Linguistische Berichte Heft 219

This article proposes definitions of two key concepts in linguistics, dialect and language, and formulates these in analogy to the biological concepts population and species. This allows for a much more coherent way of defining these concepts. In addition, a three-level model of language consisting of (i) an innate faculty of language (FL), (ii) I-language, and (iii) utterances will be proposed. Some of the more radical consequences resulting from this approach concern standard languages which are considered as entirely different entities, and languages traditionally known as Chinese, German, Dutch etc. which are claimed not to exist from a strictly linguistic point of view.

In this paper, I compare finite and infinite complements embedded under predict ate of a class I call influence predicates, for instance befehlen (order), erlauben (allow), zwingen (force), ermöglichen (enable). They share the property that under some conditions, modal verbs in the complements are semantically redundant. I assume that their core meaning provides a causal operator and a modal operator modifying the embedded proposition. Empirical studies show that with regard to the redundancy of point at a fundamental difference between finite and infinite complements. Though both are propositional, only infinite complements are modally dependent from their matrix predicate while finite clauses are optionally semantically independent.

The German particle nämlich is puzzling because it seems to have two independent semantic functions which strictly correlate with specific syntactic environments: if nämlich precedes an ,,orphan constituent" (Haegeman 1991) it specifies an underspecified discourse referent in the previous clause, and if nämlich appears in a whole clause its function is marking that the hostclause delivers an explanation to the previous clause. A polysemy- or even homonymy-analysis seems problematic precisely because of this strict correlation between syntactic environment and semantic function. In this paper we propose a unified analysis of nämlich. We argue that nämlich marks the property of the context that there is an implicit question to which the host of nämlich delivers a direct (short) answer (Jacobson 2008). Crucially, constituents are good short answers to constituent-questions (Who?), while whole clauses are only good short-answers to ,,sentence"-questions like Why p? Building on these intuitions we show how both readings of nämlich can systematically be derived and implement our analysis formally.

There are three kinds of non-phoric es: the topologically motivated prefield-es, the formal subject-es and the correlate es which is coindexed with a subject or object subordinate clause. All of them can stand in the ,prefield' of German declarative clauses. The theory of the prefield ranking should disallow the positioning of such elements in the prefield. A closer look at corpus data reveals that sentences with non-phoric es in the prefield do not contain any of the more favoured prefield-fillers. The only way to fill the prefield in such cases is by Formal Movement of the subject (es) into the prefield position, or, if even that fails, by base-generation of an expletive in the prefield, namely the prefield-es.

This article presents the results of a recall experiment on Modern Greek experiential verbs. The influence of the factors subjecthood, thematic role (agent, experiencer), and animacy on word order and their interaction is investigated with three different types of experiencer verbs, namely experiencer subject (ES) verbs, labile [±agentive] experiencer object verbs, and non-agentive experiencer object (EO) verbs. The experimental results show that while a tendency to preserve the preferred SVO order is visible with all examined verb classes, this effect is weakened by an experiencer-first preference for EO verbs. Furthermore, for EO verbs the crucial factor for the argument order preferences is the property [±agentive] of the stimulus while animacy does not exhibit an independent effect in our findings. These results support the separation of the three types of experiential verbs, which is suggested in psych-verb theories on the basis of their different syntactic behavior.