The present study aimed at investigating the effect of verbal presentation of a ‘weak’ transformation rule on verbatim memory task performance. A weak rule means a rule that is incomprehensive (referring only to limited aspect(s) of information to be learned), ambiguous (allowing multiple correspondence), and/or having exceptions. Logically, even a weak rule can partly reduce the memory load of the learner. Three lists of Japanese-German word pairs were constructed (See TABLE 1). They have 6, 28 and 9 pairs, respectively. Over these three lists, there is specific correspondence in consonants between the German words and their English counterparts, but no such correspondence in vowels and endings. Regularities of consonant correspondence can be represented in terms of 6 rules (TABLE 3) with slight ambiguity and few exceptions. Four groups of university students were presented and tested on List 1 and then List 2, given 1 a _??_minute interpolated memory task followed by a retention test of List 2, and finally tested on unlearned List 3 (TABLE 2). Procedures were different among the groups only in timing of presenting and explaining the 6 rules. Group A was given them just prior to the first presentation of List 2, Group B between the second and third (final) presentation, and Group C after the final presentation. They were allowed to keep a copy of the rules for reference up to the retention test of List 2. Group D was not presented the rules. The findings were as follows: 1) Verbal presentation of the rules inhibited learning; Groups A and B were inferior to C and D in recall and recognition tests after the third presentation.(TABLE 4) 2) Giving rules just before testing improved performance; Group C excelled D. 3) Learning under the rules facilitates generalization; Groups A and B outperformed C and D in the test of List 3. 4) Performances before and after the interpolated memory task (including test) remained nearly unchanged in all groups. 5) Analysis of errors suggests that those ignorant of the rules tended to make errors non-conforming to them, but fewer errors in vowels and/or endings alone.(TABLE 5) 6) Findings 1) and 5) combined show the advantage gained by awareness of the rules was more or less outweighed by overconformity to them and neglect of non-rule aspects.