Psychologists often hold that personal-level, conscious processing is crucial for the performance of logical reasoning (e.g., Smith and DeCoster 2000; Lieberman et al. 2002; De Wall et al. 2008; Lieberman 2009). There is experimental support for this. For instance, De Neys (2006) presented subjects with logically valid arguments that had false premises and supported conclusions that were at odds with everyday experience. When subjects were under cognitive load, i.e., when their attention was engaged elsewhere, they made logic errors based on everyday, practical knowledge. However, when they were not under load, they performed better at evaluating the logic underlying the arguments. If we accept the common assumption that cognitive load hinders conscious processing while allowing unconscious and automatic processes to operate unimpaired then these results suggest that logical reasoning depends on conscious representations.
De Wall et al. (2008) too found that cognitive load impaired performance in a logical reasoning task. They made an interesting additional finding. In one of their studies, DeWall et al. manipulated unconscious processing via priming, and imposing unconscious load (by using Wegner’s (1996) thought-suppression paradigm). It turned out that while priming the unconscious system with the goal of being logical did lead to logic-relevant ideas being more accessible, it didn’t improve performance on the logic problems. Conversely, the imposition of unconscious load didn’t negatively affect the subject’s performance in the reasoning task. That is, when conscious processing was encumbered or preoccupied, logical reasoning became worse. But neither priming nor hindering the unconscious system led to changes in logical reasoning performance in either direction. This again supports the view that logical reasoning depends upon conscious processing.
But what exactly is it about conscious, personal-level processing and representations that logical reasoning depends on? One suggestion that is sometimes mentioned is that consciousness enables the performance of functions that require multiple, sequenced rule-based operations on more than one abstract unit. Might this be a distinctive feature of conscious, personal-level processing?
Perhaps things are more complicated. In a recent study, Sklar et al. (2013) provide evidence that the semantic processing of multiple-word expressions and the performance of abstract mathematical computations can in fact take place unconsciously also. For instance, when the equation 9 minus 5 minus 1 was presented non-consciously (via a prime), subjects were quicker in pronouncing 3 than 4, even though they didn’t consciously see the equation. This indicates that the equation was represented and solved without awareness at the sub-personal level. More generally, Sklar et al’s findings suggest that multiple rule-based operations on more than one abstract unit, as, e.g., in propositional, logical reasoning, may not require conscious, personal-level representations after all. Their implementation might not be a distinctive feature of personal-level information processing.
