Unconscious Learning

Psychologists have evidence that learning starts as unconscious activity. This sounds like a contradiction of Mandler's idea that conscious control is required for learning something new, but not if an unconscious phase of learning occurs before the phase described by Mandler in which to-be-learned material is grasped in consciousness. The pre-conscious phase is called implicit learning (whether or not it is followed by conscious awareness).

What is implicit learning? How was implicit learning studied, using strings of letters?

An example comes from the work of Arthur Reber, who studied implicit learning over a period of decades. In a 1967 experiment, Reber showed subjects a set of letter strings (such as TSXS, TSSXXVPS, and PVV) generated using a hidden rule. Soon the subjects were able to judge whether new letter strings fit the rule or not. They could do this despite being unable to specify the rule. This is what people commonly call intuition, or knowledge from an unknown source. The subjects "knew" whether a new letter string fit the rule, but they did not know how they knew.

Reber carried out many such demonstrations. He used different procedures to rule out alternative explanations. In the end, he concluded that learning typically begins with unconscious processes. Brain-scanning research bears this out. For example, the anterior cingulate gyrus, an area crucial to executive control ("willpower") and planned activity, shows different responses to wins and losses in gambling before a person is conscious of them (Gehring & Wiloughby, 2002).

In what sense does it mimic our species history, in Reber's view?

Reber (1993) argued that each individual act of learning mimics our species history. "Consciousness is a late arrival on the evolutionary scene," he pointed out. "Sophisticated unconscious perceptual and cognitive functions preceded its emergence by a considerable margin." Similarly, in the individual act of learning, consciousness is a late arrival, following unconscious perceptual and cognitive functions that first detect a pattern.

How did a researcher study implicit learning with an X appearing in various parts of a computer screen?

Here is a typical experiment that supports Reber's theory of implicit learning. It comes from Dr. Pawel Lewicki of the University of Tulsa. He had volunteers try to predict where an X would appear on a computer screen, selecting one of four quadrants. The subjects pushed a button corresponding to the quarter of the screen where they predicted the X would appear next. The X followed a pattern determined by 10 simultaneous rules.

Lewicki offered $100 to anybody who could report the rules (after the experiment was over) but nobody could specify them. However, the volunteers became more and more successful with their predictions as the experiment went on. They sensed the pattern, whatever it was. Their predictions became more accurate until Lewicki suspended the rules and moved the X randomly, whereupon their performance dropped to pre-learning levels again (Goleman, 1992).

How did brain scans change as people practiced a simple motor skill?

At some point a person may grasp a pattern or make it conscious. This process can be traced in brain scans. Pascual-Leone, Grafman, and Hallett (1994) used a technique called transcranial magnetic stimulation (TMS) to study this. They used a motor (movement) task and looked for changes in the motor cortex as subjects practiced.

The transition from unconscious knowledge to conscious knowledge and then automaticity showed up as a progression of changes in the brain scans. Initially, while subjects tried to figure out what they were supposed to do, cortical areas devoted to the task grew larger. The enlargement of these "output maps" increased until subjects achieved explicit knowledge of the task, becoming conscious of the pattern. After this, their reactions became more automatic, and the areas of brain activity shrank so that only a smaller area of cortex was active.

What similar pattern was observed when subjects were asked to respond to nouns with associated verbs?

Raichle (1994) reported a very similar finding. Subjects presented with nouns had to respond with an associated verb, as rapidly as possible. For example, given the noun baseball the subject might respond, "hit." Given the noun money the subject might respond, "buy."

Nouns were presented every second and a half. At first, the subjects hesitated while they thought of an appropriate verb. After less than 15 minutes of practice, the task became easier (they "caught on"). After the subjects caught on, the areas of activity in the brain shrank. Finally the brain responses resembled the patterns produced by reading single words, a process which is automatic for most adults adults and involves only a small area of brain tissue.

The pattern revealed by this research seems sensible. Initially, as the brain grapples with a new problem, large areas are involved. Many neurons have a chance to nominate themselves for inclusion in the problem-solving activity. Some of them prove necessary for solving the task, others do not. Once the problem is solved, or the task becomes familiar, a smaller group of neurons—those most essential to the activity—take over responsibility. After this, automaticity is possible. The rest of the mind can wander to something else, while skilled performance continues among a select group of neurons.


Write to Dr. Dewey at psywww@gmail.com.

Don't see what you need? Psych Web has over 1,000 pages, so it may be elsewhere on the site. Do a site-specific Google search using the box below.

Custom Search

Copyright © 2007 Russ Dewey