Читать книгу Collocations, Creativity and Constructions - Cordula Glass - Страница 10

After over 50 years of debate, the famous words “You shall know a word by the company it keeps!” (Firth 1957/1968: 179) still seem to sum up the quintessential character of any definition of collocation quite well. But Firth was not the first to acknowledge the importance of context and lexical co-occurrence for the semantic interpretation of a word. In lexicographic description, authentic context and combinatorial restrictions were already a concern for lexicographers like Samuel Johnson1 (1747/1837) and later Harold Palmer (1933). Yet, their focus was an accurate and, in the latter’s case, learner-appropriate account of the English language (> 2.2), while Firth shifted the perception of collocation from an observable phenomenon to a linguistic principle, which he considered one of the basic relations within a linguistic system, alongside grammar:

Collocations of a given word are statements of the habitual or customary places of that word in collocational order but not in any other contextual order and emphatically not in any grammatical order. The collocation of a word or a ‘piece’ is not to be regarded as mere juxtaposition, it is an order of mutual expectancy. (Firth 1957/1968:181)

This definition of collocation is deeply rooted in Firth’s firm belief in the importance of context, which he considers the main source of any kind of meaning. Unlike for example Odgen and Richards (¹⁰1956: 10–11) and modern cognitive linguists, he claims that there is no such thing as a “hidden mental process, but chiefly […] situational relations in a context of situation […]” (Firth 1951/1964: 19). Hence, to Firth, collocation is a lexical phenomenon which operates on a syntagmatic level to create meaning through (inter-)relation and “mutual expectancy” of words. Compared to most modern definitions of collocation, this seems to be a very restricted view, but at a time when words and their meaning were nothing more than a peripheral phenomenon within linguistic research, a statement like this was a first attempt to shift the focus from a grammar-centred slot-filler perspective to a conception of language which needs both lexicon and syntax to form meaningful linguistic speech. One could even argue that Firth’s acknowledgement of meaning-building syntagmatic relations in language did in fact lay the ground for modern cognitive linguistics (> 4), since it is based on the idea that meaning arises through context, which is ultimately nothing but usage and therefore in its basic assumption quite similar to modern usage-based approaches (Bybee 2010; Goldberg 2006; Tomasello 2005). Unfortunately, Firth never gave a more detailed account of what he considers to be sufficient context for “meaning by collocation”, but it becomes apparent from his publications and a few analyses (Firth 1951/1964) that this does not simply refer to some kind of compounding:

Meaning by collocation is an abstraction at the syntagmatic level and is not directly concerned with the conceptual or idea approach to the meaning of words. One of the meanings of night is its collocability with dark, and, of dark, of course, collocation with night. (Firth 1951/1964: 196)

The example of dark night, however, seems to be somewhat misfortunate. First, because it suggests a certain spatial closeness of collocates, which, as the next example will show, is not a necessary prerequisite for the concept of “context” and “meaning by collocation”. Secondly, in this case, it is difficult to argue for “meaning by collocation”, because the co-occurrence of dark and night could also refer to real-life, extra-linguistic experience, since darkness is one of the predominant semantic components of night, even without its co-occurrence with dark (Herbst 1996: 384). Therefore, another of Firth's frequently quoted examples might be more suitable to explain the basic assumptions of this early contextualistic approach: you silly ass. Here, Firth claims that the additional meaning which arises through collocation is some kind of “personal reference” (Firth: 1951/1964: 194–195). So, when encountering a sentence like (4) one would directly conclude that the referent of ass is a person and not a donkey, because of past experiences with similar cases, such as (5) to (8)2:

(4)	BNCA0D 1316	I'm on in five minutes, that old ass slowed me down.
(5)		An ass like Bagson might easily do that.
(6)		He is an ass.
(7)		You silly ass.
(8)		Don’t be an ass.

Take another example from this chapter’s introductory quote: scornful tone. In this case, one can conclude that what the egg-shaped creature said was not very nice or polite even without necessarily being familiar with an adjective like scornful, since tone is often encountered with negatively marked premodifications such as dismissive or harsh3.

However, these examples show that, even though Firth explicitly stresses the syntagmatic character of collocations, his conception nevertheless includes an inherent paradigmatic level through the notion of “mutual expectancy”, since any expectation is based on previous experience. Therefore, in order to know the company a word keeps, one needs to have encountered this very word in various contexts and processed it with the help of some kind of cognitive storage. Yet, whether Firth would agree with this observation or not is difficult to tell, since his description of collocations remains in a rather vague state throughout his publications (Firth 1951/1964; 1968) and Lyons (1977: 612) is certainly right when he observes: “Exactly what Firth meant by collocability is never made clear.”

