Search

There is a long standing debate between aphasiologists on the essential factor that constitutes the behavioral patterns of loss and preservation in agrammatic Broca's aphasia. It has been suggested that linguistic complexity plays a crucial role: linguistically complex structures are more difficult to produce than linguistically simple ones. However, linguistic complex structures are often less frequent in a language; for example, simple active sentences are more frequently used than linguistically more complex passive sentences. Hence, it might be that it is not linguistic complexity but frequency that determines agrammatic behavior. Frequency may play a role at several levels. For agrammatic patients, for example, the frequency of sentence constructions may be crucial, whereas for fluent aphasic speakers word frequency influences performance. The present study presents corpus frequency data for constructions that have previously been used to show the influence of linguistic complexity on Dutch agrammatic speech production. These are data on: (1) verb movement; (2) object scrambling; and (3) verbs with alternating transitivity. We compared the data of our corpus research with the performance of agrammatic speakers on the constructions. The conclusion is that frequency cannot account for the data, but linguistic complexity can. It is then discussed what "linguistic complexity" exactly stands for, in terms of the word order deficit in agrammatic aphasia. (Contains 4 footnotes, 8 figures and 9 tables.)

Recent advances in multilingual language modeling have brought the idea of a truly universal parser closer to reality. However, such models are still not immune to the "curse of multilinguality": Cross-language interference and restrained model capacity remain major obstacles. To address this, we propose a novel language adaptation approach by introducing contextual language adapters to a multilingual parser. Contextual language adapters make it possible to learn adapters via language embeddings while sharing model parameters across languages based on contextual parameter generation. Moreover, our method allows for an easy but effective integration of existing linguistic typology features into the parsing model. Because not all typological features are available for every language, we further combine typological feature prediction with parsing in a multi-task model that achieves very competitive parsing performance without the need for an external prediction system for missing features. The resulting parser, UDapter, can be used for dependency parsing as well as sequence labeling tasks such as POS tagging, morphological tagging, and NER. In dependency parsing, it outperforms strong monolingual and multilingual baselines on the majority of both high-resource and low-resource (zero-shot) languages, showing the success of the proposed adaptation approach. In sequence labeling tasks, our parser surpasses the baseline on high resource languages, and performs very competitively in a zero-shot setting. Our in-depth analyses show that adapter generation via typological features of languages is key to this success. [ABSTRACT FROM AUTHOR]

In 1989, Willem Kuiper published his thesis on the Middle Dutch romance Ferguut in which he concluded that the text was written by two authors. Kuiper showed differences in writing style at all levels and concluded this could not be a coincidence. According to Kuiper, the first author translated the Old French Fergus by Guillaume le Clerc, approximately until vs. 2592, whereafter the second author completed the second half without the French example - in the spirit of Fergus, but in his own words. Whereas Kuiper had to do his quantitative style analysis by hand, today the statistical programming language R in collaboration with the stylometric program Stylo can perform the job much faster and more thoroughly. In its analysis, the software not only takes all the differences into account (like Kuiper did), but all the similarities as well, even at levels writers and readers are hardly aware of, such as word frequency, word order and the use of function words. At this level every author leaves behind his personal fingerprint. Our stylometric research on the one hand confirmed that differences in style can be detected, but on the other hand shows how the even larger number of similarities suggests that the entire text was written by single author. The difference in style should most probably be explained as a shift in register by one author, from translating a foreign language text to freely retelling a story. [ABSTRACT FROM AUTHOR]

This paper investigates agreement mismatches in Dutch relatives. While the norm is that singular neuter nouns occur with the relative pronoun dat ‘that’, it is by now quite common to find neuter nouns combining with the relative pronoun die. A large Twitter corpus is used to study which linguistic variables make die ‘that’ in this context more likely. Lack of agreement between neuter noun and relative pronoun is very frequent in this corpus (37.5% of the cases, 46.8% if the preceding determiner is indefinite). Non-agreement is most common for nouns that are high in the animacy ranking, but it also occurs with other semantic classes, and there is quite a bit of lexical variation. Young, female users have a stronger tendency to use non-agreeing relative pronouns. Contrary to what previous work suggests, we do not find that users with a Moroccan or Turkish background have a stronger tendency towards non-agreement. A comparison of tweets with agreeing and non-agreeing pronouns and a comparison of the Twitter corpus with web data both suggest that non-agreement is characteristic of informal language use. [ABSTRACT FROM AUTHOR]

In this study, we investigate crosslinguistic patterns in the alternation between UM, a hesitation marker consisting of a neutral vowel followed by a final labial nasal, and UH, a hesitation marker consisting of a neutral vowel in an open syllable. Based on a quantitative analysis of a range of spoken and written corpora, we identify clear and consistent patterns of change in the use of these forms in various Germanic languages (English, Dutch, German, Norwegian, Danish, Faroese) and dialects (American English, British English), with the use of UM increasing over time relative to the use of UH. We also find that this pattern of change is generally led by women and more educated speakers. Finally, we propose a series of possible explanations for this surprising change in hesitation marker usage that is currently taking place across Germanic languages. [ABSTRACT FROM AUTHOR]

We present a wide-coverage computational parser and grammar for Dutch. The grammar is developed within the tradition of Head-driven Phrase Structure Grammar and provides detailed and linguistically motivated accounts for most syntactic phenomena in Dutch. A broad overview of the grammar and lexicon is given, as well as a more detailed discussion of the analysis of comparatives and partitive constructions. Parsing and disambiguation uses statistical information, which is derived in part from a syntactically annotated treebank. The accuracy and coverage of the grammar is evaluated on representative portions of the tree-bank and a number of problematic constructions for the current grammar are identified. [ABSTRACT FROM AUTHOR]