1
‘Formulaic language’ will be used as the neutral mass (uncountable) noun, while ‘formula’ is used as the neutral count noun (with the plural ‘formulas’, other than in quotations where ‘formulae’ occurs in the original). A new term, ‘morpheme equivalent unit’ (MEU), is introduced in Chapter 2. It is theory-sensitive, and used to refer to the items in the lexicon of an individual speaker or hearer. ‘Formulaic sequence’, a term introduced in my earlier work (Wray 1999, 2002b; Wray and Perkins 2000), has a specific definition (see Chapter 8) and is used in that sense only. In practice this means it is used to refer to items observed in text that are inferred to be MEUs for (usually) large groups of individuals in a speech community. Detailed discussion of definitions is provided in Chapter 8.
2
A fourth conclusion, that access to the lexicon is distributed across the brain (Wray 2002b: Chapter 14), is less central to this book and so is not used as an anchoring conceptual claim here. It is, however, mentioned later, as it becomes relevant in specific discussions.
3
The opposition of ‘novel’ with ‘formulaic’ is customarily made in discussions about formulaic language, and it is a useful shorthand here. However, as the three claims described here unfold, it will become clear that the distinction is not sustainable, since it is no less novel to construct an utterance out of multiword lexical units than out of single words or morphemes. Of course, in the former the holistically-retrieved units may well provide runs of familiar configurations, but they are really no different from familiar words – just longer.
4
Some versions of Construction Grammar allow for everything to be lexically stored – see Chapter 7– since constructions can be partly or completely lexically unspecified. The point being made here is that the capacity for novelty must be accounted for somehow. In models that accommodate the existence of grammatical rules, albeit sometimes insertional ones that complete lexically stored frames, novelty is achieved by activating the rules to produce word strings that are not independently stored in the lexicon.
5
It is probably fair to say that the multiword strings most likely to be systematically excluded from an atomic lexicon are those containing a specified verb, for example, ‘take the trouble’; ‘make one’s way’; ‘haul over the coals’; ‘see to’. Almost all verbs inside formulaic sequences can be varied morphologically, and where the verb has a wider use than just within that sequence, there is a rationale (within that type of model) for not treating the word string as a lexical unit.
id="litres_trial_promo">
