(Interactional) Construction Grammar
The term Construction Grammar does not denote a single theory but rather - in the sense of the plural form Construction Grammars (Ostman and Fried 2005) - refers to what Fischer and Stefanowitsch (2006: 3; my translation) call a “family of theories that all share the conviction that human language consists of signs (i.e. form-meaning pairings) on all linguistic levels”. This focus on the sign- based nature of language goes back to the very foundations of Construction Grammar. Fillmore (1988: 36), for example, defines grammatical constructions as syntactic patterns which are “assigned one or more conventional functions in a language”. For Croft (2002: 21), too, what is decisive about Construction Grammar is the fact that it “treats grammatical units as fundamentally symbolic, that is, pairings of grammatical form and the corresponding meaning or semantic structure”.
This extension of the (Saussurean) sign to all linguistic levels, such as phonology, morphology, syntax, text and discourse, is not just shared by those approaches that include the words Construction Grammar in their name (e.g. Croft’s 2002 Radical Construction Grammar, Bergen and Chang’s 2005 Embodied Construction Grammar, Sag’s 2011 Sign-based Construction Grammar or Van Trijp’s 2008 and Steels’s 2011 Fluid Construction Grammar) but also by many of those within the field of Cognitive Grammar. Langacker’s (1987) Cognitive Grammar is indeed so similar to Construction Grammar approaches that Goldberg (1998: 205) opts for an interchangeability of both terms: “Construction Grammar (also Cognitive Grammar)”. In a textbook about Cognitive Grammar, Taylor (2002: 20-21) provides the following basic definition of what Cognitive Grammar is about:
Cognitive Grammar is driven by the idea that language is essentially and inherently symbolic in nature. Linguistic expressions symbolize, or stand for, conceptualizations. I shall refer to this basic assumption as the symbolic thesis. [...] The symbolic thesis actually amounts to little more than the claim that language is in essence a means for relating sound and meaning. [...] What is special about the Cognitive Grammar approach is that syntax itself is regarded as inherently symbolic, and is therefore handled in terms of symbolic relations between phonological and semantic structures. (Taylor 2002: 20-21)
Despite some recent criticism (e.g. Jacobs 2008) of attempts at viewing the whole of language as signs, the “symbolic thesis” (Taylor 2002: 38-60) is still one of the mainstays of both Construction and Cognitive Grammar.
The reason why the idea of treating everything from morphemes to texts or discourse patterns as constructions appears so attractive to many linguists is that the concept of meaning has been extended considerably in Construction Grammar. While for Saussure signs simply consist of form and meaning, constructions consist of pairings “of form with meaning/use” (Goldberg 1996: 68). This extension implies that a construction not only contains information about its semantic (and maybe functional) properties, but also includes every relevant fact about the context it usually occurs in. In other words, “facts about the use of entire constructions, including facts about registers, restricted dialect variation etc., are stated as part of the construction” (Goldberg 1996: 69).
Especially for empirically oriented linguists, for whom most theories of syntax are problematic because of their tendencies to attempt maximally context- free descriptions of rules and patterns, Construction Grammar may offer a way to include any type of information into a construction that empirical analyses prove important. Because of this advantage, there have been some attempts in recent years within the field of Conversation Analysis and Interactional Linguistics to establish an approach within Construction Grammar which could be called Interactional Construction Grammar (Auer 2006b; Deppermann 2006; Gunthner 2006a, b, c, Gunthner and Imo 2006; Imo 2006, 2007a, b, 2008, 2009, 2011a, b, 2012; Zima and Brone 2011). In an article published in 2006, for example, Deppermann (2006: 1) asks whether Construction Grammar could be expanded into a “grammar for interaction”. The reason why Construction Grammar might indeed be a useful theory of syntax for Interactional Linguistics is that three basic tenets are shared by both approaches. First, the idea of viewing the basic units of language as holistic gestalts - i.e. constructions - which include all relevant information about their morphological, syntactic, prosodic, functional, sequential, situational and genre-related properties. Second, the symbolic thesis, which claims that there are no merely formal structures but that there is always a combination of formal and functional aspects. Third, the usage-based approach that both Construction Grammar and Interactional Linguistics/Conversation Analysis share, i.e. the idea that constructions emerge out of repeated use in interactions via routinization processes. Drawing on converging interests between these approaches, Deppermann (2006: 43) arrives at the following conclusion:
Commonalities of Construction Grammar and Conversation Analysis with respect to these three claims are sketched. As a conclusion, the paper argues for the combination of detailed sequential analysis in a CA mode with corpus-linguistic methods, and it pleads for the integration of cognitive and interactive perspectives on the meaning and use of grammatical constructions.
