This volume, a revised and updated version of a 2005 PhD thesis by the same title at the University of Southern Denmark, takes an in-depth look at how to (better) theorise collocations. The book takes as its starting point what one might term a traditional phraseological approach to collocations. The various facets of and influences on this approach are insightfully explored and lead the author to outline a range of shortcomings that her work proposes to address via a new theorisation of collocations based on concepts drawn from work in Cognitive and Functional Linguistics. The most prominent shortcomings identified are the presumption of full compositionality as the norm from which collocations deviate, the categorisation of collocations and where to place the boundaries between collocations and other phenomena. These are indeed issues that have been noted as problematic for some time, so a book-length investigation into how such issues might be resolved is welcome. Although this book has the ambition to theorise a very wide array of phenomena labelled collocation, the subject is significantly narrowed by the early declaration that combinations of verbs and their nominal objects will be the exclusive testing ground for the new theorisation and narrowed again when it transpires later that the work rests on the in-depth analysis of only a single collocation (break an appointment).
The content is presented in four chapters. Chapters 1 and 4 serve as introduction and conclusion respectively, chapter 2 covers the in-depth critique of the conception of collocations in traditional phraseology, but the bulk of the book (200 pages) is contained within chapter 3, which lays out the new theorisation of collocations in detail, based on a case study.
Chapter 1 (pp. 1–26) introduces the work by giving the reader a whistle-stop tour of the various theoretical traditions and perspectives that will feature in the treatment as well as previews of the main lines of argument and, crucially, the delineation of what counts as a collocation for the author. As mentioned, this is oriented toward including everything that has been discussed under the label of collocation, including concepts somewhat removed from a phraseological understanding of the term. This clashes to some extent with the narrowness of the analysis in chapter 3, however.
In chapter 2, titled ‘The foundations of the phraseological approach’ (pp. 27–88), the reader is presented with two main sections. The first explores influences and background concepts relevant to what is termed ‘the phraseological approach’ to collocations. The second section then critically assesses how this approach conceptualises collocations. Throughout, Poulsen uses ‘the phraseological approach’ mainly as a shorthand for the work of A. P. Cowie, Peter Howarth and to some extent Rosamund Moon. These are influential figures in the field of phraseology of the English language, but their contributions date back to the last century and there would be newer work to draw on. Nevertheless, doing so provides Poulsen with a stable base to work from, which has the advantage of clarity.
In terms of influences on the phraseological approach to collocations, the book traces early treatments of collocation in English back to the 1930s and the work of Palmer and Hornby (both focused on L2 pedagogical concerns). Subsequent major influences stem from what is termed ‘the Russian tradition’ (though this will surely encompass work from many parts of Eastern, and later also Western Europe). Insightfully, Poulsen also points out what are described as more hidden influences on traditional phraseological theory. These include a Saussurian commitment to a strict division between synchronic and diachronic analysis and to the arbitrariness of the sign. Poulsen argues that these commitments prevent phraseologists from recognising that there may be ways in which collocations and similar expressions are motivated, for example in that collocations may follow general linguistic tendencies including regularities in contextual configuration (as explored in frame semantics) or regularities in the metaphorical structuring of domains (as in the Cognitive Metaphor Theory). These aspects are incorporated into Cognitive Linguistic treatments to show that many expressions in fact follow general patterns of conceptualisation and contextual adaptation, whereas phraseologists seem to rely only on a determination of synchronic syntactic and semantic regularity when analysing and classifying collocations. Further, Poulsen detects influences of generative linguistics in the implicit suggestion of the phraseological approach that full compositionality ought to be the norm against which phraseology can be defined. It is correctly pointed out that full compositionality is more a generative ideal than a reflection of how language is analysed in other frameworks, such as cognitive, functional or Firthian linguistics. Poulsen here makes a distinction between compositionality (the idea that composite structures have qualities that stem only from their constituent elements and the manner of their combination) and analysability (which Poulsen uses interchangeably with motivation; an individual's awareness of the contribution of component structures to the whole). Poulsen claims that the two are ‘independent’ of each other. The figurative–literal distinction, which interacts with compositionality in phraseological classifications of collocations, is another contrast that Poulsen feels is difficult to maintain as a sharp division. This is doubtlessly correct and has been recognised for a long time within the phraseological approach, e.g. by Fleischer (Reference Fleischer1982: 38). A final influence on phraseology that Poulsen detects is the classical view of categories as discrete and established through a set of criterial attributes, in contrast to prototype categories as used in Cognitive Linguistics. Poulsen concedes that in the phraseological literature, sharp categorical boundaries are sometimes relaxed to allow for some fuzziness, but superimposed on a classical view of categories, this only leads to endless discussions of where to draw the boundaries between what is phraseological and what is not, according to Poulsen.
