Book review

Deconstructing Constructions

Christopher S. Butler and Javier Martín Arista. 2009. Deconstructing Constructions. Amsterdam and Phildelphia: John Benjamins

Reviewed by Florent Perek, Freiburg Institute for Advanced Studies / Université Charles de Gaulle Lille 3.

Introduction: functional and cognitive approaches

Deconstructing constructions is a collective volume edited after the eponym workshop organized in May 2006 by the Functional Grammars Research Group at the University of La Rioja in Logroño, Spain. Six of the nine chapters are derived from contributions presented at the workshop. As the editors put it, the aim of the volume is to "explore the concept of constructions from different angles and try to arrive at a better understanding of what a construction is, and what roles constructions play in the frameworks which constitute functional-cognitive space" (p. XVII). Quoting the philosopher David Allison, the editors define the concept of deconstruction as "a project of critical thought whose task is to locate and 'take apart' those concepts which serve as the axioms or rules for a period of thought". This volume thus takes an obvious epistemological stance, illustrated by actual analyses of various linguistic issues, most of which will be valuable to scholars with a general interest in the topic of argument realization (especially in English and Spanish). In this review, I will first introduce the epistemological background which underlies this volume. I will then discuss the contents of the book and relate each chapter to the general aims of the volume, emphasizing what I take to be the main input of the volume as a whole. I will intentionally not summarize each contribution thoroughly; if necessary, I will just briefly comment on the general topic of the article to show how it relates to the general concerns of the volume. For more details from the authors themselves, I refer the reader to the abstracts which can all be found online here.

The past thirty years have seen a proliferation of linguistic models for grammatical representations, from the various functional grammars to the many variants of construction grammar and the like in cognitive linguistics, not to mention the many more formally-oriented approaches. While these frameworks can be classified in terms of the theoretical commitments they make, there still remains great variety within each trend. It remains to be evaluated whether and how the most valuable contributions of each model can be integrated into a single coherent and up-to-date framework. Discussing this question in the light of several theoretical issues and providing such a framework are the primary goals of this book.

Different frameworks usually do not put the same emphasis on different aspects of grammatical organization, and while such variety can surely be beneficial to the field, it is also a source of confusion for the language scientist, who could wish for more efforts towards unity. As the editors point out in the introduction, this is particularly true of functional and cognitive approaches, which have a lot in common. First, they are both opposed to formal approaches in that they reject the conception of grammar as a self-contained formal system. Moreover, they also share as a number of basic assumptions, namely that linguistic structures are better explained and motivated through communicative function, usage and socio-cultural context (In addition to these three, cognitive linguists add a fourth dimension, cognitive adequacy, i.e., the commitment that linguistic structure should not be different from other mental representations and should comply to what we know about cognition in general.). Functionalist and cognitivist models mainly differ in the way they represent grammatical knowledge. Functionalists structure the grammar in terms of functions fulfilled by linguistic elements at the levels of syntax, semantics and pragmatics. Most functional grammars are highly lexicalized and therefore projectionist in essence, which means that (i) they assume that grammatical constraints are better expressed in terms of syntactic and semantic requirements of lexical items, and (ii) they emphasize the role of lexical semantics and incorporate detailed semantic representations. Cognitivists, on the other hand, describe grammar on the basis of a single basic construct: the linguistic symbol or sign (cf. Langacker's content of requirement), a pairing of phonological form with linguistic function, also often called constructions especially when such symbols refer to grammatical patterns. The grammar is a vast network of phonological structures, semantic structures (and combinations thereof), and symbolic structures pairing them.

Despite these minor differences, the conclusions of functionalism and cognitivism are largely compatible thanks to the shared assumptions of the two approaches, which are mutually recognized. Thus the concept of construction put forward by cognitive approaches was adopted in Role and Reference Grammar, a functionalist framework. In the last decade, research on the syntax-semantic interface in cognitive linguistics has shown that constructionist accounts appear to succeed where projectionist explanations fail, but often also incur a drawback by reducing the grammatically relevant facets of verb meaning to too strong a minimum. As many studies show (Boas 2003, Iwata 2008 and Lemmens 1998, to name only a few), lexical semantics and world knowledge play an important role and interact in non-trivial ways with constructions. This body of research suggests that constructional and lexical accounts are not mutually exclusive, but rather complementary. These early convergences show that it is not only possible but also desirable to take the strengths of functional and cognitive models and build them into a unified framework, which I believe is a worthwhile enterprise.

