Constructions and Frames - Volume 6, Issue 2, 2014

Turkish spoken in the Netherlands (NL-Turkish) sounds different in comparison to Turkish spoken in Turkey (TR-Turkish) due to Dutch influence. In addition to borrowed Dutch words/phrases, Dutch influence on NL-Turkish is also observed through literally translated constructions. This study investigates the Dutch influence on NL-Turkish constructions with subject pronouns. Analyses of NL-Turkish and TR-Turkish spoken corpora do not reveal any significant differences in terms of subject pronoun frequency. However, qualitative analysis of the data reveals some unconventional cases of subject pronoun use in NL-Turkish. In these cases, subject pronouns do not lead to unconventionality on their own but as parts of larger constructions that are copied from Dutch as chunks. Following the principles of usage-based approaches, these unconventional constructions are further analyzed in terms of their level of schematicity and flexibility.

This paper investigates the cross-linguistic applicability of the concept of frame as developed in the Berkeley FrameNet project. We examine whether the frames created for the annotation of English texts can also function as a tool for the assessment of the accuracy of English-to-Japanese translations. If the semantic structure of a source text is analyzed in terms of the frames evoked by its constituent words and the ways in which the elements of those frames are realized, then those frames, their constituent elements, and their interconnections must somehow be present in the translation. The paper concentrates on passages involving causation, as causal relationships are considered by many to exhibit the most salient differences in rhetorical preference between the two languages.

Usage-based CxG approaches share the central assumption that any grammar has to be acquired and organised through input-based abstraction and categorisation. Diasystematic Construction Grammar (DCxG) is based on the idea that these processes are not sensitive to language boundaries. Multilingual input thus results in multilingual grammars which are conceived of as constructicons containing language-specific as well as language-unspecific constructions. Within such systems, phonological structures play an important part in the identification of schematic constructions. However, the status of phonology in DCxG, as in CxG in general, yet remains unclear. This paper presents some arguments for including phonological elements systematically in the construction-based analysis of (multilingual) constructional systems.

The goal of the present study is to examine whether clause-combining rhetorical preferences that differentiate between Hebrew and English are maintained across grammars, specifically, in the context of text production in a non-native language. It examines the usage of various bi-clausal constructions marking different levels of event integration in texts written by advanced speakers of English, all native monolingual Hebrew speakers. The data analyzed consist of personal experience narratives that were collected from high-school and university-level students. These texts are compared to narratives that were collected from native speakers of both languages following the same design of study. Quantitative and qualitative analyses show differences and similarities between the three populations in terms of clause-combining strategies. They reveal that not only the constraints of the L1 but mainly those of the L2 guide non-native speakers in their choice of bi-clausal constructions, as devices expressing event integration. Results further show that event integration is reflected by constructions at different levels of the grammatical system, and that constraints on bi-clausal constructions at the more local, morpho-syntactic level are echoed by constraints at the level of discourse itself as a construction.

The present paper investigates the question whether different languages can be categorized into ‘constructionally tolerant’ languages, which grant speakers considerable freedom to combine syntactic constructions with lexical items in non-conventional ways, and ‘valency-driven’ languages, which impose stronger restrictions on the way in which constructions and lexical items can be combined. The idea of such a typological distinction is sketched for instance by Rostila (2014). In order to explore possible effects of constructional tolerance, a grammaticality judgment task is administered to speakers of English and French, which are two languages that differ with regard to this phenomenon: English verbs can be used across different argument structure constructions with relative ease, French verbs are more constrained. Both populations of speakers are exposed to stimuli sentences of varying creativity in a second language, namely German. The paper advances the constructional tolerance hypothesis, which states that speakers of a constructionally tolerant language should judge non-conventional examples in an L2 with more lenience than speakers of a valency-driven language. The experimental results are in line with this hypothesis, but they also suggest that grammaticality judgments are influenced by the availability of a productive L1 construction that shows functional overlap.

In research on bilingualism it is often assumed that linguistic structures can be shared across languages. The emphasis on generalization and categorization in construction grammar also seems to imply that speakers can develop cross-linguistic representations. This contribution argues that generalizations can occur only on the semantic level. Data from typologically distinct languages shows that generalizations over form are not likely to play a role in language processing. It is further argued that neither syntactical nor grammatical form is needed in order to explain syntactic transfer.