"Discovery learning" or "learning by discovery" is usually defined as a way of teaching where the students are expected to "discover" or draw out principles from the facts that are presented to them. As soon as the definition is spelled out in these terms, it becomes immediately clear that there is a paradox in the concept of discovery learning. For, apparently, in discovery learning the students are expected to learn something which they are not taught; if they are taught to discover the principles, then the principles are not what they discovered for themselves. The present paper attempts at elucidating, without any pretension of solving, the paradox inherent in the concept of discovery learning. The paradox becomes clearer and perhaps more pressing when we distinguish two types of content of teaching: facts and principles. Indeed, we usually apply the word "teaching" indiscriminately to facts and principles in everyday parlance; we talk of "teaching facts" and "teaching principles." However, it should be noted that the word "teaching" has entirely different meaning in the two cases. In relation to facts, teaching means "telling the students," whereas in the case of principles, it means "making the students see." This distinction is a reflection of, and is reflected by the corresponding difference in the status of facts and principles as content of instruction. The facts are outside the students, in the sense that they can be told to the students by the teacher, whereas the principles, if the students are said to have learned them, must be within the students, in the sense that the principles now constitute the "mind's eye" of the students. Then the question arises: Can principles be taught? Or, more specifically, what is there for the teacher to do in teaching principles, in addition to telling the students the facts? Three articles in The Concept of Education edited by R.S. Peters are dealing with more or less similar questions. Oakeshott, in his "Learning and Teaching," identifies two kinds of cultural heritage to be transmitted through education, i.e., information and judgment, which roughly correspond to facts and principles as distinguished above. He stresses, however, that information and judgment represent not so much the types of educational content as the• modes of transmission. Teaching, according to him, is a twofold activity of instructing information and imparting judgment, and correspondingly, learning is a twofold activity of acquiring information and coming to possess judgment. If judgment can be taught, it can be taught only obliquely in conjunction with instructing information, possibly by the model of the teacher. Ryle, in his "Teaching and Training," tackles in a rather straightforward way the question: Is it possible that the students learn principles untaught by the teacher? His answer is that the students learn general modus operandi (knowing-how) at the same time they learn specific information(knowing-that) . He then proposes that we shift our educational emphasis away from teaching-that and learning-that, and more to teaching-to and learning-to; more toward teaching "abilities, competences, and skills" and less toward teaching "propositions." Dearden, in his "Instruction and Learning by Discovery," warns us against the common sense notion that the two ways of teaching, "instruction" and "learning by discovery," are distinguished in terms of the presence or absence of actual teaching, and, by implication, that instruction is the only type of teaching. The so-called "pre-school model" and "abstraction ism" are criticised accordingly. A more adequate model for discovery learning is what Dearden calls "problem solving model," where the teacher is expected to actively guide learning experience "by the subtle use of language." According to Dearden, therefore, "discovery" in education is not to be too literally interpreted. Would the teachers be satisfied with the answers provided in the three articles? Probably not, for "obliquely imparting," "teaching modus operandi," and "guiding experience through the subtle use of language" may be short of giving specific guidelines for, the teachers. Dearden's very criticism of the pre-school model and the abstractionism reveals a most important epistemological point that the students must be equipped with necessary concepts before they can benefit from the teacher's "subtle use of language." This point reminds us of the Socrates' mythical explanation of "recollection" by the eternal transmutation of soul. A more down-to-earth version of this explanation may be found in Piaget's theory of intellectual development. In Piaget's theory, the precondition of discovery is ensured by the continuous and cumulative process of assimilation and accommodation. However it may be hastily added that the teacher's role in this process' of intellectual development is far from clearly defined.