'In two minds': consideration of evolutionary precursors permits a more integrative theory.
Trends in Cognitive Sciences, 8(2),
Evans  describes a dual-process account of human reasoning, involving an evolutionarily old System 1, common between humans and various non-human species, and a System 2 that, he proposes, is distinctively human.
This communication is designed to raise two issues: (i) How uniquely human is System 2?; and (ii) can the fundamental dichotomy between systems be revealed in more than just reasoning? In humans, System 2 doubtless exhibits distinctive characteristics in terms of some of the types of reasoning that it is able to perform. However, there is considerable evidence that a dichotomy very similar to the one described can also be applied, not only more broadly to the controls of behaviour in humans, but also to those of a range of non-human species [2 and 3]. These controls can compete for expression in behaviour.
In a classical paper concerning rat learning, Hirsh described a comparable distinction between ‘performance-line’ and ‘off-line’ controls . The latter were disrupted by hippocampal damage whereas the former remained intact. Expressed in various terms, a similar line of theorizing has been followed by a number of researchers in the area of animal learning and memory [5 and 6]. The distinction means that some of the old debates between cognitivists and behaviourists are now more correctly seen as an argument over which process dominates behavioural control rather than over incompatible theoretical perspectives [2 and 7]. Similarly, with the help of a comparable dual-systems approach to vision, Norman has recently argued convincingly that both Gibsonian and constructivist perspectives map onto two processes present within the same brain .
Under some conditions, in rats as well as humans, the control exercised by the performance-line system is inhibited by cognitive controls. For example, the tendency to repeat a behaviour pattern can be overridden when circumstances suddenly change. For rat as well as human, a novel situation requires the synthesis of a novel behavioural solution, presumably involving something that has features in common with Evans' System 2.
For too long, ‘human’ and ‘non-human’ researchers have tended to work in isolation. Of course, we need to avoid forcing a single model uncritically onto all species. However, by considering a broader comparative approach, common ‘design’ problems and the evolutionary precursors of the human dichotomous reasoning system, we might gain valuable sources of insight and better theoretical integration.
Actions (login may be required)