Description
Timing how long readers take on matched text versions that fit or overturn an evoked frame, to measure the processing cost of reinterpretation.
These paradigms compare reading times (or ERPs) for text versions that either fit or overturn the frame a reader has built. Coulson and Kutas took one-line jokes, removed the final word and normed the fragments on a cloze task; the top completion reveals the default frame, and its probability splits fragments into high (over 40 per cent) and low constraint. Each fragment receives a joke ending and a frame-congruent ending matched on cloze probability (about 3.6 per cent each). Self-paced reading showed longer final-word times for joke endings, but only under high constraint; the ERP version found larger N400s. A non-humorous frame-shift variant added about 100 ms on the shift word. Earlier text-level studies (Rumelhart; Garrod and Sanford's scenario titles; Pickering and Traxler's garden paths) timed matched story versions between groups, corrected for baseline speed. Matching on cloze is what makes the contrast rigorous.
Attributes
- Discipline
- Linguistics, Semiotics & Discourse
- Loss type
- Unmigrated
- Stage
- Design, Evaluation
- Data
- Quantitative
- Effort
- Weeks
Steps
- Collect items in which a single word or sentence forces reinterpretation of earlier content.
- Remove the critical word and norm the fragments on a cloze task with a separate sample of about 20-40 people.
- Split fragments into high and low constraint by the top completion's cloze probability (for example above 40 per cent).
- Write a reinterpretation ending and a frame-congruent ending matched on cloze probability.
- Present text word by word (self-paced, eye-tracked, or fixed rate with EEG) with comprehension probes, counterbalancing versions across lists.
- Correct for baseline reading speed and compare critical-word reading time or N400 by ending type and constraint.
- Analyse separately by whether the reinterpretation was understood, using probe accuracy.
Sources
- Semantic Leaps (Coulson, 2001)
Also known as
- Frame-shift processing-cost test
- Frame-shifting reading-time and ERP paradigm
- Joke versus non-joke ending contrast
- Coulson and Kutas frame-shift study
- Text-level reading-time studies of frame and scenario effects
- Scenario title paradigm
- Pet versus fur story
Categories
- Content
- Linguistics, Semiotics & Discourse