Comp-trace effects: a comparative and psycholinguistic approach

• Applicant: Dr. Ankelien Schippers
• Subject Area: General and Comparative Linguistics, Experimental Linguistics, Typology, Non-European Languages
• Term: from 2019 to 2023
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 424443710

This project investigates the so-called COMP-trace effect. This refers to the fact that in English and several other languages, a complementizer (e.g. that, if, whether) cannot be spelled out when it immediately precedes an empty subject position. This happens, amongst others, when the subject is a question word which is fronted to a structurally higher clause: [Who do you think [that met Tom]] = ungrammatical, [Who do you think [met Tom] = grammatical. The COMP-trace effect is considered to be a universal effect, but has claimed to be absent in German and Dutch. Furthermore, up to this date, no satisfactory, crosslinguistic explanation for the COMP-trace effect has been given. There is evidence that the COMP-trace effect is not the result of a single factor (be it syntactic or non-syntactic), but of an accumulation of several factors, some of which are processing-related. The goal of this project is to uncover the factors that contribute to the COMP-trace effect, leading to a conclusive analysis of this effect. This will be done by comparing the closely related languages German, English and Dutch. One hypothesis that will be investigated is that German and Dutch show a COMP-trace effect as well, albeit in a milder form. A second hypothesis that will be tested is that the strength of the COMP-trace effect is dependent on processing-related factors, which in turn are dependent on typological properties, in particular case-marking and word order. These hypotheses will be investigated by combining experimentally gathered quantitative judgement data with online and offline processing measures, specifically reading time and comprehension data. With regards to English, a novel hypothesis will be tested, namely that the embedded clauses in questions like Who do you think [met Tom] is a type of pseudo-relative clause, of which the question word (who) is the head. One of the predictions of this analysis is that the question word should be able to surface as whom, under the assumption that it is a matrix clause argument. This will be tested by means of a grammaticality judgement task and a production task.

• DFG Programme: Research Grants