Collocation Dictionary Creation
1) A. P. Cowie
Cowie takes this a step further, and categorises phraseological units (collocation dictionary) not only on the basis of their semantic transparency, but also according to their grade of lexical and grammatical variability in the collocation dictionary (cf. Nesselhauf 2004: 10).
First of all, Cowie (1994) distinguishes ‘composites’ from ‘formulae’ – a categorisation which goes hand in hand with the Russian tradition – the former being word combinations below the sentence level. According to semantic transparency and (lexical and grammatical, for collocation dictionary) variability composites are further classified into the following categories which should be seen on a scale from (1) to (4) in collocation dictionary:
a) pure idioms – they are completely opaque and invariable, e.g. red tape
b) figurative idioms – they are slightly variable and have one figurative and oneliteral interpretation, e.g. green fingers
c) restricted collocations – they have one figurative and one literal element and there is an arbitrary combinability restriction on one of the elements with other elements outside the combination, e.g. dry cow
d) open collocations – they consist of elements which are both used in their literal sense and they are freely combinable, e.g. thunderous applause.
You can find all these in collocation dictionary.
(2) I.A. Mel’cuk
Following the Russian classification model, Mel’cuk also distinguishes between sentence- and word-like units.
He refers to word-like units as semantic phrasemes (as opposed to pragmatic phrasemes) and subdivides them as follows (cf. Nesselhauf 2004: for collocation dictionary Skandera 2004:
a) idioms – their meaning as a whole is not included in any of the constituents, e.g. red tape
b) quasi-idioms / quasi-phrasemes – they have a literal meaning and implicitly
express an additional one, e.g. bacon and eggs – where the meaning of ‘fried’ is part of either element (Nesselhauf 2004: 12)
c) collocations / semi-phrasemes
Mel’cuk further classifies collocation in four major categories, which Nesselhauf (2004: summarises in a less formalised language as (1) collocations including a delexical verb, e.g. take a step, (2) collocations (collocation dictionary) in which the meaning of the dependent lexeme is expressed in combination with only a few others, e.g. black (‘without milk’) coffee, (3) collocations whose dependent lexeme can be used in the same sense in combination with a lot of other lexemes, e.g. strong coffee, and (4) collocations in which the dependent lexeme includes the meaning of the other lexeme, e.g. the meaning ‘horse’ is included in neigh. These are included in collocation dictionary as well.
What Mel’cuk is best known for is his model of ‘lexical functions’ (LFs), which is part of his ‘Meaning-Text-Theory’ and comprises well over 50 different types in his collocation dictionary. He sees collocations as an oriented relationship where the so-called ‘keyword’ selects the dependent lexeme, the ‘value’, i.e. in a combination like heavy smoker, smoker is the keyword and selects the value heavy through the lexical function of Magn (‘magnitude’ = intensifier in collocation dictionary).
The Lexicographical Approach to Collocations
The lexicographical approach deals with the question of how collocations should be presented in collocation dictionary. The prime practitioners are Morton Benson and Franz-Josef Hausmann, although also A.P. Cowie, I.A. Mel’cuk and H.E. Palmer must be mentioned here.
Cowie implemented his phraseological approach in the two volumes of the Oxford Collocation Dictionary of Current Idiomatic English (1975 and 1983), Mel’cuk compiled the Dictionnaire explicatif et combinatoire du français contemporain: recherches lexicosémantiques (1984-99, collocation dictionary in French), so far in four volumes and, together with A.K. Zholkovsky, the collocation dictionary creator.
Explanatory Combinatorial Collocation Dictionary of Modern Russian (1984) using his model.
This conception of the term collocation dictionary can be compared to W. Porzig’s “Wesenhafte Bedeutungsbeziehungen” (1934) – this is basically an approach to meaning concerned with syntagmatic relations, e.g. wiehern – Pferd (viz. neigh – horse) where the semantic relevance of the syntactic relation is obvious since what can be connected with wiehern is implicit in it:
Anecdotally, Mel’cuk invented his concept of “lexical functions to label systematic and re-current lexical-semantic relations” collocation dictionary, during a storm in the Russian countryside, when reflecting on the English language: the fact that “rain combines with heavy while light combines with bright led him to invent the first lexical function which he called Magn (from Magnitude) to refer to an intensifying meaning.” (cf. Evens 1998: 11, as quoted in Fontenelle 1997: collocation dictionary)
5) and Palmer took his work on the core vocabulary of English as the basis for his collocation dictionary A Grammar of English Words (1938).
M. Benson, being an American Slavonic linguist, is also influenced by the Russian tradition in way that he defines collocations with respect to other word combinations in collocation dictionary. His approach to the presentation of collocations in a collocation dictionary mainly had its basis in practical reasons for the purpose of compiling, together with E. Benson and R. Ilson, the BBI9.
As summarised in collocation dictionary Nesselhauf (2004: 14-15), he divides lexical combinations (word-like units in the Russian tradition) into the following categories:
a) idioms – they are defined as (Benson et al. 1986a: 252)
b) transitional combinations – are more transparent than idioms but less variable than collocations
c) collocations
d) compounds
Collocations, for Benson, are “recurrent, semi-fixed combinations” and they can be further subdivided into grammatical and lexical collocations (Benson et al. 1997: ix; cf. collocation dictionary; e.z.)
section 1.1. of the collocation dictionary:
The four volumes cover only a selection of some 500 entries scattered throughout the alphabet, of varying complexity.
Interestingly, Benson recognised the need for a collocation dictionary for English learners when he compiled his Serbo-Croatian – English collocation dictionary in the mid-to-late 1960s (Gabrovsek 2004: personal communication).
