Read about the latest research published in Psychological Science:
Still Suspicious: The Suspicious-Coincidence Effect Revisited
Molly L. Lewis and Michael C. Frank
Imagine hearing someone call a particular Dalmatian a “dax.” You will probably generalize this novel word to all dogs (the basic level). But if you hear “dax” applied to three identical dalmatians, you will probably generalize the word to other dalmatians (the subordinate level). Researchers label this latter generalization the suspicious-coincidence effect, in which learners generalize novel words more narrowly if they encounter several identical examples of that word rather than just one. To test the boundaries of this effect, Lewis and Frank presented pictures of exemplars associated with novel words (e.g., “dax”) and asked participants to classify the words as either subordinate or basic. Across 12 experiments, they varied whether exemplars were presented simultaneously or sequentially, whether participants were first exposed to one exemplar or three exemplars, whether the sequences of one or three exemplars were blocked by category or randomly, and whether the names of the one and three exemplars were the same. Results indicated that people tend to be suspicious when they encounter several items defined by the same word, assuming that it is not a coincidence, and that the word must represent a basic level of the category. Thus, probabilistic reasoning seems to influence learners’ ability to make inferences on the basis of sparse linguistic data
You That Read Wrong Again! A Transposed-Word Effect in Grammaticality Judgments
Jonathan Mirault, Joshua Snell, and Jonathan Grainger
People take longer to recognize a pseudoword when it is formed by transposing letters in a real word (e.g., “gadren” from garden) than when it is formed by replacing letters in the same word (e.g., “gatsen” from garden). This effect has been seen as evidence for flexibility in the way the position of letters in a word is encoded, which might inform how skilled readers process words. Mirault et al. investigated whether such positional flexibility also characterizes how people encode words in a sentence. They presented sentences that were grammatically incorrect because two words taken from a correct sentence were transposed (e.g., “The white was cat big” from “The white cat was big.”) or because two words taken from an incorrect sentence were transposed (e.g. “The white was cat slowly” from “The white cat was slowly.”), and asked participants to rapidly judge whether each sentence was grammatically correct. Participants took longer to decide and made more mistakes when the grammatically incorrect sentences derived from a correct sentence than when they derived from incorrect sentences. These results suggest that words might be processed in parallel rather than sequentially in a sentence, showing a flexible encoding of word-order information while reading. Future studies might help clarify the mechanisms used to represent word order in written-sentence comprehension.