Despite all criticism, Firth’s thoughts on collocation and its influence on the meaning of a lexical unit did inspire his students, John Sinclair and Michael Halliday. Based on Firth’s first sketch of these special phenomena of partly restricted co-occurrence, two approaches towards collocation have emerged: a frequency-oriented approach with corpus linguists like John Sinclair and Göran Kjelmer at its centre and a text-oriented approach, which was primarily developed by Halliday. While Sinclair’s definition of collocations as an observable phenomenon of statistical significance within a linguistic system is still the basis for collocational research today (among others Hanks 2013; Moon 2009; Bartsch 2004; Stefanowitsch/Gries 2003; Hunston/Francis 2000), Halliday’s concept of collocation as a meaning-creating phenomenon is frequently rejected as a misnomer for a special form of cohesive tie within textlinguistics (Herbst 1996: 381). However, especially in his early publications, it was Halliday who explicitly added a paradigmatic level to a contextualist perception of lexical co-occurrences. He was, therefore, the first to acknowledge an experience or usage-based dimension within lexis.

Like his approach to grammar, Halliday’s general conception of collocation is strongly influenced by Firth’s notion of the importance of context. Yet, employing a more systematic definition of basic principles within language, he tries to base the relations of lexical as well as syntactical linguistic elements on conceptually more solid ground. He introduces a two-dimensional model to illustrate the most prominent interdependencies of language (Halliday 1966: 152–153).

Table 2.1: Lexical and grammatical relations (based on Halliday 1966: 152–153)

Following de Saussure (1916/1967: 147–148) he argues for a lexico-grammatical as well as a syntagmatic-paradigmatic level of analysis and positions collocation at the intersection of syntagmatic, lexical phenomena (table 2.1). By definition, a paradigmatic level is excluded here. This would make collocation a rather broad category for all things within one string of words (= chain), but not every linear co-occurrence would count as a collocation in Halliday’s books. Again, the notion of “mutual expectancy” or “significant proximity” plays a crucial role:

First, in place of the highly abstract relation of structure, in which the value of an element depends on complex factors in no sense reducible to simple sequence, lexis seems to require the recognition merely of linear co-occurrence together with some measure of significant proximity, either a scale or at least a cut-off point. It is this syntagmatic relation which is referred to as ‘collocation’. (Halliday 1966: 152)

Coming back to the example of Humpty Dumpty, what Halliday would presumably regard as collocations are expressions like neither more nor less, do work or make a remark, not so much because of their reoccurring character throughout a text or discourse, as would be the case for Firth’s concept of “meaning by collocation”, but rather because of their restricted nature. He therefore stresses that “[i]n lexical analysis it is the lexical restriction which is under focus: the extent to which an item is specified by its collocational environment.” (Halliday 1966: 156). Neither more nor less, for example, could not, or at most fairly rarely, be expressed as neither less nor more4, do work is much more usual than make work and vice versa, and one makes a remark instead of does a remark. The adjective scornful in a combination like scornful tone, on the other hand, might be mutually expected in terms of meaning or, to be more precise, the likelihood of a negatively marked adjective to occur with the nominal use of tone, but it is not an item which shows “significant proximity” with the noun, since the same concept could be equally well expressed with condescending or sarcastic5. To analyse likely candidates for lexical co-occurrence, Halliday establishes the concept of lexical sets, which is of course closely related to collocation, since its members are selected based on “the similarity of their collocational restriction” (Halliday 1966: 156) Later he continues: “If we say that the criterion for the assignment of items to sets is collocational, this means to say that items showing a certain degree of likeness in their collocational patterning are assigned to the same set.” (Halliday 1966: 158). The value of this distinction between paradigmatic and syntagmatic level is that it clearly shows two principles which apply to any word co-occurrence with an intuitively – at least for native speakers of English – close relationship: the fact that the constituents cannot be substituted by any random word with roughly the same semantic concept (collocation) and a certain underlying meaning value which words with similar collocation share (lexical set). This is strikingly similar to the principles of cognitive linguistics in general and construction grammar in particular, since more abstract constructions, like argument structure constructions (Goldberg 2006: 19–44), are at their heart very much like Halliday’s collocations and lexical sets. They open up slots on a paradigmatic level, which influence word choice based on certain meaning restrictions, not unlike collocations, so ultimately this leads directly to a notion of collocations as constructions (> 2.3; 4). Admittedly, in later publications like his 1976 co-authored book on Cohesion in English (Halliday/Hasan 1976), Halliday focuses more strongly on Firth’s original concept of “meaning by collocation” and draws away from his general thoughts on the structure of lexical relations within language.

