Non-classical science gives a very specific answer to the question of scientific errors and their epistemological value. But for all the specificity of this answer, it casts light on a problem that remains with us century after century, the historically constant problem of truth and error—one of the most fundamental problems of knowledge. At first sight, these two poles have always stood opposite each other, like good and evil, beauty and ugliness. But moral and aesthetic theories have long since left behind this initial conception, and shown that the two polar concepts are in fact inseparable. As regards truth and error, their indivisibility has only become apparent in the context of non-classical science. Truth has ceased to be the absolute contradiction of error. Contemporary science finds it to be something relative, inseparable from its opposite pole. But the non-classical, retrospective approach, a re-evaluation of the past of science in the light of its present, and indeed even more in the light of its prognosis, its future, allows one to see that scientific truths that are adequate for the object of knowledge always reflect only one side of the objective world, a world which has an infinite number of sides. The irreversible process that leads from relative error to relative truth has an absolute character: an irreversible process of generalization, concretization and complication of a Weltanschauung that more accurately reflects the infinite complexity of life.