It's a Jungle in There: How Competition and Cooperation in the Brain Shape the Mind by David A. Rosenbaum Page B

inclination to say “green” when seeing “Green” or the powerful urge to say “red” when seeing “Red,” the effect washes out. From this it follows that susceptibility to the Stroop effect indexes reading skill.
    Given the link between the Stroop effect and skill level, other cognitive psychologists have used Stroop or Stroop-like interference to investigate the development of automaticity. Such an approach was taken by Daniel Reisberg, Jonathan Baron, and Deborah Kemler, working together at the University of Pennsylvania. 29 They showed participants strings of digits and asked the participants (university students) to call out the number of digits in each display. Seeing “2 2 2” required “3,” seeing “4 4” required “2,” and so on. After a little practice, the participants quickly and accurately reported the number of numerals they saw rather than the names of the digits to which they were exposed.
    Inviting people to report the number of digits they see provides a convenient medium for studying the development of automatic responding. Initially, when you’re shown “2 2 2” you count the number of numbers. You might go so far as to say to yourself, “Let me count those digits: 1, 2, 3…OK, there are 3 digits, so I’ll call out ‘3.’” Soon you don’t need to count the digits. You just rely on a shortcut: “Whenever I see ‘2,’ I’ll call out ‘3.’” What began as a deliberate process turns into a prepackaged rule for relating a stimulus to a response—what cognitive psychologists call a
production
. 30
    A production is a condition-action instruction. Its canonical form is “If X, then Y,” where X specifies a
condition
, such as “If the stimulus is 2…,” and Y specifies an
action
, such as “…say 3.” You can store many productions. The strengths of the productions can vary.
    At first, when you participate in an experiment like the one run by Reisberg, Baron, and Kemler, you have a strong production at the ready: “If I see 2, then say ‘2.’” But after saying “3” over and over again in response to the sight of “2,” a new production takes hold: “If I see 2, then say ‘3.’” Now you’ve got
two
productions that share conditions but dictate different actions. Because you can’t say “2” and “3” at the same time, you have to make up your mind which response to produce. Deciding between the responses constitutes the moment of truth for the attention system, the moment when attention—whatever it is and whatever it does—gets you to do one thing rather than another. When multiple productions are activated, the sorting-out process gets difficult, but as a production is called for more often, it gets stronger than its competitors.
    Such strengthening and weakening of productions provides a way of accounting for the data of Reisberg, Baron, and Kemler, who found that reaction times for correct completions of required productions decreased the more often the productions were called for. If subjects saw “2 2 2” several times with few intervening items—and, especially, with
no
intervening stimuli that had different numbers of 2s in them, such as “2” or “2 2” or “2 2 2 2”—the participants got faster and faster at saying “3” to the trio of deuces. But if the composition of the stimuli changed, so deuces were shown as singletons rather than as trios (“2”…“2”…“2”) and so the necessary response was “1,” the reaction times were long relative to what they had been earlier.
    Such patterns of data are easily explained with the jungle principle. You simply need to appeal to the notion that just as the fortunes of creatures rise and fall depending on how well they fit their environments, so do the fortunes of productions within the brain.
More on Automaticity
    When stimuli are responded to automatically, they’re responded to with considerable speed, with little need for attention, and with little chance of slowing