This paper presents the results of a pilot study concerned with regional variation in the acceptability of long distance movement ('Satzverschränkung') in German. We show that the likelihood of the production of such constructions decreases from South to North. Places with a Bavarian dialect background show a particularly high productivity of long distance movement, while places with a Saxonian dialect background turned out to be very restrictive. Apart from the latter exception, the relative strength of a dialectal background in a given area seems to correlate well with the acceptability of long movement.

Production data from native speakers of Swiss German, Standard German and Dutch show that they employ different means to realize phrase-medial pretonic stop contrasts. Swiss German uses a quantity contrast, whereas Standard German and Dutch employ different laryngeal contrasts. Two perception experiments reveal (i) the location of the respective categorical boundaries for native listeners, (ii) the difficulty to distinguish the Swiss German quantity contrast for non-native listeners, and (iii) the ease to distinguish the German laryngeal contrast for Swiss German but not for Dutch listeners. We find that a native length contrast does not inhibit categorical perception of a laryngeal contrast whereas the presence of prevoicing hinders the perception of a deviant laryngeal contrast (long vs. short voicing lag).

Empirically based research on second language acquisition (further L2A) has proven to be full of possibilities for delving further into the acquisition of functional L2 categories. The current paper discusses general considerations concerning the extent to which the second language learner (L2er) is influenced by the first language (L1) grammar, the transfer – or the rise – of L2 functional properties, and the phenomenon of optionality in interlanguage grammars (ILGs), particularly focusing on the nature of linguistic knowledge that is available at different stages beyond the commencement of L2A. Drawing on the data from various levels of L2A, I have found ample empirical support for the claim that the reparametrisation of [+strong] L1 to [–strong] L2 features is a recalcitrant task to accomplish even in advanced ILGs, whereas the V2-parameter (motivated by verb-raising which is also available in the L1) is accurately acquired. Thus, the data obtained on the development of CP shell and expletives in TP has turned out to be consistent with my assumption that there is a kind of impairment in the syntactic module of L2 faculty, specifically associated with L2 uninterpretable features.

We present the rationale and the procedures of a corpus-based study designed to extend the 'interactive alignment' theory of conversation (which posits that automatic convergence of representations within and between interlocutors captures a significant aspect of dialogue structure) to mediated communication. The corpus is multimodal in that natural dialogue is supplemented by interlocutors' eye movement data plus audio and video footage. Findings from ongoing analyses suggest that local alignment (as indicated by unsolicited repetitions, collaborative completions, joint attention, etc.) is a prominent feature of dialogue tactics while global alignment (as indicated by long-term consolidation of discourse structure) merely is to be regarded as one among several dialogue strategies.