The sentence 'What did you believe the claim that she bought?' is ungrammatical. According to subjacency, why?
AThe verb 'believe' cannot take a wh-complement under any syntactic conditions
BWh-movement is never permitted when the gap is located more than one clause away from the landing site
CExtracting 'what' requires crossing two bounding nodes—the NP containing the relative clause and the higher S—in a single derivational step, violating subjacency
DThe object position of 'bought' is not a valid extraction site because relative clauses have no specifier position
Subjacency prohibits any single movement step from crossing more than one bounding node (typically NP and S/IP). 'The claim that she bought ___' is a complex NP—the relative clause is embedded inside an NP, itself inside a higher clause. Extracting 'what' from the object position of 'bought' would require crossing both the NP boundary and the S boundary in one step: a subjacency violation. Compare the grammatical 'What did you believe she bought?' where there is no complex NP and movement can proceed via successive-cyclic steps through intermediate positions.
Question 2 Multiple Choice
Which statement best characterizes island constraints across human languages?
AIsland constraints are categorical and equally strong in all languages, confirming that they are a universal syntactic primitive
BIsland constraints apply only in languages that front wh-words to sentence-initial position
CIsland sensitivity varies cross-linguistically and shows gradient acceptability within languages, suggesting the constraints may be violable principles rather than absolute rules
DAll languages obey the coordinate structure constraint but differ only in whether they observe complex NP islands
Cross-linguistic variation is one of the central challenges for purely syntactic accounts of islands. Italian and Spanish permit some extractions from relative clauses that are blocked in English. Within languages, experimental work shows that acceptability judgments are gradient—not binary—and vary with context and processing load. These gradations are hard to reconcile with islands as inviolable syntactic prohibitions. They have motivated Optimality-theoretic accounts (violable ranked constraints) and processing-based explanations (island effects as working-memory costs), neither of which treat islands as universal categorical rules.
Question 3 True / False
According to subjacency, extraction from a relative clause is blocked because the moved element would need to cross two bounding nodes—the NP containing the relative clause and the higher S—in a single derivational step.
TTrue
FFalse
Answer: True
This is correct and is the core of the subjacency account of relative clause islands. A relative clause is embedded inside an NP (as in 'the man [NP who bought ___]'), and that NP is inside a higher sentence (S/IP). Movement out of the relative clause in one step crosses both the NP boundary and the S boundary simultaneously. Subjacency allows crossing at most one bounding node per step. Successive-cyclic movement can avoid violations in other contexts but cannot rescue movement from inside an embedded NP—there is no intermediate landing site within the NP that would allow incremental escape.
Question 4 True / False
Ross's (1967) empirical island generalizations are fully and adequately explained by Chomsky's Subjacency Condition, leaving no residual counterexamples or competing theoretical accounts.
TTrue
FFalse
Answer: False
False. Subjacency provided an elegant unified account of several island types but faced immediate empirical problems: cross-linguistic variation in island sensitivity (Italian allows extractions that English blocks); gradient rather than categorical acceptability within languages; and difficulty specifying which nodes count as bounding nodes across languages. These failures motivated successor theories: the Phase Impenetrability Condition in Minimalism, violable-constraint accounts in Optimality Theory, and processing-based explanations that attribute island effects partly to working memory costs rather than syntactic prohibition. The empirical landscape remains actively theorized.
Question 5 Short Answer
What is the core empirical phenomenon that island constraints describe, and what evidence challenges the view that island sensitivity is purely a syntactic phenomenon?
Think about your answer, then reveal below.
Model answer: Island constraints describe the observation that syntactic movement (wh-movement, topicalization, etc.) is systematically blocked from certain structural domains—relative clauses, coordinate structures, adjuncts—even when the same type of movement is grammatical elsewhere. Evidence against a purely syntactic account includes: (1) cross-linguistic variation (Italian allows relative clause extractions that English blocks); (2) gradient acceptability judgments within languages (island violations are not uniformly impossible but vary in severity); (3) processing accounts showing that many 'island effects' correlate with working memory demands of maintaining long-distance dependencies through complex structures.
The key insight is that subjacency sought to derive multiple distinct island types from one formal principle—an elegant explanatory strategy. But the empirical departures from categorical ungrammaticality across languages and contexts reveal that 'island sensitivity' is not a single, purely syntactic phenomenon. Modern theories treat it as a cluster of effects with both syntactic and non-syntactic contributors, making islands a revealing window into how grammar and processing interact.