No escape from the island: On extraction from complement wh-clauses in English

Abstract

In theoretical syntax, English complement wh-clause are considered syntactic islands which block extraction in an asymmetric way: Argument extraction is more acceptable than adjunct extraction. Though this pattern is often assumed to be universal, studies have shown that Danish (and other Mainland Scandinavian languages) may be exceptions. It has also been argued that the patterns of (un)acceptability are biased by expert intuitions. We present data from 100 native speakers of English which confi rms (i) that English complement wh-clauses are islands, (ii) that there is a (subtle) argumentadjunct asymmetry, and (iii) that this acceptability pattern is not due to participant bias. Together with earlier fi ndings on Danish, these results are compatible with an island account that relies on parametric variation in the possibility of CP-recursion. 1 We would like to thank Sten Vikner for many years of interesting discussions on comparative generative syntax, movement and islands, and on the nature of language in general. It has been our pleasure and privilege to have him fi rst as our teacher, then as our supervisor, and eventually as our colleague and friend. Many thanks to Hubert Haider for his constructive review and to the participants at the

The Island Is Still There: Experimental Evidence For The Inescapability Of Relative Clauses In English*

“…The target stimuli consisted of 48 sentences embedded in a total set of 140 sentences: 12 sets of four target types (quadruples) as illustrated in (6) below: baseline, wh ‐extraction, topic extraction, and an ungrammatical control (an anomalous version of wh ‐movement plus a superfluous bare NP object) (plus 10 fully grammatical fillers, 10 completely ungrammatical fillers, and 72 other sentences that form the basis for a different experiment, see Christensen & Nyvad 2019).…”

<em>Hvis</em>, <em>da</em> og <em>fordi</em> ekstraktion fra adverbielle ledsætninger var mulig

“…Ekstraktion fra adverbielle ledsaetninger forekommer således umiddelbart at vaere mere acceptable på engelsk end dansk, helt kontra vores forventning (Nyvad m.fl. 2023), da vi i tidligere studier har fundet, at engelsk er mere restriktiv end dansk, når det kommer til ekstraktion fra relativsaetninger ) og hv-spørgsmål (Christensen & Nyvad 2019). Så hvordan skal vi forklare denne uventede tvaersproglige forskel?…”

“…Observationerne af tilsyneladende grammatikalitet af eksemplerne i (3) er blevet genspejlet i eksperimentelle studier af saetningsknuder i dansk, der har vist, at det i nogen grad er muligt at ekstrahere ud af indlejrede spørgsmål (Christensen m.fl. 2013) og relativsaetninger , modsat hvad vi finder for engelsk (Christensen & Nyvad 2019. På baggrund af disse studier kunne man have en formod-ning om, at dansk er mindre restriktivt end engelsk, når det kommer til ekstraktion ud af indlejrede spørgsmål og relativsaetninger.…”