Goals and contents of the book

The book is organized into three parts. The first part deals with a number of theoretical issues from the viewpoint of functionalist and constructionist approaches, which relate to the aforementioned issue in that they present evidence of the relevance of constructions in linguistic description, while also showing their limits. The second part presents the Lexical-Constructional Model (hereafter LCM), a new framework explicitly aimed at bridging the gap between lexicalist and constructionist approaches. The third part presents three detailed analyses of specific constructions, couched in construction grammar (Gonzálvez-Garcia and Guerrero Medina, with an additional emphasis on verbal semantics in the latter case) or in the LCM (Rodríguez).

While the selection of articles in this volume might seem heterogeneous at first sight, they all relate to the question of how the concept of constructions can be profitably applied to various theoretical issues, with an emphasized synergy with functional frameworks. Each contribution addresses to varying extent either of the two following questions:

1.How the concept of construction can benefit functionalist frameworks towards a better account of some linguistic phenomena?

2.Where are constructional approaches limited and, in turn, how insights from functional frameworks can complement them to fill in for these limits?

The first question is addressed mainly by Pedersen's, Martín Arista's and Guerrero Medina's contributions.

Martín Arista applies the concept of constructions to the domain of morphology in a Role and Reference Grammar framework in order to develop a typology of morphological constructions illustrated by data from Pitjantjatjara/Yankunytjatjara. He defines morphological patterns as including general positions and language-specific parameters on these positions. This model gives the morphological component of RRG a more integrated situation in the syntax, in contrast to previous positions situating morphology in a "pre-lexicon".

Pedersen makes a compelling contribution to linguistic typology, by recasting Talmy's traditional classification in terms of information structure. His formulation relies on the concept of construction as a unit, associated with particular communicative functions, in this case encoding main versus supportive information.

Gonzálvez-Garcia's contrastive study of secondary predication constructions in English and Spanish shows that while in both languages similar constructions must be posited to account for the partial productivity of these patterns, corpus data shows that each language displays its own distributional properties, providing further evidence that argument structure is not only construction-specific but also language-specific.

Guerrero Medina argues that lexical semantics does not account for the choice of get-passives (over regular passives) in English and shows the importance of pragmatic and contextual factors to motivate his own account in terms of a family of related constructions.

The second question is addressed by García Velasco's, Fernández', and Rodríguez' articles.

García Velasco deals with eponyms, a particular class of innovative coinage taking a proper noun as input, and argues that coercion as it is implemented in a construction grammar framework does not appropriately account for eponyms, since they necessitate such a great deal of contextual information, and suggests that the contextual component of Functional Discourse Grammar could fill this gap. Besides, García Velasco's data represent an interesting challenge for usage-based theories of language.

Fernández rejects a purely constructional account of the so-called caused-motion construction in English (as discussed by Goldberg 1995) and offers an alternative account based on Pustejosky's (1995) Generative Lexicon, that demonstrates the importance of a detailed semantic representation to predict the meaning of expressions seemingly instantiating the caused-motion construction. She argues for the existence of two different interpretations of such expressions, which she explains by two distinct mechanisms of meaning composition (co-composition and merge), the choice between either reading being determined by the lexical representation of the verb.

Rodríguez' article nicely illustrates the LCM introduced in Part 2 (see below) by providing an analysis of the English inchoative construction couched in this framework. In line with the goals of the framework, Rodriguez analysis shows the importance of detailed lexical semantics to predict the occurrence of a verb in the construction.

The Lexical-Constructional Model

I devote an entire section of this review to what I think is the centerpiece of this volume: the introduction of the Lexical Constructional Model, presented in Section 2. In an attempt at fulfilling the epistemological requirement that I summarized in the introductory part of this review, the LCM aims to bring together functionalist approaches, mostly projectionist in orientation, and cognitive approaches, mostly constructionist, towards a better understanding of phenomena so far mostly relating to argument structure. The LCM thus effectively conciles the best of both worlds: the commitment to finely defined lexical semantics of functionalist models and the commitment to cognitive adequacy of constructionist approaches. The two chapters in part 2 define the LCM in more details, adopting different perspectives.

The first chapter by Butler gives the genesis of the model by showing what each parent theory contributes to the LCM and details the successive steps of its emergence. More than yet another framework competing with the many others in cognitive and functional linguistics, the LCM integrates a number of ideas from the three past decades of research in various fields: Dik's Functional Grammar, Coseriu's Lexematics, Van Valin and La Polla's Role and Reference Grammar, Wierzbicka's Natural Semantic Metalanguage, Mel'Cuk's Meaning-Text Theory and finally Construction Grammar. As Butler puts it, the LCM "capitalizes on the similarities between functionalist and cognitivist approaches and achieves a synthesis of the two strands of what we might call functional-cognitive linguistics" (p. 119).

