As well as trying to characterise meaning or representational content in relatively simple information processing systems in the brain, we’re also interested in identifying ways that information processing is different in more complex cases. Clearly consciousness may make a difference, as we’ve discussed already. Another popular way of distinguishing between types of information processing carried out by cognitive systems is the popular ‘dual process’ view, which sees the mind as operating with two systems, ‘system 1’ and ‘system 2’.
Even if there’s a lot of disagreement about whether the human mind really has two different systems, there is broader acceptance that there are two characteristic modes of operation, or a continuum that can be characterised by these two extremes. System 2, or type 2 cognition, is the kind of thinking that we’re all familiar with from our everyday reasoning. We use it when we work out what to do in a slow, serial way, often involving language. System 1, or type 1 cognition, is fast and automatic – it does not depend on deliberate reasoning. Crucially, type 2 processes interfere with one another (try solving an arithmetical problem while counting backwards from 20) whereas type 1 processes don’t. Type 1 processes can go on in parallel in their fast automatic way, without interfering with each other. The interference between type 2 processes is called ‘cognitive load’. Type 2 processes are subject cognitive load whereas type 1 processes are relatively unaffected by cognitive load.
Interestingly, theorists have often put consciousness into the mix, characterising type 2 processing as conscious and type 1 processing as non-conscious. In a recent paper, Chris Frith and I take issue with doing that. We argue that these are two different distinctions and that it is important to keep them apart. The standard dual process distinction is – as it says – about processes: whether a cognitive process is deliberate and subject to cognitive load; or automatic and relatively immune to cognitive load. The other distinction is about representations – whether the representations involved in a piece of processing are conscious or not.
Drawing this distinction should be relatively uncontroversial. In the paper we review some scientific findings to show that both distinctions are needed to understand what is going on in some of the experiments.
We are ultimately interested in the functions of consciousness, and we argue that, to home in on what consciousness does for us, it is important not to elide the type 1 / type 2 distinction with the conscious / non-conscious distinction. Indeed, there are cases where automatic, load-insensitive processes take conscious representations as input and produce conscious representations as output. Importantly, there are also lots of cases where automatic, load-insensitive processes take place entirely on non-conscious representations. We argue that it is important to identify this as a separate category of cognitive processing. We label it ‘type 0’ cognition. In cases of type 0 cognition consciousness is entirely absent. By contrasting these cases with the cases where consciousness is involved, both of type 1 and type 2, we get a clearer picture of what difference consciousness is making to the way information processing takes place.