Can People Learn without Realizing It?
As you enter the laboratory reception area, you are met by an experimenter, taken to a small room, and told that you will be taking part in a study of human memory. Next you’re told that you’ll have seven minutes to examine the following strings of letters and that you should simply try to ‘learn and remember as much as you can.’ Ready, begin:
PVV TSSSXXVV TSXXTVPS PVPXTVPS TSSXXVV
TSXS PTVPXVV TXXTVPS TXXTTTVV PVPXXVV
TSSXXVPS TXXVPXVV PTVPS PTTTVPS PTVPXV
You look at these meaningless items one by one, trying to memorize. P-V-V. T-S-X-S. You just keep repeating them under your breath. After seven minutes, the experimenter stops you. Time’s up. You can’t wait to write down what you remember before you lose it – but the
experimenter has other plans. He or she now reveals that the items were formed according to a set of rules, an ‘artificial grammar’ if you will, and that he or she is interested in whether you know what the rules are. As the experimenter explains the test you’re about to take, you get a
sinking feeling in the pit of your stomach. ‘Uh-oh. There were rules? I wasn’t looking for
The test is straightforward. On slides, you’ll see one hundred new items made up of the same
letters as before, one item per slide. Half the items will be grammatical (according to the rules); half will not. For each one, you are to press a button marked YES if you think the item is grammatical or NO if you think it’s not. You should also rate your confidence in each judgment
on a scale marked from 1 to 5. Oh, one more thing: Your responses will be timed.
The slide projector is turned on, the overhead lights are shut off, and you’ve got your fingers on the buttons ready to fire. First item: PTTTVPVS. It looks okay, as it contains the usual letters, and all. But are these letters ordered in a way that fits the grammar? ‘If I don’t know,’ you ask, ‘should I just guess?’ Instructed to respond to every item, you press a button and state your
confidence. The next one is PVTW. Same routine. PVPS. SXPTXTW. SXXVPS. Sometimes you answer quickly; at other times, you stare at the screen for a while before making a response. By the hundredth slide, you’re ready for a cold drink and a nap.
Make a Prediction
How well do you think people fare in this task? Having been in this experiment myself, I can tell you that many subjects shrug their shoulders in confusion. Based on what you’ve read, what do you think is the average test score? If all you did was guess, you would get roughly a 50% accuracy rate. If you came up with rules that were wrong, your score would be lower. If you knew the right rules, you would do better. So, what is your estimate, 10%? 30? 55? 90? Circle your prediction:
0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100
For thirty years, cognitive psychologist Arthur Reber has used experiments like this one to study ‘implicit learning’ – the tendency for people to acquire complex, abstract concepts without awareness or intention. Consistently, Reber has found that subjects cannot describe the grammar that they use to form the letter strings, nor can they explain the reasons for their YES and NO judgments. Yet in the study just described, subjects made the correct response 77% of the time –
and usually did so quickly.
What does it all mean?
Implicit learning, which occurs without our awareness, is a primitive but powerful form of adaptation. Indeed, people learn this grammar not by actively searching for rules or by receiving explicit instruction but simply through exposure to properly formed letter strings. Without really trying, subjects learn the grammar the way we learn to speak in our native tongue, figure out how to behave properly in a new setting, or ‘calculate’ the trajectory of a ball in fight in order to make the catch. Implicit mental processes such as these are common. To learn from experience, as these subjects did in this study, people must be attuned to associations between stimuli in their environment and between their behavior and its consequences.