This study investigated whether a word-learning method in which learners infer the meaning of unknown words from the context, subsequently verify the meaning with the aid of a word list, and finally memorize the meaning ("meaning-inferred method") leads to better retention than one in which the meaning of unknown words is given in the form of a translation so that learners can immediately start memorizing ("meaning-given method"). Additionally, the learning effect of the various stages of the meaning-inferred method (inferring, verifying, and memorizing) was investigated. In all cases the amount of time invested was recorded. The most important findings were: (a) The meaning-inferred method leads to a similar level of retention as the meaning-given method, but the former is considerably more timeconsuming and therefore less efficient; and (b) each separate stage of the meaning-inferred method leads to retention, but the learning effect of memorizing is the greatest, and the learning effect of verifying is about the same as that of inferring.