The purpose of this article is to introduce Tulving's hypothesis on encoding process of memory and to make some comments on the hypothesis. Tulving et al. have conducted a series of experiments on the effect of extralist cue, and have obtained a few phenomena which could not be explained by the existing theories. For instance, recall was sometimes higher than recognition, and strong extralist cue could not necessarily facilitate recall. These results are contradictory to "generation-recognition" theory which was most popular explanation of cueing effect. Tulving proposed "encoding specificity" hypothesis to explain these phenomena. In this articale "enconding specificity" hypothesis is discussed in six sections. In 1 Tulving's studies on extralist cue effect are summerizied. The results of this kind of studies lead the distinction between "availability" and "accessibility of memory. In 2 several types of encoding and decoding processes are illustrated with simple examples. The relation between two processes entails that retrieval cue can facilitate recall only if the encoding and the cue are semanticall consistent in subjects' verbal system. This viewpoint is background of Tulving's hypothesis. In 3 Tulving's "encoding specificity" hypothesis is depicted in detail and his experiments which verified the hypothesis are shown. In 4, 5, 6 affirmative and negative responses to "encoding specificity" hypothesis are cited. Specially the controvesy with Santa et al. is reviewed minutely. Lastly in 7 author's impression and opinions concerning Tulving's hypothesis, etc. are described breifly.