Following on from this analysis of influences on the phraseological approach, the second part of chapter 2 outlines the phraseological approach's theorisation of collocations which is taken as the basis for the retheorisation in chapter 3. Poulsen most closely follows Howarth (Reference Howarth1996) here and starts with the observation that, unlike Firthian or Sinclairian collocations, collocations of the phraseological approach consist of words that are in a structural relationship with each other, following Hausmann (Reference Hausmann, Bergenholtz and Mugdan1985). The collocational pattern on which the book focusses (verb + nominal object) is an example of this, but forming a syntactic unit is only a preliminary condition for classification as a phraseologically interesting collocation. Again following Howarth, Poulsen explains that restricted collocations have a figurative (or at least semantically bleached) element to distinguish them from free combinations, but also a literal element to distinguish them from idioms. Unlike free combinations (which are fully compositional) and idioms (which are not compositional at all), restricted collocations are partly compositional. Poulsen rightly draws attention to some of the questionable borderline cases that this produces: foot/pick up/meet/settle the bill are all restricted collocations on account of the figurative verbal component and consequential partial compositionality, but pay the bill, which is the most common form among the cluster and clearly an institutionalised expression, would have to be considered a (phraseologically uninteresting) free combination on account of being fully compositional and literal. As Poulsen argues, each of the criteria on which the distinctions are based is difficult to pin down (most of all perhaps the literal/figurative distinction), but as soon as grey areas are permitted, the totality of the categorisation schema starts to fall because it is built on clear distinctions being possible. It is difficult to argue with Poulsen's criticisms here and they ring true in relation to more work in the phraseological tradition than just Howarth (Reference Howarth1996), but taking Howarth's work, almost thirty years old, as the best exemplar for ‘the phraseological approach’ arguably does not do full justice to what is a much more diverse and up-to-date field.
Chapter 3 proceeds to present the promised reconceptualisation of the notion of collocation, in ‘Collocations in a functional and cognitive framework’ (pp. 89–292). This is delivered, in the main, via the use of an in-depth case study of the collocation break (an) appointment. The chapter starts, however, with a couple of sections discussing philosophical-theoretical concepts of relevance to the enterprise: Poulsen characterises the study of collocations as the study of expressions that are entrenched (a psychological phenomenon) as well as conventionalised (a social phenomenon) and that serve the function of reproducing connections between language and contexts of situation. The latter is theorised via frame semantics (Fillmore Reference Fillmore1982), which provides the overall analytic structure for the case study.
The case study proper gets underway from the third section of chapter 3 with the analysis of break an appointment, a collocation exhibiting a syntactic relationship between its constituents and containing, according to Poulsen, a basic-level verb (an aspect that will assume significance later in the analysis). Concordances of the lemma break with the lemma appointment as its nominal object are investigated using data from the British National Corpus (BNC). The analysis proceeds in a ‘semasiological’ fashion, that is, by starting with the constituent lexical items rather than the collocation itself. This arguably has a decisive influence on the findings and the opposite perspective is never considered along the way. For break, a sample of 1,000 occurrences in the BNC are investigated, out of over 10,000 occurrences. For appointment, all 908 occurrences of the lemma are considered. Only then are combinations of break with appointment analysed – in Poulsen's sample there are five such occurrences, though there would have been nine or ten more in the BNC as a whole.
