Garden-path sentences like 'The horse raced past the barn fell' are initially misinterpreted because:
AEnglish grammar is ambiguous and incoherent
BReaders commit to an initial analysis (racing horses) that violates working memory constraints, then must reanalyze
CThe sentence is genuinely ungrammatical
DReaders are ungrammatical
Garden-path effects reveal interface constraints. The parser prefers the simpler, more frequent analysis (raced as main verb) even though the correct analysis (raced as participle in a reduced relative) is available. This preference reflects limits on working memory and incremental processing, not grammar alone.
Question 2 Multiple Choice
Why do right-branching structures (constituents attached to the right) appear more frequently in languages than left-branching structures?
ARight-branching is inherently more logical
BLeft-branching violates grammar
CRight-branching imposes lower working memory loads during incremental parsing; information comes in order needed
DLanguages arbitrarily prefer right-branching
Right-branching structure (where modifiers follow) allows incremental processing: the head is available early, and modifiers augment it. Left-branching requires holding unintegrated modifiers in memory until the head appears. Working memory constraints explain the cross-linguistic frequency pattern.
Question 3 Multiple Choice
Resumptive pronouns (pronouns in dependencies, as in 'The boy that I saw him') violate island constraints in formal syntax, yet they appear in many languages. Interface constraints explain this because:
AResumptive pronouns are grammatically correct in all languages
BResumptive pronouns reduce processing load in center-embedding; they're licensed by interface constraints even if they violate core grammar
CResumptive pronouns are ungrammatical everywhere
DFormal syntax is irrelevant
Deep center-embedding (the boy that [the girl that [I saw]] left) causes garden-paths. Resumptive pronouns ('the boy that I saw him') circumvent this by eliminating the gap. They may violate island constraints but satisfy interface constraints (keep dependencies shallow). Their presence reflects interface licensing.
Question 4 True / False
Understanding language-cognition interfaces explains linguistic universals only if we assume all languages have identical cognitive architecture.
TTrue
FFalse
Answer: False
Interfaces can explain universals (all languages prefer right-branching, all languages prefer active over passive) while allowing variation in how constraints are satisfied. Cognitive constraints are universal; languages differ in how they navigate those constraints.
Question 5 Short Answer
Explain how working memory constraints could explain why languages limit center-embedding to one or two levels, and what linguistic structures arise as a consequence.
Think about your answer, then reveal below.
Model answer: Working memory has limited capacity (roughly 4-7 items). Center-embedding (NP [VP [VP ...]]) requires keeping multiple incomplete dependencies in memory until integration. Nested embeddings exceed capacity, causing breakdown. Languages minimize center-embedding through extraposition, raising, and resumptive pronouns — structures that keep dependencies shallower. This interface constraint shapes grammar.
Interface constraints often explain linguistic structure. Why certain movements exist, why certain islands are universal, why certain structures appear or disappear — often reflects memory, attention, and processing constraints, not arbitrary grammar.