The second chapter by Mairal Uson and Ruiz de Mendoza Ibanez presents the LCM in more ("technical") details. It presents the overall architecture of the model, which consists of four levels corresponding to successive levels of meaning construction, from 'core' grammar to discourse integration. The focus moves to Level 1 of the model, i.e., the level of argument structure. The basic unit of argument structure in the LCM is called a template, containing argument realization information. The model recognizes the existence of both lexical templates, projected by the verbs, and constructional templates, which roughly correspond to Goldberg's (1995) concept of Argument Structure Constructions. The licensing of argument structure lies on the unification of a lexical with a constructional template. Both types of templates are described with the same metalanguage, which clarifies and makes the unification phenomena easier to describe formally and allows the formulation of testable hypotheses on the integration of lexical semantics with argument realization. The authors then turn to the pragmatic and discursive dimensions by evoking other types of pragmatically grounded constructions. They conclude the article with a description of the two cognitive mechanisms at work at all levels of the architecture: cued inferencing and subsumption. The former accounts for how we make sense of linguistic expressions by integrating cues from different sources that constrain interpretation. The latter accounts for the integration of semantic structures by constraining the integration of lexical templates with constructional templates in two ways: internal constraints determine to what extent the meaning of a lexical predicate can be internally adapted when it is combined with constructional templates, and external constraints specify in which ways a constructional template changes the semantic type of a predicate when the latter violates the semantic requirements of the former (a phenomenon sometimes called coercion).

The LCM is still in development, yet promising; to this day, only level 1, i.e., 'core' grammar, has been characterized, levels 2 (pragmatic implications), 3 (illocutionary constructions) and 4 (discourse integration) being still only programmatic. The LCM is thus aimed at eventually being an integrated model of meaning construction at all levels of interpretation.

Conclusion

This volume makes a significant contribution to the field of cognitive-functional linguistics. The contributions explain why it is desirable to make further move towards convergence between projectionist and constructionist approaches on the topic of argument structure, and how it is possible. The Lexical-Constructional Model, presented in the middle section of the book, is precisely designed to achieve such a goal and to capitalize on the strength of both functional and cognitive models, further enriched by the contribution of other parent theories, especially as far as it pertains to the representation of lexical semantics.

I do believe that the LCM is a promising approach that adequately address issues in argument realization. It deals with old problems using old solutions, albeit a combination thereof. Its main merit is to rehabilitate the crucial role of lexical semantics in predicting argument realization, without abandoning the idea of constructions as the locus of argument structure flexibility and creativity. The idea that fine-grained lexical semantics are likely to refine constructional accounts has been around for a while, but the novelty of the LCM lies in the powerful tools it provides to formalize and analyze these phenomena. Another promising feature is the incorporation of implicational, illocutionary and discourse levels, which to this day have yet to be developed, to provide an integrated model of meaning constructions. I am looking forward to the developments of the model in this direction.

My only regret regarding the presentation of this model in the volume is that we would have to see more case studies showing the model at work. The LCM is only illustrated by a few in-text examples and Rodriguez' contribution. A few additional case studies exemplifying the treatment of key issues in argument realization by the LCM would have been beneficial to show the superiority of the model to purely constructional or purely projectionist approaches.

In conclusion, this book should be of concern to all scholars with an interest in the topic of argument realization, especially those working in a functional or cognitive linguistics tradition. The various analyses presented therein address issues that can also concern a larger audience, especially researchers in English and Romance linguistics. Language typologists will also find a few relevant contributions (notably Pedersen's, Martín Arista's and Gonzálvez-Garcia's contrastive study).

References

Boas, H. C. (2003). A constructional approach to resultatives. CSLI Publications.

Goldberg, A. (1995). Constructions: a construction grammar approach to argument structure. Chicago: University of Chicago Press.

Iwata, S. (2008). Locative alternation: a lexical-constructional approach. Amsterdam: John Benjamins.

Lemmens, M. (1998). Lexical perspectives on transitivity and ergativity: causative constructions in English. Amsterdam: John Benjamins.

Pustejovsky, J. (1995). The Generative Lexicon. Cambridge: MIT Press.

Links

© 2002-present ICLA; all rights reserved.
Site admin information and contact