Philosophical presuppositions in ‘computational thinking’—old wine in new bottles?

‘Computational thinking’ (CT) is highlighted in research literature, societal debates, and educational policies alike as being of prime significance in the 21st century. It is currently being introduced into K–12 (primary and secondary education) curricula around the world. However, there is no consensus on what exactly CT consists of, which skills it involves, and how it relates to programming. This article pinpoints four competing claims as to what constitutes the defining traits of CT. For each of the four claims, inherent philosophical presuppositions are identified concerning the nature of thinking, problem-solving, and human–computer relationships. It is argued that most of these philosophical presuppositions are very similar to ones that Hubert Dreyfus and Donald Schön addressed forty to fifty years ago. Therefore, claims about the power of CT raise old discussions in a new disguise. Dreyfus and Schön argued that the presuppositions were wrong. Do their arguments hold and thus defeat contemporary claims about CT? Alternatively, has the technological development since the time of their writings proven them wrong? This article argues that it is necessary to heed the arguments of Dreyfus, Schön, and—later—Bent Flyvbjerg to ensure that CT curricula are built in accord with the value-rational nature of human practice, rather than on misconceived rationalizations of problem-solving, computer use, and computational aspects in the world. However, the pervasive integration of information technology in today’s world raises new issues concerning human–machine collaborations that sidetrack the arguments of Dreyfus and Schön. A revised view of CT is required which focusses on articulating how humans can design for, partake in, and critically appraise human–machine collaborations.