One of the more enduring ideas in psychology, dating back to the time of William James a little more than a century ago, is the notion that human behavior is not the product of a single process, but rather reflects the interaction of different specialized subsystems. These systems, the idea goes, usually interact seamlessly to determine behavior, but at times they may compete. The end result is that the brain sometimes argues with itself, as these distinct systems come to different conclusions about what we should do.
The major distinction responsible for these internal disagreements is the one between automatic and controlled processes. System 1 is generally automatic, affective and heuristic-based, which means that it relies on mental “shortcuts.” It quickly proposes intuitive answers to problems as they arise. System 2, which corresponds closely with controlled processes, is slow, effortful, conscious, rule-based and also can be employed to monitor the quality of the answer provided by System 1. If it’s convinced that our intuition is wrong, then it’s capable of correcting or overriding the automatic judgments.
One way to conceptualize these systems is to think of the processes involved in driving a car: the novice needs to rely on controlled processing, requiring focused concentration on a sequence of operations that require mental effort and are easily disrupted by any distractions. In contrast, the well-practiced driver, relying on automatic processes, can carry out the same task efficiently while engaged in other activities (such as chatting with a passenger or tuning in to a radio station). Of course, he or she can always switch to more deliberative processing when necessary, such as conditions of extreme weather, heavy traffic or mechanical failure.
In terms of decision-making, the description of System 2 bears a close resemblance to the rational, general-purpose processor presupposed by standard economic theory. Although these economic models have provided a strong and unifying foundation for the development of theory about decision-making, several decades of research on these topics has produced a wealth of evidence demonstrating that, in practice, these models do not provide a satisfactory description of actual human behavior. For instance, it’s been recognized for several decades the people are more sensitive to losses than to gains, a phenomenon known as loss aversion. This doesn't fit with economic theory, but it appears to be hard-wired into the brain.
A major cause of these observed idiosyncrasies of decision-making may be that controlled processing accounts for only part of our overall behavioral repertoire, and in some circumstances can face stiff competition from domain-specific automatic processes that are part of System 1. One recent compelling demonstration of this phenomenon comes from Princeton University psychologist Adam Alter and colleagues, who examined how subtle changes in contextual cues, such as altering the legibility of a font, can facilitate switching between System 1 and System 2 processing.
In a series of clever experiments, the authors manipulated the “perceptual fluency” of various sets of stimuli. In other words, they made it harder for people to understand or decipher the scenarios they were asked to judge. For example, in one experiment participants were asked a series of questions, known as the Cognitive Reflection Test, designed to assess the degree to which System 1 intuitive processes are engaged in decision-making. In this test the gut reaction answer is invariably incorrect. (An example: if a bat and a ball together cost $1.10, and the bat costs $1 more than the ball, how much does the ball cost? If you find yourself wanting to shout out “10 cents, of course,” then you’re in the majority, but sadly also wrong.) Alter et al. found that by making the problem simply more difficult to read (by using grayed-out, reduced-size font), participants seemed to shift to more considered, System 2 responses, and as a result answered more of the questions correctly.
The authors repeated this effect in various situations. For example they degraded the byline of the author on a review of an MP3 player. As a result, participants were less influenced by the apparent competence of the reviewer, which would have been based on viewing a picture of him or her, and more by the actual content of the review. In an additional scenario, they ask participants to either furrow their brow or puff their cheeks while assessing statistical information. The former activity is a cue for cognitive effort and as such led to decreased reliance on (incorrect) intuition, and more on dispassionate analytic thinking.
These examples are important for several reasons. Most trivially, they are a good example of the ingenuity of researchers in finding interesting new ways to demonstrate the existence of the two purported systems. More important however, they begin to address the issue, largely ignored until now, of exactly why and when the various systems are employed in judgments. The work can lead towards more accurate predictions of when the respective Systems may be engaged.
Finally, the examples illustrated here have the potential to contribute to how these systems may be usefully applied to construct environments that foster more sensible decisions. In a similar vein, a recent movement in behavioral economics seeks to acknowledge the limitations of everyday decision-making (such as the apparent reluctance of workers to contribute to 401K plans) and therefore design institutions in such a way as to ‘encourage’ better choices (such as introducing default options for retirement savings). Work led by Richard Thaler has demonstrated that, when people are asked to commit to saving money in the distant future (as opposed to right now), they end up making much more economically rational decisions. This is because System 2 seems to be in charge of making decisions that concern the future, while System 1 is more interested in the present moment.
Of course, there are still many outstanding questions regarding the multiple-system model, not least the degree to which these proposed systems actually exist and are truly separable. The welcome integration of neuroscience with traditional experimental psychology has led to some debate about how, and where, exactly, these systems are instantiated in the brain. Although there is a good deal of evidence for some level of dissociation between multiple systems that approximate controlled and automatic processing respectively, with parts of the brain such as areas of frontal cortex (controlled) and limbic regions (automatic) implicated in these processes, it seems highly unlikely that there are dedicated, independent, sub-systems at the neural level that are specific to these modes of processing. Therefore, one important question is whether the types of systems that have been described at the psychological level are a good analogue for the way information is organized and processed in the brain. Research such as Alter et al.’s work points to the importance of becoming increasingly more specific about the situations and conditions that engage these distinct systems, which will prove to be essential in understanding how these multiple systems interact at a neural level.