The paper addresses the question of how the unity of science can adequately be characterized. A mere classification of scientific fields and disciplines does not express the unity of science unless it is supplemented with a perspective that establishes a systematic coherence among the different branches of science. Four ideas of this kind are discussed. Namely, the unity of scientific language, of scientific laws, of scientific method and of science as a practical‐operational enterprise. Whereas reference to the unity of scientific language and of scientific laws does not provide a viable basis for the unity of science, the methodological and practical unity might. The unity of science can be characterized by the way in which methodological criteria enter into the assessment or evaluation of theories, and, moreover, by a transdisciplin‐ary approach to problems. Accordingly, the unity of science is not expressed by theoretical uniformity but by the unity of scientific practice.