[Preprint] In: Yearbook of the German Cognitive Linguistics Association, 3/1, 169-194. Berlin: New York: de Gruyter. The constructional patterns of L2 German meteorological events by native French, Dutch- and Italian-speaking L1 learners Sabine De Knop 1 , Paul Sambre 2 , Fabio Mollica 3 1 Université Saint-Louis Brussels (Belgium) 2 University of Leuven (Belgium) 3 University of Milan (Italy) Abstract Weather expressions can be seen as a relevant topic in foreign language (FL) teaching. Traditional teaching methodology focuses on fixed FL expressions like impersonal verbs discarding both the rich array of verbal and nominal constructions for typical weather states, as well as the learners’ needs for their production. Construction grammar may fill this gap in two respects, by focusing on weather expressions produced by second language learners (L2). This study describes: (1) the variation in the preferred L2 constructional patterns for weather forecasts in three samples of non-native speakers of German (French, Italian and Dutch); (2) an inventory of constructional and associated lexical deficits in German-L2, compared to the participants’ L1. Descriptive results lead to constructional recommendations for language pedagogy of L2 German. From a contrastive perspective, we explore the influence of L1 on L2-constructions, in order to show possible learner interferences. More specifically, we contrast the Dutch learners as representatives of a Germanic language, with the speakers of the Romance languages. Finally, we propose some guidelines for construction-based teaching as an alternative to the conventional lexical approach to meteorological expressions. Keywords: construction grammar, weather expressions, German, constructional and lexical deficit, error analysis 1 A rationale for weather expressions in foreign language teaching as a constructional issue Small talk about the weather belongs to the well-known discursive strategies (Coupland and Ylänne-McEwen 2000) used by speakers to socialize and to start communication with others. Weather discourse does not simply depend on a specific meteorological vocabulary, but also on formulaic or idiomatic expressions. For Wray (2002: 79) weather expressions even constitute “an interesting case for research into formulaicity” (Wray 2002: 79). Her claim is based on Crystal’s (1995) observation that the fluency of the delivery of weather forecasts strongly depends on the “broadcaster’s ability to rely on formulaic phrasing (with light winds and largely clear skies, blue skies and sunshine, widespread frost) and on standard sequences of locations” (1995: 385). In spite of the fact that meteorological discourse is a limited and well-defined discursive domain (see also Koptjevskaja-Tamm 2015; Pennesi 2007), it constitutes a major