Languages in Contrast - Online First

Based on the description of EU legislative varieties covering EU directives and their national laws of implementation in 11 languages, we are interested in the extent to which Eurolects are similar to each other, above and beyond trivial genealogical similarities. We thus utilise a variation-oriented aggregative analysis technique to address these questions: (a) What is the precise extent to which Eurolects are similar to each other? (b) Are similarities predicted by extra-linguistic affinities? (c) Do factors such as EU accession dates and language policy play a role in shaping the Eurolect clusters? Our methodology starts out from a meticulously catalogued list of corpus-based and corpus-driven lexical and grammatical features. Through the observed presence or absence of these features, we calculate in a second step the aggregate linguistic distances between all of the Eurolects. Finally, in step three, we use a well-established technique, Multidimensional Scaling, to visualize and interpret the Eurolect landscape.

This article presents a contrastive analysis of the English copular subschema [become + past participle] and the equivalent copular subschema [devenir + past participle] in French, based on web data. It is shown that both patterns are almost equally productive at the subject complement level. Furthermore, a more in-depth analysis demonstrates that, in the segment of participles with a high adjectival potential, devenir accumulates more participle tokens than become. Conversely, the reverse holds true for participles with a high verbal potential, in which case become is characterized by more participle tokens than devenir. This high amount of combinations between become and eventive participles also suggests a higher degree of passivity for become. However, in the segment of participles with an intermediate verbal potential, devenir is slightly more type frequent than become, which hints at an emerging productivity in this area for devenir as well.

This study focuses on French and English abstract nouns denoting properties that can be ascribed to humans, such as beauty, carefulness and anger. Previous research showed that some but not all of these nouns are licensed in both locative existentials (e.g., There’s an intense anger in Isabella) and possessive existentials (e.g., Isabella has an intense anger). What remains unclear is how these and other patterns correlate among themselves depending on how easily they host such nouns. We here use speaker ratings of these nouns in different constructional environments. A principal component analysis suggests that the main dimension underlying native speakers’ ratings of these abstract nouns in six different patterns is temporal limitability. This gradable distinction, strongly correlated with the locative existential, holds for both the French and English data and outweighs any French-English contrastive differences in how acceptable human property nouns are considered to be in the patterns studied.

Old (Medieval and Classical) Spanish permitted finite enclisis and as such is classified as a strong-F language, as are many archaic varieties of Romance languages. Notable about Old Spanish is that, prior to the 1500s, interpolation arrangements were acceptable and rather common, as is still the case of Galician and some dialects of Portuguese. However, from the 1500s onward, interpolation in Old Spanish was no longer productive, much like modern Asturian. This is evidence that the “strong-weak” dichotomy of FP is insufficient to explain the situation of the languages. I argue that the strength of FP should be described as not only “weak” or “strong,” but instead on a gradient scale to distinguish languages that permit a range of possible clitic arrangements.