The investigation of break takes into account three main aspects: the construction types in which break occurs (only 29 per cent of occurrences in the sample instantiate the transitive verb + nominal object construction; others are intransitive, phrasal verbs, etc.), the semantic domains to which occurrences of break belong, as well as the image schemas associated with each occurrence. This makes for an exceptionally rich and detailed analysis which is distilled into a well-argued and carefully supported presentation of the lemma break as a complex category encompassing a network of twenty linked readings. As it happens, ten readings fall into sensorimotor domains of physical experience, linked via conceptual mappings to ten metaphorical readings in non-sensorimotor domains. For example, the most basic reading, damaging a physical object (e.g. ‘somebody has broken the glass door’), has a metaphorical extension into a non-sensorimotor domain: violation of social institutions/constructs (e.g. ‘RPR had broken the initial agreement’). There are also extensions within sensorimotor domains, such as when readings overcoming physical barriers (e.g. ‘once the blockade of the river … was broken by English slips’), opening physical containers (e.g. ‘Madge's house was broken into’) and damaging a body part (e.g. ‘neighbour broke his collar bone’) are presented as extensions off the source reading damaging a physical object via specific conceptualisations of object as barrier, container and body part. In this way, Poulsen shows that although readings of break are diverse, they can be arranged into a network of readings that all form a single category involving conceptualisation via image-schematic structure, regardless of whether break occurs as a phrasal verb or in one of the other structural configurations, and regardless of whether or not break is part of an entrenched and conventional expression.
In similar fashion, occurrences of appointment are analysed and distilled into a set of readings, including (1) arrangement for a meeting, (2) placing somebody in a position (i.e. an appointment to a position) or (3) the position itself (e.g. to fill an appointment), (4) accessories for cars, rooms or people. Emphasis is placed on how the different readings co-occur with verbs that elaborate salient substructure of the frames evoked: ‘While the noun evokes the dominant [but schematic] frame …, the function of the verbs is to provide a perspective on the frame by profiling a specific frame …. In the case of polysemy, the verb may help to identify the dominant frame’ (p. 230). Crucially, Poulsen shows how this works regardless of whether the verb + noun combination is conventional and entrenched or ad hoc. For example, within the dominant frame evoked by the first reading (‘arrangement for a meeting’), there are a number of basic-level verbs that each function as the default verbs in the specific frames of what people do with appointments (break/keep/have/get/give/make). A further set of verbs in each specific frame are said to be at a subordinate level (for the frame evoked by break + appointment, for example, those are listed as miss and cancel, which specify the frame in even more detail, while still being conventional). Beyond basic-level and subordinate-level verbs, there are also ad hoc combinations in the specific frames which are used to further specify the frame (e.g. arrive late for/forget an appointment, all within the break + appointment frame).
Looking at verbs that occur with appointment foreshadows the analysis of the integration of break and appointment as a collocation in the final subsection of chapter 3. Here, the detailed work on break and appointment as separate words is leveraged to show that the collocation, far from instantiating an arbitrary or ‘irrational’ (Jespersen Reference Jespersen1904: 17) combination, is in fact one that fits very well with the behaviour of its constituent words and general mechanisms that kick in when two items and their frames are integrated. Poulsen is careful to say that this does not make the conventionalised and entrenched status of the collocation predictable, nor is the collocation simply the result of an additive process (one word added to another). But by showing in great detail how the composite structure of break an appointment follows from the general mechanisms of integration applied to the two constituent words and their frames, Poulsen does make a strong case for the collocation being motivated.
