In a visual world eye-tracking experiment, participants look at a display of objects while hearing 'Pick up the can...' Their eyes move toward the candle before the word is completed. What does this demonstrate?
AComprehension is sequential — listeners wait for full words before processing meaning
BPrediction operates on partial phonological input, activating candidates before a word is identified
CListeners guess randomly based on visual salience, not linguistic prediction
DThe brain processes syntax before semantics in online comprehension
Anticipatory eye movements to phonologically consistent objects before a word is completed demonstrate that the comprehension system activates multiple candidates based on partial input — not just after word recognition is complete. This is a signature of predictive processing: the system is running ahead of the input, committing to likely interpretations before they are confirmed. Option A describes a sequential bottom-up model inconsistent with this finding.
Question 2 Multiple Choice
A perfectly grammatical sentence ends with a semantically unexpected word ('The detective examined the potatoes'). The N400 amplitude for 'potatoes' is larger than for 'evidence'. What does this pattern reveal?
AThe brain flags 'potatoes' as ungrammatical and generates an error signal
BThe brain had pre-activated semantically likely continuations; the unexpected word requires revising a prior prediction
CThe N400 reflects the number of syllables in a word, which is greater for 'potatoes'
DThe finding shows that syntactic and semantic processing are completely independent
The sentence is grammatically correct, so the N400 cannot be a grammaticality signal. The N400 scales with the violation of semantic expectation — the degree to which the word conflicts with what was pre-activated. The brain already committed to likely continuations (evidence, witness, clues) before 'potatoes' arrived. The large N400 reflects the cost of updating a prior prediction. This is the key evidence that comprehension involves continuous prediction, not reactive word-by-word analysis.
Question 3 True / False
The N400 is an most-or-very little signal that fires primarily when a word is semantically anomalous or ungrammatical.
TTrue
FFalse
Answer: False
The N400 is a graded signal, not binary. Its amplitude is inversely proportional to the predictability of a word in context — even highly predictable, perfectly acceptable words produce a smaller N400 than moderately unexpected but acceptable words. This grading shows that the brain is continuously tracking probability distributions over upcoming words, not just detecting errors. If it were binary, all acceptable words would produce no N400 at all.
Question 4 True / False
Language comprehension involves active prediction: the brain pre-activates likely upcoming words based on syntactic, semantic, and discourse constraints before those words arrive.
TTrue
FFalse
Answer: True
This is the central claim of the predictive processing account of language comprehension, supported by N400 amplitude patterns, visual world eye-tracking, and reading time data. The evidence consistently shows that unexpected words — even when grammatical — cost extra processing time and produce larger neural responses, demonstrating that predictions were made and must be revised. Comprehension is not merely reactive integration of each word as it arrives.
Question 5 Short Answer
How does the N400 EEG component provide evidence that language comprehension is predictive rather than purely reactive? What specific feature of the N400 response is key?
Think about your answer, then reveal below.
Model answer: The N400 amplitude is inversely graded by how predictable a word is in context — not just whether the word is semantically anomalous. Highly predictable words produce smaller N400s; moderately unexpected but acceptable words produce larger ones. If comprehension were purely reactive, all acceptable words should produce the same response. The grading shows the brain pre-activates candidates and shows less neural effort when predictions are confirmed, more when they must be revised.
The graded nature of the N400 is what distinguishes prediction from mere anomaly detection. An anomaly detector would fire only for violations; a predictor generates a continuous probability distribution and signals surprise proportional to how unexpected the input was. The N400's gradient across a range of predictability values is direct evidence for the latter model.