In the introduction to the volume Contrastive Studies in Construction Grammar Boas (2010: 11) underlines the need for “creating a complete record of constructional inventories of languages that we want to include in our cross-linguistic constructional investigations”. Construction Grammar linguists have addressed this request by developing so-called “constructicons”, i.e. repositories of the lexicogrammatical constructions of a language based on Semantic Frames (see among others Fillmore et al. (2012) for Framenet and Boas & Ziem (2018) for the German Gfol). In a cross-linguistic or contrastive perspective, most studies have started from English constructional patterns for the description of equivalents in other languages (Boas & Gonzalvez-Garcia (2014: 2)).
My study adopts a different approach as it wants to establish an inventory of authentic and frequent German constructions in contrast with Romance languages, and more specifically French, without referring explicitly to Frames. Starting from examples collected in German press articles and from the German deTenTen20 corpus of the SketchEngine, the aim is, first, to define typical German constructional patterns for which no or hardly any equivalents exist in Romance languages and, second, to discuss the motivation behind the issues related to the contrastive analysis German–French.
Drawing up an inventory of German constructional patterns and their possible equivalents in French is important for translation studies but also for foreign language learning and teaching.
One of the most important characteristics of Contrastive Linguistics, in comparison to other comparative approaches to the study of language, is its connection to applied fields of linguistics, such as foreign language acquisition and teaching (Lado 1957, Kortmann 1998) and translation and interpreting studies (e.g. Hansen-Schirra and Czulo 2017). While contrastive linguistics has not so far been prominently linked to psycholinguistics, there are moreover indirect relations to this field via the aforementioned disciplines (e.g. Lörscher 1991), not least because the study of bilingualism is directly relevant to contrastive linguistics (e.g. Grosjean 2024). One of the central challenges that all of these approaches face is crosslinguistic variation in form-function mapping. Language learners need to learn that expression x (e.g. although) from language A (English) is not fully equivalent to expression y (obwohl) from language B (German), translators and interpreters face the challenge of finding linguistic near equivalents (under different constraints), and the bilingual brain in general needs to manage the n-to-n relations between meanings and forms from different languages. A key challenge for all of these fields is, therefore, to model form-function mappings within as well as across languages. This challenge falls squarely into the field of Contrastive Linguistics, where evidence from language acquisition and learning, translation and interpreting studies and psycholinguistics as well as bilingualism can be brought together.
In this talk I will propose a model of form-function mapping in the domain of concessivity, based on data from English and German (cf. also Gast 2019). The distribution of concessives is captured by a set of domain-general variables, i.e., variables that are relevant beyond the domain of concessivity. Generally speaking, there are four types of relations between two expressions x (from language A) and y (from language B) if x and y have overlapping distributions, (i) dist(x) ~ dist(y), (ii) dist(x) ⊂ dist(y), (iii) dist(x) ⊃ dist(y), and (iv) dist(x) ∩ dist(y) (where dist(e) maps expression e to the set of context types where e is used). From an applied point of view, I will focus on cases (ii) and (iii), which amount to explicitation and implication, respectively. I will show that explicitation is prevalent in written translation because translators avoid information loss, whereas interpreters, working under time pressure and high cognitive load, tend to choose more general correspondences in the target language, minimizing the risk of mistranslations.
The Universal Dependencies (UD) framework aspires to offer a cross-linguistically consistent syntactic annotation scheme. In doing so, it inevitably clashes with concepts defined in traditional descriptions of individual languages (which are often mutually incompatible with traditions held in other languages). This talk focuses on one such problematic concept, the indirect object – a grammatical relation whose presence in the UD taxonomy was inspired by traditional grammars of some Indo-European languages, yet its classification as core argument makes a UD-style “indirect object” quite different from the (intuitively expected) RECIPIENT role in, e.g., Spanish or French. Using data from a range of languages, I show how the ambiguity in the guidelines leads to divergent annotation practices, particularly in constructions involving recipients, benefactives, and goal arguments. While this obviously creates challenges for typological comparison, I will also discuss possible workarounds, as well as potential future rethinking of the UD guidelines with respect to modeling of indirect objects.
