When speakers produce and interpret language structures, they rely on a structured inventory of patterns that save them from having to memorize every individual expression. These patterns or grammatical constructions are productive to varying degrees. E.g., the transitive construction X+Verb+Y (to eat an apple), which accepts many verbs, is far more productive than the Nominative-Genitive construction in German, which is restricted to a handful of verbs (e.g. gedenken: Wir gedenken der Opfer ‘We remember the victims’). Productivity, defined as the domain of application of a grammatical pattern, constitutes the very nature of grammar. It plays a fundamental role in language use and acquisition (where increase in productivity is the driving force), and it is a key concept in capturing language change (i.e. the constant interplay between grammatical patterns exhibiting (gradually) increasing or decreasing lexical scope, and hence increasing or decreasing productivity).
Still, until now the phenomenon of productivity is poorly understood, especially with regard to syntactic constructions. Productivity is an abstract property of linguistic structures that forms part of the implicit knowledge speakers have about a language. It is a theoretical construct that refers to what is possible in a language, but this ‘virtual openness’ of constructions can only be observed indirectly, when productivity is “at work”. However, here is the main difficulty: productivity manifests itself “at work” in multiple facets. First, a construction’s productivity is reflected in the sum of utterances produced by the speakers of the language, i.e. in language usage. Word counts in language corpora can measure the lexical scope of a construction (= realized productivity). In the case of corpora from different historical stages of a language, the evolution of this scope can thus be charted. A fundamental conceptual debate in this respect revolves around the issue that realized productivity does not suffice to define a construction’s productivity, since constructions may be extendible beyond necessarily closed-ended corpora, and constructions with fewer types may have a large number of original applications, i.e. hapaxes or one-offs and thus be very extendible too. A second guise of productivity is at work in the mind of the language user when (s)he applies grammatical patterns in on-line speech, i.e. in production and comprehension, or when (s)he is asked to judge such acceptability off-line. Productivity in the mind can be measured through psycholinguistic experiments that tap into on-line language processing (e.g. reading times, EEG) or off-line acceptability judgments. In addition to the tension between productivity in attested usage and in the mind, there is a third dimension coming into play, since productivity is likely to be influenced by individual differences related to socio-biographical factors and personality traits, which may make speakers more or less prone to extending patterns.
The ground-breaking nature of this project lies in the fact that attested discourse, on-line (and off-line) processing and individual differences are integrated in a data-driven approach that is inter-disciplinary, integrating synchronic and diachronic linguistics and psycholinguistics.