This paper is a contribution to that part of science studies known as ‘the cognitive study of science’. The general goal of such studies is to understand cognition as it occurs in modern science. Attention is restricted to higher energy physics (HEP) practiced at a research laboratory such as the European Centre for Nuclear Research (CERN). The dominant paradigm in cognitive science would have us understand the acquisition of knowledge regarding elementary particles in terms of a cognitive agent that performs computations on a symbolic representation. In the case of HEP, however, it is difficult to identify a cognitive agent or the required symbolic representations and algorithms. The notion of ‘collective cognition’ provides a step in the right direction, but does not go far enough. I argue that we need to invoke a more radical notion of ‘distributed cognition’ that explicitly includes artefacts. However, the inclusion of artefacts in scientific cognitive systems makes application of the cognition as computation paradigm even more problematic. I conclude that scientific cognitive systems are hybrid systems including both computational and dynamical artefacts as well as human agents. Scientific knowledge itself is best regarded as the impersonal product of a whole scientific community.