(2) F.J. Hausmann
Hausmann, first of all, distinguishes collocation dictionary fixed from non-fixed combinations. Fixed combinations, for him, are idioms, compounds, etc., while non-fixed combinations can be further classified as (Hausmann 1984, as quoted in Nesselhauf 2004: 16 and Bahns 1996: 23; collocation dictionary II):
a) co-creations – are free combinations that are creatively combined by the speaker,
e.g. une maison agréable (viz. a pleasant home)
b) collocations – are not creatively combined but put together out of some convention, e.g. ton péremptoire (viz. peremptory tone)
c) counter-creations – are words that do not usually combine, they are mainly found in literature and advertisements to create a special effect, e.g. jour fissuré (viz. cracking day).
A collocation, according to Hausmann, can be defined as: “la combinaison caractéristique de deux mots dans une des structures suivantes (for French collocation dictionary):
a)substantif + adjective (épithète) b) substantif + verbe c) verbe + substantif (objet) d) verbe + adverbe e) adjective + adverbe f) substantif + (prép.) + substantif.”
(F.J. Hausmann 1989: 1010)
According to Nesselhauf (2004: 16-17), Hausmann’s most important contribution to.
Collocation dictionary theory is his view what the status of the two elements in the collocation is not the same. While one of the elements, the ‘base’, is semantically autonomous, the other one, the ‘collocator’ is chosen according to this base and only get its precise meaning from that combination10 (cf. Hausmann 1989: 1010).
Lexicographically speaking, he put forward that collocations should be listed under their bases in collocation dictionary, rather than their collocators, since “la fameuse recherché du mot propre est celle du collocatif.” (Hausmann 1989: 1010). He is of the opinion that listing the collocation in the collocator entry merely serves the testing of a hypothesis.
This phenomenon is referred to as ‘semantic tailoring’ by DJ. Allerton (1982 and 1984). this?”), but listing the collocation in the base entry allows the dictionary user to hit upon an unknown collocation or a collocation dictionary they might have forgotten (ibid.).
Rule 2.2.5 of collocation dictionary – The Linguistic (Frequency-Based) Approach to Collocations
In contrast to the Russian, the phraseological and the lexicographical approaches, the linguistic approach does not characterise collocations with respect to other word combinations on a certain cline. Collocations in the linguistic app. (collocation dictionary I.e).
Heliel (1990: 134-35) points out that the collocator is the unpredictable element in a Collocation dictionary and that the dictionary user, in encoding, usually knows the equivalent ofthe base (or they can easily be found in the bilingual dictionary). He or she is mainly concerned with finding the collocates and thus needs the onomasiological approach.
In contrast to Cop, Heliel (ibid.) sees the semasiological approach relevant only for decoding purposes since in comprehending a foreign-language collocation dictionary it is primarily the collocator that causes difficulties – which is, again, due to semantic tailoring.
Hausmann (1988: 149; cf. Bahns 1996: 47) notes that, basically, it makes no difference to the translator where he or she will find the collocation, under the base or under the collocator (applicable for any collocation dictionary). Still, when the collocation is only listed in one entry, this often results in time-consuming double look-ups. Hausmann thus arrives at the conclusion that a collocation should ideally be found in all four entries: base and collocator of the source language and base and collocator of the target language. However, limited space inprinted collocation dictionary makes this an impossible option since “general bilingual collocation dictionary has a much larger job than just presenting collocations”
The space problem can be easily overcome by the implementation of the dictionary on a CD-ROM. Furthermore, I am of the opinion that electronic versions of bilingual dictionaries should have a special collocation feature, such as the one in the Longman Collocation Dictionary of Contemporary English on CD-ROM.
Looking at the four major German-English / English-German collocation dictionary on the market (cf. sections 1.3. and 3.5.1.) we can observe that, first of all, on a purely quantitative basis, in the best of all cases in a German-English encoding translation task of adjective-noun collocations, six German collocations out of 38 are listed under their noun bases on the German side. 16 collocations were found in the adjective, the collocator, entry though. For the encoding translation of 29 English-German collocations only one collocation was listed explicitly in the noun entry on the English side of the bilingual collocation dictionary in question, while the best collocation dictionary managed 15 collocations given in the adjectives entries.
Thus, we might assume that in German and English collocation dictionary, lexicographical practice the semasiological approach to the representation of collocations prevails; however, if we look at the different user’s guides, we will learn otherwise.
Collocators are nouns which the translations typically qualify. They are normally placed after the translation.” (Thyen et al. eds. 1999: 25). In the German-English part, collocators are in English; in the English-German part they are in German.
We see that DUOX treats adjectives as bases and nouns as collocators. This could lead to the incorrect assumption that in DUOX collocations are listed under their bases – a vital prerequisite for an encoding dictionary as suggested by Hausmann (1988), Heliel (1990) and Cop (1991). Without wishing to belittle Thyen et al.’s work, I would prefer them to have applied standard terminology – since, in adjective-noun collocations, the noun IS the base specifying the meaning (and translation) of the pre-modifying adjective in the first place.
Likewise, in the user guide to COGER, Terrell et al. (eds. 1999: xiv-xv) point out, under the heading Explanatory Material, that collocators and typical complements are, in contrast to sense indicators, not in brackets but in italics. In adjective entries the user will find typical nouns that the adjective modifies. They are usually found before the translation. Again, adjectives are treated as bases and nouns as collocators. Next part of collocation dictionary is coming soon!
Get back to the homepage of Collocation Dictionary!
collocation dictionary