The Collins Birmingham University International Language Database (COBUILD), the result of cooperation between the University of Birmingham and Collins Publishers, started in 1980. The aim of this project was to build a large corpus of contemporary text (100 million words) in order to analyse the lexical and grammatical foundations of the English language on a more systematic basis. The pilot project, the OSTI Report (Sinclair/Jones/Daley 1970/2005), had already yielded some interesting insights into the pervasive nature of collocational phenomena, so with this follow-up project Sinclair and colleagues concerned themselves once more with a detailed, corpus-driven6 analysis of lexical items. The result was, as intended, a comprehensive corpus-based dictionary of the English language, the Collins COBUILD English Language Dictionary (COBUILD 1, Sinclair 1987a), but the OSTI report and the conception of the COBUILD 1 also influenced research on collocations in general. Until then, analysis of collocations was based on a rather random sampling of examples and seemed to be somewhat eclectic at times. So, following Firth’s thoughts on the co-occurrence of words, Sinclair took further steps towards the operationalization of the concept of collocation. In order to do so, he labelled its components node, for the lexical item under investigation, and collocate, for “items in the environment” of the node which co-occur in a text or sentence within a certain distance, the span7 (Sinclair 1966: 415). Leaving introspection behind, all these elements can then be retrieved and correlated with the help of a corpus, which could be regarded as a kind of replica of adult human linguistic experience. In principle, this leads to a very simplified basic assumption: the more frequent, the more important. Unsurprisingly, the notion of frequency is also at the centre of Sinclair’s concept of collocation:

Collocation is the cooccurrence of two items in a text within a specified environment. Significant collocation is regular collocation between items such that they co-occur more often than their respective frequencies and the length of text in which they appear would predict. Standard statistical tests can be used to tell whether the association between word A and word B is a significant one. (Jones/Sinclair 1974: 19)

Once again, true to contextualist tradition, the term is defined against the background of co-occurrence and context (“specified environment”). A new aspect is the additional dimension of observable frequency and with it the distinction between significant and casual collocations at its centre (Sinlcair 1966: 418). While basically any co-occurrence of lexical items could be considered a casual collocation, significant collocations are regarded to be special since two items co-occur more often than expected by chance. This correlation, however, can only be established by employing, as Sinclair points out, statistical tests. However, the nature of these tests and their results are, still today, a source of multifaceted debate (Evert 2005: 137–164; Sinclair/Jones/Daley 1970/2005: xxi).

Hence, a point of criticism which eventually arose (Granger/Paquot 2008; Herbst 1996) is that corpus research, by definition, depends to a large extent on the mathematical methods applied. Since each statistical test yields different results and suggests different interpretations, a more or less subjective factor as far as the decision for or against a methodology is concerned, cannot be denied. This ultimately means that whether a combination of lexical items is indeed significant or not still depends, to a certain degree, on the choice and intention of the author of a study8. Similar objections could be raised with regard to the corpus as such. Not every corpus is suitable for any kind of research question. Corpus size and design affect the outcome of most statistical tests, and some corpora might simply be too small to answer certain kinds of question, like, for example, a comprehensive collocational set for less frequent words, such as scornful. Furthermore, corpus evidence and statistically significant collocations need to address the question of relevance, because a significant correlation might also occur because of the topic and/or genre of a text and not because a certain combination is particularly salient (> 4.2.1). The same might be true for collocations which describe real world phenomena, like build, buy and sell as the most frequent verbal collocates for house (Herbst 1996: 388–389). However, Sinclair and his colleagues claim that “[…] at this moment it is impossible to prove or disprove this because the examination has been confined to text of one particular kind.” (Sinclair/Jones/Daley 1970/2005: 76) and therefore suggest that “[a]ll collocations must, therefore, be accepted at their face value, since they have all actually occurred more than three times in the sample of spoken language and have passed a fairly stringent significance test.” (Sinclair/Jones/Daley 1970/2005: 76).

Table 2.2 lists just a small selection of statistical measures, namely Mutual Information (MI), z-score, t-score and log-likelihood, for Humpty Dumpty’s “collocations” to demonstrate their respective effect on a potential interpretation. The results are taken from the BNC and, based on Sinclair’s suggestion (Jones/Sinclair 1974: 21–22), the span was set at four, with a minimum of occurrence of three (Sinclair/Jones/Daley 1970/2005: 42).

