The fingers don’t lie
Typing experiments show brain has at least two copy editors
The brain uses two different checks to guard against sloppy copy, a new study finds. By using a doctored word processor to sneak errors into typed words and surreptitiously fix typists’ real errors, researchers teased apart the various ways people catch their own mistakes. The study, published in the Oct. 29 Science, highlights the complexity of performance monitoring.
Psychologist Gordon Logan and his colleague Matthew Crump of Vanderbilt University in Nashville recruited skilled typists — people who typed more than 40 words a minute using all of their fingers. These subjects were able to type a paragraph about the merits of border collies with over 90 percent accuracy.
As the typists pecked away, researchers introduced common typing errors into about 6 percent of the words that appeared on a screen (changing sweat to swaet, swerat or swet, for instance). The program also corrected about 45 percent of the typists’ true errors.
In questionnaires after the typing test, subjects by and large took the blame for the introduced errors and took credit for the researchers’ corrections. No matter what he actually typed, when the typist saw that the word on the screen matched the word he had intended to type, he assessed his own performance as accurate.
But the speed of the typists’ keystrokes revealed something else. After hitting the wrong key, a typist’s fingers slowed down for the next keystroke, even if the researchers sneakily fixed the error so that the typist didn’t notice it. In these cases, a typist wasn’t explicitly aware of the mistake, but the brain’s motor signal changed nevertheless.
Logan says that this change in timing reflects a kind of automatic assessment of performance. “The body is doing one thing and the mind is doing another,” he says. “What we found was that the fingers knew the truth.”
Many psychologists thought that the mind was capable of detecting errors in several ways, but “nobody had pinned it down,” says cognitive neuroscientist Jonathan Cohen of Princeton University. “Here, they developed a very clever set of experiments to tease the types of system apart.”
The results may reveal a hierarchical method of error correction — with a “lower” system doing the actual work and a “higher” system assigning credit and blame, Logan suggests. These multiple layers of control may be evident in tasks such as playing music, speaking and walking to a destination, Logan says. As a man heads toward a new restaurant, his brain is noticing landmarks and keeping on the right course. Meanwhile, his feet steadily plod along, navigating the terrain automatically.
Whether the two types of error-catching systems operate in tandem or one is subservient to the other isn’t yet clear, Cohen says. Having the automatic, fingers-level system answer to the higher system “has a sort of intuitive appeal,” Cohen says, “but whether that idea is a convenience until we have a better idea or it’s true remains to be seen.”