With Construction Grammar, it is possible to reformulate the findings of Interactional Linguistics/Conversation Analysis - which are always based on detailed, highly qualitative, sequential and context-sensitive analyses - within the framework of a theory of syntax that, in turn, opens up new vistas for Interactional Linguistics/Conversation Analysis. These new vistas concern the morphological, semantic and general cognitive aspects of constructions about which the strictly empirical approach of Interactional Linguistics/Conversation Analysis has to be silent for methodological reasons:
On the other hand it is clear that Conversation Analysis is not able to explain quite a lot of properties of constructions. [...] Most of all, this concerns morphological and semantic aspects that only rarely provide the basis for conversation, and if they do, they do it only in a very superficial way. Therefore, it is necessary to draw on cognitive concepts [...]. (Deppermann 2006: 61; my translation)
Many of the cognitive concepts mentioned by Deppermann can be provided by Construction Grammar. One important concept that allows for many of the aforementioned aspects to be explained is the idea of a network of constructions, which is taken as the repository of the organization of the grammatical (i.e. constructional) knowledge of a language (see, for example, Boas 2010). Usually, the inventory of constructions is conceptualized as some kind of a “structured inventory” that contains - in a very broad sense - the “speaker’s knowledge of the conventions of their language” (Croft 2002: 25). Croft’s wide focus on the “conventions” of a given language has the advantage that it is possible to include any kind of information that is necessary to describe the form and function of a given construction as well as all types of relations to more or less similar neighboring constructions into the description of a construction.
The concept of a structured network has another great advantage. It makes it possible to explain processes of the “blending” (Fauconnier 2004) or “amalgamation” (Gunthner 2006c; Imo 2007b) of constructions as well as processes of grammaticalization, pragmaticalization and lexicalization. All of these phenomena can be viewed as results of a shifting of positions of single constructions within the whole network, the construct-icon. It is important to keep in mind this constantly emerging structure of the construct-icon and not to fall into the trap of viewing this network as a permanently fixed inventory: “Our characterization of schematic networks has emphasized their ‘static’ properties, but it is important to regard them as dynamic, continually evolving structures. A schematic network is shaped, maintained and modified by the pressures of language use.” (Langacker 1987: 381)
In spite of the commonalities between Interactional Linguistics and Construction Grammar - the focus on processes of routinization as a driving force of language and the demand for sign-based, holistic descriptions including information about prosody, context etc. - many questions still remain unanswered. One of the most important problems that emerges when Construction Grammar and Interactional Linguistics are combined is how to reconcile the rather schematic, pattern-based approach of Construction Grammar with the process-oriented approach of Interactional Linguistics, in which syntax is viewed as an open, temporally emergent structure: Auer (2000) uses the term “on-line syntax” to refer to the fact that much of language is produced incrementally and that these increments are often triggered by factors outside of “syntax proper”. For example, a sentence can be expanded by adding new material in a piecemeal fashion if the recipient does not react or shows a lack of understanding (e.g. Auer 2007b and the whole special issue of Pragmatics 2007 edited by Couper-Kuhlen and Ono on “Turn continuation in cross-linguistic perspective”). The problem of incremental utterance expansion will be addressed in this paper by contrasting the use of wide appositions in monological stretches of talk and incremental, apposition-like structures in interactional stretches of talk.
-  For an extension of the range of Construction Grammar to text and discourse patterns seeOstman (2005) and to genres Gunthner (2006b; 2010) and Imo (2010).