Questions: Language Comprehension and Sentence Processing

Question 1

What do garden-path sentences (e.g., 'The horse raced past the barn fell') primarily demonstrate about sentence processing?

• A. That comprehension is a passive decoding process
• B. That syntax and semantics are processed in completely independent serial stages
• C. That initial syntactic parses are built rapidly and incrementally, and must sometimes be revised when later words disambiguate
• D. That readers process only the final word of a sentence before assigning structure

Answer: (C)

Garden-path sentences produce a characteristic 'garden path' effect: readers initially commit to a plausible parse (e.g., 'raced' as the main verb), then hit a word ('fell') that forces reanalysis. This shows parsing is incremental and committed, not wait-and-see — a hallmark of online sentence processing.

Question 2

Language comprehension is fundamentally a passive process: listeners decode a message that was fully encoded by the speaker.

• T. True
• F. False

Answer: False

Comprehension is constructive, not just receptive. Listeners make rapid predictions, use context and prior knowledge to go beyond what is literally said, and draw pragmatic inferences (e.g., understanding 'Can you pass the salt?' as a request, not a question about ability). The final representation often contains information not explicitly present in the signal.

Question 3

What does it mean to say that syntactic and semantic processing are 'interactive' rather than serial?

Model answer: Meaning and context influence grammatical analysis as it unfolds — syntax does not have to be fully resolved before semantics begins. For instance, semantic plausibility can bias which parse a reader initially commits to, and world knowledge can help resolve ambiguous grammatical structures before the disambiguating word appears.

Serial (modular) models predict that syntax delivers a fully parsed structure to semantics, which then interprets it. Interactive models allow information to flow in both directions simultaneously. Evidence from reading-time studies shows that semantically implausible continuations slow processing even before syntactic disambiguation, supporting interactivity.