In Brief

Psychologists report a previously unobserved counterintuitive effect involving how people produce words in this month's Journal of Experimental Psychology: General (Vol. 132, No. 2).

In the study, Michele Miozzo, PhD, of Columbia University, and Alfonso Caramazza, PhD, of Harvard University's Cognitive Neuropsychology Laboratory, tease out the interplay of two well-known phenomena:

  • People are slower to say the name of a pictured object if a to-be-ignored word is superimposed over the picture.

  • People read familiar words such as "house" faster than rare words such as "fern."

The researchers examined whether common and rare words superimposed over the pictures of line-drawn objects had different effects on how long it took for a total of 185 participants to name the object pictured.

From the researching findings above, many psychologists theorized that common words superimposed over pictures would interfere with naming the picture more than rare words. They conjectured that, because it would take longer to recognize the superimposed rare word, people would process the picture first--but that common words, which are read faster, would be processed at about the same rate and therefore would interfere more with naming the picture.

However, Miozzo and Caramazza found the opposite; participants took longer to name pictures superimposed with rare words--a finding they termed the "distractor frequency effect."

In follow-up experiments, they found that the distractor frequency effect remained when they tried to hinder participants' recognition of the distractor words by presenting the words in alternating upper and lower case; for example, hAnD. There was also no effect when they tried to facilitate word recognition by visually presenting the distractor words before the picture-naming task began.

However, participants responded faster to the rare words when they read the distractor words aloud several times before performing the picture-naming task.

In other words, simply reading the distractor words was not enough; it wasn't until participants spoke the distractor words that they reduced the time they took to name the pictured objects.

In three additional experiments, Miozzo and Caramazza paired pictures with words that were either semantically related, such as a picture of a bench with the word "chair," or phonologically related, such a picture of a ball with the word "wall."

They found that the distractor frequency effect remained during the semantically related-word trial. However, the effect disappeared with the phonologically related pairs; in that trial, participants took longer to name objects superimposed with common words than rare words.

That finding--coupled with the finding that speaking distractor words aloud shortens the time to name the object--indicates that the interference effect arises at a particular level of lexical processing, the authors say.

Specifically, they theorize that the interference occurs at the point when people start to name a word aloud, rather than when they're simply processing the information by, for example, looking at a picture of a dog.

Miozzo and Caramazza say their findings are at odds with current explanations of the picture-word paradigm--none of which predict the distractor frequency effect.

"These data are going to force necessarily the reconsideration of a large literature on lexical access," says Caramazza. "Our results show that one can't use the picture-word interference paradigm to make claims about lexical access in the way that people have done up to now."

But perhaps just as interesting to Caramazza is the fact that he and others studied the picture-word paradigm for so long without discovering the effect.

"This shows the power of theoretical blinders," he explains. "How when we become convinced of a particular idea--in this case, the theoretical expectation that high-frequency words should be more potent distractors than low-frequency words--we tend to ignore certain things."

--D. SMITH