CLEFS

Cross-linguistic experiments on fragmentary sentences

01/10/2023 to 01/10/2026

Linguistic fragments, such as ‘I think himself’ as a response to the question ‘Who shaved Hans?’, are seemingly nonsentential utterances that nevertheless convey full-fledged sentential meaning. The prevailing view in the 2000s-2010s was that the most conspicuous formal restrictions on fragments are grammatical in nature. In recent years, results obtained via corpus and experimental judgment studies have shown that the empirical foundations underlying the prevailing view – which come from informal data collection methods – are an idealization, and that the true empirical picture is more complicated than previously thought. This project's starting interpretation of these results is that the core insight of the prevailing view was partly correct, but that the grammar asserts its influence indirectly, by limiting the discourse contexts from which a fragment’s meaning can be retrieved. This project tests various predictions that arise from this interpretation. Specifically, this project will conduct 26 formally oriented acceptability judgment experiments on native speakers of English, German, and Turkish (~9 experiments on each). The project predicts specific patterns of variance in acceptability judgments when the following variables (among others) are modulated: (i) the clause-type of the fragment’s antecedent, (ii) whether the fragment shows P-omission, (iii) whether the language permits P-stranding (English=yes, German/Turkish=no), (iv) whether the fragment’s correlate is island-bound in the antecedent, (v) whether the language forms wh-questions via wh-fronting (English/German=yes, Turkish=no), and (vi) whether the fragment bears presentational or contrastive focus. Our experiments will test these variables (and others) in balanced and controlled web-hosted experiments whose participants will be crowdsourced wherever possible.