The problem of transfer of learning has been historically of great concern to educators.
For them, it constitutes the very important practical question of how the school curricula should be arranged to ensure maximum positive transfer. Does learning algebra help in the learning of geometry? Which of the sciences should be taught first to ensure maximum transfer to other science courses?
One of the earliest notions of transfer of learning prevalent among educators around the turn of the century maintained that the mind was composed of faculties that could be strengthened through exercise, much as individual’s muscles can be strengthened.
This notion, known as the doctrine of formal discipline, was advanced in support of keeping such studies as Latin and Greek in the high school curriculum. It was argued that the study of Latin, for example, trains a student’s powers of self- discipline, reasoning, and observation.
The doctrine of formal discipline has been largely discredited by experiments. Some transfer does take place but, it depends much less on formal mental training than on learning for specific purposes.
For example, the study of Latin does indeed improve the understanding of English words but, only those with Latin roots. It does not improve the understanding of words of Anglo-Saxon origin. And the extent to which improvement occurs depends upon the way the Latin is taught: the gain in English vocabulary is much greater when the course is taught with emphasis on word derivation than when taught by more conventional methods.
Learning to Learn:
A special example of transfer of training is a phenomenon that psychologists have labeled learning to learn. Subjects who learn successive lists of verbal materials over a period of days are able to increase the speed with which they learn subsequent lists. Positive transfer occurs even though the lists are not similar. The Subjects apparently learn-a technique or an approach to the task that facilitates their performance on later tasks of the same sort.
Another example of learning to learn is provided by an experiment in which monkeys are presented with a series of discrimination problems. For each problem the animal is shown two objects – for example, a red triangle and a green circle – and is reinforced with food if it selects the correct objects, which might be the red triangle.
Object position is alternated in a random order from trial to trial so that sometimes the triangle is on the right and sometimes on the left. The animal must learn to ignore positional cues in selecting the correct object. After the monkey has learned consistently to select the correct object, it is given a problem involving a different pair of objects.
In one experiment monkeys were given several hundred such problems, each problem using a completely new pair of stimulus objects. The learning curves for two problems that occurred early in the series of problems and two that occurred later are presented in.
All the curves begin at the 50 percent level, because, on the first trial of each problem the objects are new and the animal has to guess. By the 10th problem the monkeys are making only about 54 percent correct choices on the second trial. By the 300th problem they were responding at the 98 percent level on the second trial.
In the beginning the animals made little use of the information provided by the first trial but, by the 300th problem they utilized this information to obtain almost perfect performance, on the second trial. The monkeys learnt that, if the object they select on the first trial is rewarded they should pick the same object on the second trial regardless of its position; if the first trial choice is not rewarded they should select the other object on the second trial. The monkeys have learnt how to learn for this particular class of problems and can now proceed on the basis of insight as opposed to trial-and-error behavior.
Learning to learn has been extensively investigated not only with monkeys but, also with young children. The finding indicates that the phenomenon is general and involves a number of different factors. One factor may be learning to relax in the experimental situation; another, to ignore distracting noises and other irrelevant stimuli.
Most important is learning to identify he relevant cues in the situation; example, in the experiment described above the monkey learns that the important cue is the quality (shape or colour) of the object, not its position on the display board. In a sense this involves learning principle. And as we shall see, learning principles rather than specific responses constitutes one the chief ways in which learning transfers.
Transfer by mastering principles:
One factor that makes transfer possible is the appropriate application to new situations of principle learned in old situations. The Wright brothers applied the principles they learned in flying kites to building an airplane. Principles of reasoning learned in logic are equally applicable in mathematics. The following experiment demonstrates the advantage of learning principles.
Two groups of boys shot with rifles at a target submerged under water. Prior to the target practice the experimental group studied an explanation of the theory of refraction of light so that they understood the apparent displacement of objects viewed under water. The control group received no explanation.
The experimental group learned to hit the target in about the same number of trials as the control group; thus, learning to hit the target was not influenced by whether or not the subject had studied the theory of light refraction. However, after the boys had become proficient at hitting the target, the depth of the target was changed. Both groups showed positive transfer from the first to the second task but the experimental group evidenced the greatest amount of transfer.
Their knowledge of the principle of refraction enabled them to master the new task in significantly fewer trials than the control group (Hendrickson and schroeder, 1941).