Table 2.2: Corpus-based association measures for “Humpty Dumpty’s collocations”

What becomes apparent immediately is that the result for each of the items selected differs depending on the respective statistical measurement. In other words, in a ranking scornful tone, for example, scores quite high when it comes to MI but has a relatively low t-score and log-likelihood value. This is, of course, due to the fact that each method focuses on a different parameter and, therefore, yields different results. While Mutual Information and z-score highlight collocations with a rather low frequency but strong likelihood to co-occur, t-score and log-likelihood show high-frequency pairs. For MI and t-score a result of two or three is considered to be statistically significant (Oakes 1998: 11–12.; Hunston 2002: 69–75; McEnery/Xiao/Tono 2006: 56). Chapter 5 will provide a more thorough discussion of tests for statistical significance; the crucial point to make at this stage is that depending on the method chosen, scornful tone, do work and make a remark could be regarded as statistically significant collocations.

None of these methods, however, brought to light the most frequent verbatim word combination neither more nor less. Of course, one could argue in favour of neither more nor less as a word-like, lexically stored chunk. Yet, there are possible variations, like I neither knew nor cared or neither the TV nor the video (OALD 7: neither), which suggest an underlying constructional principle, like for example [neither X nor Y]. Scornful tone, on the other hand, might not be as relevant as the corpus analysis suggests, since, upon closer examination, three of the four hits are quotes of the very same passage, namely Humpty Dumpty’s and Alice’s conversation in Lewis Carroll Through the Looking-Glass. Despite the biased data for this co-occurrence, scornful tone could nevertheless qualify as a linguistically interesting word combination, or, to be more precise, the combination scornful and a noun phrase [scornful+N], since, according to the BNC, all nouns with the premodification scornful can be subsumed under the semantic headline of “visual and auditive senses” and tone, in this case, behaves right along the line. There seems to be a kind of semi-lexically filled collocation, like [scornful+N_{<visual and auditive senses>}] supporting the interpretation of a word combination like scornful tone. This phenomenon of underlying semantic properties of certain, partially fixed word combinations has also been observed by John Sinclair (1996, 1998, 2004) and Bill Louw (1993), who coined the term semantic prosody, which Gavioli later describes as “[…] the way in which words and expressions create an aura of meaning capable of affecting words around them.” (Gavioli 2005: 46; > 3.1).

Working with the COUBILD corpus, the Bank of English (BoE), Sinclair further realised that patterns like these are, in fact, themselves more pervasive in the English language than they would have been expected to be. He calls this the idiom principle or principle of idiom:

The principle of idiom is that a language user has available to him or her a large number of semi-preconstructed phrases that constitute single choices, even though they might appear to be analysable into segments. To some extent, this may reflect the recurrence of similar situations in human affairs; it may illustrate a natural tendency to economy of effort; or it may be motivated in part by the exigencies of real-time conversation. However it arises, it has been relegated to an inferior position in most current linguistics, because it does not fit the open-choice model. (Sinclair 1991: 110)

This observation stands in stark contrast to a traditional conception of language, which sees grammar and lexis as more or less detached from each other and lexical items as inferior to grammatical structures for which they simply serve as fillers for the slots they open up. In this respect, the notion of statistically significant collocations and the idiom principle can again be closely related to concepts within construction grammar, since both approaches take authentic language as a starting point and deduce more general linguistic principles from it. The purpose of the identification of these patterns and/or constructions in Sinclair’s research is, of course, very different from modern, cognitive-oriented corpus studies, which are mostly concerned with the mental representation of linguistic concepts and its consequences for language acquisition and learning (> 4). But the corpus-driven, inductive nature of corpus linguistics as a methodology makes it a valid tool for cognitive studies focusing on aspects like frequency of encounter, salience, pre-emption and entrenchment (Ambridge/Lieven 2011; Goldberg 2006; Stefanowitsch/Gries 2003). Furthermore, comparing potential collocates of a node word, looking at collocational sets, and identifying the most likely patterns of co-occurrence yields very similar results to what construction grammar has to offer in terms of underlying, partly delexicalised constructions within the English language system (> 5.1.2). Moreover, the functional side of semi-abstract constructions, like [scornful+N], is also very similar to Sinclair’s notion of semantic prosody.

To sum up the general assumptions which most context-oriented approaches share, it could be concluded that this categorisation implies that every dimension of meaning is assumed to be context-derived. Therefore collocates should be regarded as meaning-creating context (syntagmatic relations). Furthermore, previous relations to other lexical items also shape the two dimensions of collocational meaning (paradigmatic relations). Here, syntagmatic relations refer to directly encountered co-occurrences, which do not necessarily need to be special or noteworthy in any way. It is rather the level of previously experienced context which makes a combination of two (or more) words significant and might hence potentially contribute to the constructional process of delexicalised or semi-delexicalised patterns. Moving away from sole introspection, more recent context-oriented approaches predominantly rely on results gained through corpus research or similar sources for extensive data collection, like large-scale surveys. In the analysis of corpora, syntagmatic relations become obvious through concordance lines, while reoccurring patterns throughout different texts and genres help with analysis of paradigmatic relations.