This leads on to the promised retheorisation of verb + nominal object collocations via additional observations around the focus expression break an appointment. Poulsen firstly proposes a binary distinction between verb + nominal object combinations with basic-level verbs, on the one hand, and those with more specific verbs, on the other. Both of these combinations are said to have a high likelihood of entrenchment. Combinations with basic-level verbs (in the case of appointment these are break/keep/have/get/give/make) map out the general and highly schematic possibilities allowed by the frame of the nominal object and are sometimes described as support-verb constructions. Combinations with verbs that are somewhat more specific and therefore at a medium level of schematicity (e.g. cancel/miss an appointment as more specific elaborations on break an appointment) are more characteristic in their combination with the noun compared to basic-level verbs. This is because basic-level verbs, due to their high frequency, are bound to occur with a very wide array of nominal objects, whereas subordinate-level verbs are more tied to the noun. Both of these types are labelled entrenched collocations. They are in turn distinguished from open collocations that are ad hoc, rather than entrenched and conventional combinations. The verb in open collocations still elaborates one of the actions expected to be associated with the noun, e.g. arrive late for/forget an appointment are still specific ways to elaborate on the basic-level break an appointment. Although they are not likely to be entrenched and conventional, open collocations are perfectly normal combinations according to Poulsen and, as with entrenched collocations, it is the noun (rather than the verb) that evokes the relevant frame during integration. This latter feature distinguishes open collocations from the final category of verb + nominal object combinations: in free combinations, the verb does not elaborate an aspect of the frame evoked by the noun, but rather evokes its own frame (e.g. discuss an appointment). Here, the frame of a discussion is evoked, not one of the things usually done with appointments. Poulsen summarises: ‘in a given usage situation, there is typically freedom both to choose from a range of conventional and entrenched collocations and to choose a more specific expression, which may not be conventional but still perfectly normal’ (p. 252). All of this remains within the domain of what is to be theorised as a collocation. Poulsen's taxonomy of entrenched and open collocations versus free combinations ‘stresses the continuity between conventional, entrenched collocations and open collocations, which both categorize intrinsic frame knowledge’ (p. 255), against non-collocations (free combinations) where verbs do not elaborate aspects of the frame evoked by the noun. Poulsen claims that this remedies the mistaken attempt of traditional phraseology to reduce a continuum of word combinations to classical categories. Instead, frame-related continuities across conventional and non-conventional expressions are given greater weight in the delimitation of the domain of collocations, so that collocations can be recognised as an important language resource allowing the combination of ‘convention with flexible lexicogrammatical conceptualisation’ (p. 277), a point that is given prominence in the title of the book.
The remainder of chapter 3 makes a number of connections to other theoretical frameworks and then finishes on a discussion of a set of key statements that summarise the chapter (pp. 278–91). The brief final chapter, titled ‘Collocations as a language resource: Winding up’ (pp. 293–308), restates the main findings and summarises the proposed model as ‘a functionally and cognitively based framework [that] is descriptively more adequate and has greater explanatory power than the phraseological approach to collocations’ (p. 294).
Overall, this book does not always make the usual concessions to readability; the continuous engagement with an exceptionally wide range of theoretical frameworks is admirable and shows a desire to find new connections, but it is also somewhat exhausting and possibly disorienting for the reader who is continuously asked to consider ideas from different and not always readily compatible frameworks. There are additionally a very large number of cross-references, particularly in chapter 3, but only to sections (no page numbers given), which makes it laborious to follow them. Section headings are not always as helpful as they might be when wishing to navigate quickly, and within-section turns are sometimes sharp, as when a discussion of background turns into one of methodology mid-section (p. 146). If the reader is happy to be taken on a journey of discovery, is in no particular hurry, trusts their guide to get to the destination in the end and does not mind having various interesting matters pointed out along the way, this is an enjoyable and educational read. Readers who feel they have a limited time available for the journey and would like to have a clear sense of where they are headed at all times, may not derive the same sense of enjoyment.
Despite the general care and attention to detail that is evident in analyses, there are occasionally places where one might have quibbles. The stated reading of appointment as accessories is only available in the plural form of the noun and seems now antiquated, for example. Similarly, at one point frequencies of the expression take on a mantle are reported, but the conventional expression uses the definite rather than the indefinite article. The interpretation of break as a basic-level, but cancel as a subordinate-level expression in relation to appointment seems contentious; arguably, breaking an appointment is always a negative construal, but cancel is neutral so seems to present an alternative rather than a more specific conceptualisation. Overall, some might feel that a theorisation of collocations based on frame semantics mainly replaces a delimitation of collocations based on conventionalisation with one based on how frames are evoked, and therefore deals above all differently with the challenge of drawing boundaries, rather than necessarily in a superior way, as Poulsen claims. Perhaps the most central difficulty, however, is the assumption that insight derived from the study of a single collocation generalises not only to other collocations of the verb + nominal object type, but makes possible a retheorisation of all collocations. Clearly, the case study does make pertinent and well-founded suggestions, but it represents a study of one particular collocation. Arguably, significantly more breadth would have been needed to make more than a speculative suggestion about collocations in general.
Without a doubt, however, there are treasures and genuinely fresh insights to be enjoyed by readers of this book. Principally, these are perhaps the continuities shown between conventional expressions and those that are less conventional but still rely on the same frame integration mechanisms and that (some) collocations may not be as arbitrary as they might at first appear. Consequently, this book makes a welcome contribution to the discussion around the theorisation of collocations.
