In an experiment reminiscent of French Connection's successful FCUK advertising campaign, psychologists in America have documented a new word illusion using what they call the "fast pairs" method.
Catherine Caldwell-Harris and Alison Morris have found that participants presented very briefly with familiar word pairs in the unfamiliar order, tend to report that the words appeared in their usual order (read the title to this post quickly enough and you might experience a similar effect).
For example, in an initial experiment, when 22 Boston University students were presented first with the word "day" for 45 milliseconds and then with the word "every" for the same length of time, and asked to report what they'd seen, they tended to get the order the wrong way around, believing that they'd seen the words in the order "every day" rather than "day every". By contrast, they made far fewer reversal errors when the words were actually presented in their usual order.
Analysis of the frequency with which word pairs occur in a given order, using a Google search, showed that the participants made the reversal mistake more often for pairs that have a greater tendency to appear in a given order.
Intriguingly, people don't seem to realise that they're making this error. In fact, a pilot study in which participants were asked to report their confidence in the word order had to be abandoned because they grew so irritated with having their confidence repeatedly questioned.
The researchers believe the technique provides a great example of how entrenched expectations can bias our perception in conditions where sensory information is weak. "This illusion is relevant to theories of top-down effects in recognising words in context, and suggests that familiar word pairs may come to have unitised storage," they said.
C CALDWELLHARRIS, A MORRIS (2008). Fast Pairs: A visual word recognition paradigm for measuring entrenchment, top-down effects, and subjective phenomenology☆ Consciousness and Cognition, 17 (4), 1063-1081 DOI: 10.1016/j.concog.2008.09.004
Post written by Christian Jarrett (@psych_writer) for the BPS Research Digest.