A usage-based model of m/ftying+NP as a construction
Up to this point, it has been shown how non-attributive and attributive instances of mittying are used to tie the local context to a topical antecedent. This section aims to show that the occurrences of mittying + NP can be analyzed by means of a construction in the sense of a usage-based approach to Construction Grammar.
Construction Grammar from a usage-based point of view
In a series of seminal studies, Fillmore, Goldberg, Croft and other advocates of Construction Grammar (CxG) argue that the primary units of the linguistic competence are surface-near constructions rather than atomic units and abstract autonomous rules (see Fillmore 1988, Fillmore, Kay, and Catherine O’Connor 1988, Goldberg 1995, 2006 and Croft 2001, amongst others). Constructions are considered to be entrenched symbolic pairings of form (phonological, morphological and syntactical representations) and meaning (usually both semantic and pragmatic representations) which can differ with regard to internal complexity and specification and which are created and shaped by actual language use. Most of the branches of Construction Grammar are usage-based (cf. Lan- gacker 2000, Barlow and Kemmer 2000, Bybee and Hopper 2001, Tomasello 2003 and Bucker 2012, for example) and reject the distinction between autonomous structure- and rule-based parts of grammar (substantially brought about by an innate universal grammar) on the one hand and the lexicon (as the host of all idiosyncrasies in a certain language) on the other. Instead, language is seen as constantly emerging within and through the actual use of symbolic pairings of form and meaning.
Recent usage-based studies in Construction Grammar have shown that the use of constructions as constructs (Fried and Ostman 2005: 18) in talk-ininteraction is both complex and context-driven since it strongly interacts with sequential positions, genres and sometimes also social registers (cf. Selting and Kern 2009). For example, Gunthner (2010, 2011a) notes that dense constructions are particularly common in conversational narrative genres, while Imo (2010a) considers German mein Problem ist/mein Thema ist ‘my problem is/my topic is’ to be a conventionalized means to initiate topic talk in German radio phone-ins. Such studies reveal that constructions can emerge in virtually every “ecological niche” of language use - such “niches” being sequential positions, genres and other sociolinguistic constellations of all kinds - as the result of specific communicative conditions and particularities which characterize these “niches”.
Due to this, it is by no means unlikely that there are still many patterns in talk- in-interaction which have not been recognized yet as corresponding to entrenched pairings of form and meaning. In my opinion, one of these patterns is mittying + NP. However, in order to avoid a naive exemplar-based approach with a behavioristic bias, not every pattern of language use should directly be identified with a construction (see Bucker 2014). It needs to be shown instead that the degree of structural and functional “autonomy” a certain pattern of language use has is really high enough to justify the postulation of a cognitively entrenched construction.
-  See Birkner (2006), Imo (2007), Gunthner (2009) and Bucker (2009) with regard to furtherexamples of constructions which have not been discussed as entrenched form-meaning pairingsbefore.