A study of transfer between isomorphs of the Tower of Hanoi problem

The study concerns transfer between Tower of Hanoi (TH) isomorphs. The trai ning problem was a Move problem and the transfer problems were three Change of size problems varying according to the type of change evoked in the ins tructions. There was no transfer effect on the first trial of the transfer problem. A positive transfer effect only developed progressively and was ma nifest on the third trial. Verbal reports produced after solution show that there was no simple relation between explicit understanding and (1) the pr oblem just solved and (2) performance in the transfer problem. It is sugges ted that the major progress in learning concerns goal ordering. Subgoal ord ering learned in the training problem cannot be transferred immediately to the new problem, perhaps because to do so would involve a very abstract und erstanding of the structural role of each dimension, which is beyond the re ach of participants solving a single problem for a few trials. The appropri ate ordering has to be relearned in the new problem, something which requir es time and may explain that transfer is based more on implicit learning an d less on explicit learning.