (1995): “Une étude de la combinatoire des noms de sentiment dans une grammaire locale.” – Langue Française 105, 88-97. Beacco, Jean-Claude (2004): Niveau B2 pour le français – un référentiel. – Paris: Didier. Blank, Andreas (2001): Einführung in die lexikalische Semantik für Romanisten. – Tübingen: Niemeyer. Blumenthal, Peter (2006): Wortprofil im Französischen. - Tübingen: Max Niemeyer (= Beihefte zur Zeitschrift für Romanische Philologie 332). Blumenthal, Peter (2008): “Französische Kollokationen in Lexikografie und Forschung.” – Lexicographica 24, 21-38. Cruse, Alan D. (1986): Lexical Semantics. – Cambridge: Cambridge University Press. Dubois, Jean (1969): “Énoncé et énonciation.” – Langages 13, 100-110. Firth, John R. (1957): “Modes of Meaning.” – In: J. R. Firth (ed.), Papers in Linguistics 1934-1951, 190-215. London et al.: Oxford University Press. Firth, John R. (1968): “A Synopsis of Linguistic Theory 1930-1955.” - In: F. R. Palmer (ed.), Selected Papers of J. R. Firth 1952-1959, 168-205. London et al.: Longmans (Longmans’ Linguistics Library). Goddard, Cliff (2007): “A culture-neutral metalanguage for mental state concepts” - In: A. C. Schalley, K. Drew (eds.): Mental States. Volume 2: Language and cognitive structure, 11-34. Amsterdam: John Benjamins. Hausmann, Franz Josef (1999): “Le dictionnaire de collocations - Critères de son organisation” – In: N. Greiner; Joachim Kornelius; Giovanni Rovere (eds.): Texte und Kontexte in Sprachen und Kulturen. Festschrift für Jörn Albrecht, 121-139. Trier : WVT Wissenschaftlicher Verlag. Heringer, Hans-Jürgen (1999): Das Höchste der Gefühle: Empirische Studien zur distributiven Semantik. – Tübingen: Stauffenburg (= Stauffenburg Linguistik). Hesselhauf, Nadja (2003): “The use of collocations by advanced learners of English and some implications for teaching” Applied Linguistics 24/2, 223-242. Hoey, Michael (2005): Lexical Priming: A New Theory of Words and Language. – London, New York: Routledge. Formatiert: Englisch (USA) 26 Dirk Siepmann Jakobson, Roman (1959/1966): “On linguistic aspects of translation” In: R. A. Brower (ed.): On translation, 232-239. New York: Oxford University Press. Jones, Steven (2002): Antonymy: a corpus-based perspective. – London: Routledge. Katz, Jerrold I./Fodor, Jerry A. (1963): “The Structure of a Semantic Theory” - Language 39, 170210. Keller, Rudi (1995): Zeichentheorie: Zu einer Theorie semiotischen Wissens – Tübingen, Basel: Francke. Kimmes, Anne/Kornelius, Joachim (2010): “Oftmals trügt der Schein. Bemessungen des collocational range und ihre Bedeutung für die Übersetzungspraxis”. T21N – Translation in Transition 12. (http://www.t21n.com/homepage/articles/T21N-2010-12-Kimmes, Kornelius.pdf) (19.03.2014) Kövecses, Zoltán (1986): Metaphors of Anger, Pride, and Love: A Lexical Approach to the Structure of Concepts. – Amsterdam: John Benjamins. Leeman, Danielle (1995): “Pourquoi peut-on dire Max est en colère mais non *Max est en peur ? Hypothèses sur la construction être en N.” - Langue française 105, 55-69. Leisi, Ernst (1953): Der Wortinhalt – Seine Struktur im Deutschen und Englischen. – Heidelberg: Carl Winter. Mathieu, Yvette Yannick (1995):” Verbes psychologiques et interprétation sémantique” – Langue française 105, 98-116. Mathieu, Yvette Yannick (2000): Les verbes de sentiment: de l’analyse au traitement automatique. – Paris: Editions du CNRS. Mel’čuk, Igor A./ Clas, André/Polguère, Alain (1995): Introduction à la lexicologie explicative et combinatoire. – Louvain-la-Neuve: Duculot (Champs linguistiques: Manuels). Meyer, Charles F. (2010): Introducing English linguistics. – Cambridge: Cambridge University Press. Plantin, Christian/Doury, Marianne/Traverso, Véronique (eds.) (2000): Les émotions dans les interactions. – Lyon: PUL. Pottier, Bernard (1963): Recherches sur l’analyse sémantique en linguistique et en traduction mécanique. – Nancy: Faculté des Lettres et Sciences Humaines. Proost, Kristel (2001): “Zum Lexikalisierungsraum kommunikativer Konzepte.“ – In: G. Harras (eds.): Kommunikationsverben: Konzeptuelle Ordnung und semantische Repräsentation, 77129. Tübingen: Narr, 2001. Putnam, Hilary (1988): Representation and Reality. – Cambridge, MA: The MIT Press. Ruwet, Nicolas (1994): “Etre ou ne pas être un verbe de sentiment” – Langue française 103, 4555. Ruwet, Nicolas (1995): “Les verbes de sentiments peuvent-ils être agentifs?” – Langue française 105, 17-27. Scherer, Klaus R. et al. (eds.) (1986): Experiencing Emotion: A Cross-Cultural Study. – Cambridge University Press. Siepmann, Dirk (2003): “Collocations in Trilingual Perspective: New Evidence from Large Corpora and Implications for Dictionary Making.” – Cahiers de Lexicologie 83/2, 173-196. Siepmann, Dirk (2005): "Collocation, Colligation and Encoding Dictionaries. Part I: Lexicological Aspects" –International Journal of Lexicography 18/4, 409-444. Sinclair, John M. (1991): Corpus, Concordance, Collocation, – Oxford et al.: Oxford University Press (= Describing English Language). Sinclair, John M. (2004): Trust the Text: Language, Corpus and Discourse, Ronald Carter (Ed.): London et al.: Routledge. Strohmeyer, Fritz (1910): Der Stil der französischen Sprache. – Berlin: Weidmann. Feldfunktion geändert Formatiert: Englisch (USA) Collocations across languages: unity in diversity? 27 Tutin, Agnès, Iva Novakova, Francis Grossmann & Cristelle Cavalla (2006): "Esquisse de typologie des noms d’affect à partir de leurs propriétés combinatoires" – Langue française, 150, 32-49. Wierzbicka, Anna (1996): Semantics: Primes and Universals, – Oxford et al.: Oxford University Press.