Let us now return to the question of intuition. Optimal guidance by "evaluative mapping" of future possibilities must combine stored and computed information of a large variety of kinds with the representations furnished by perception and spatial mapping. For greatest functionality, efficiency, and effectiveness in regulating the organ- ism/environment interchange (e.g., guiding action selection), the predictive and explanatory models thereby created must forge together evidence from:
- • Our different senses
- • Information about causal relationships or similar patterns of events, drawn from memory of particular incidents (episodic memory) or general tendencies (semantic memory)
- • "Interoceptive" information about the internal state of the organism, including available resources, more or less urgent needs, and longer-term goals
- • Representations of the physical and social surroundings, whether proximate or distal
- • Indicators of personal location and trajectory
- • Imaginative representation of an array of possible actions and outcomes and the ways in which values or likelihoods depend on which action is taken
Could all of this really be going on as you sit at your kitchen table, pondering how to respond to your coworker's invitation? Don't we know that such responses tend to be based on relatively simplistic thinking and blunt "gut feelings?" The reader versed in contemporary social psychology will know of the huge body of research based on "dual-process" models of the mind (Bargh & Chartrand, 1999; Haidt, 2001; Kahneman, 2011). In this model, system 1 is intuitive. It is largely an implicit system, fast, automatic, and affective rather than calculating, and it is dominated by such simplifying heuristics as "anchoring," "representativeness," and "salience," with "little understanding of logic and statistics" (Kahneman, 2011, pp. 21, 24). System 2 is conscious and deliberative. It is the "declarative" system of conscious thought; it is capable of logical and statistical reasoning, but it is slow and effortful, placing demands on attention. In ordinary life system 2 plays a much more restricted role in our behavior. How does the dual-process picture square with our picture of the implicit, intuitive system as actively engaged in prospective assessment via expected value representations and causal modeling?
We have already seen that although our affective system is "ancient" and grounded in brain systems we share with our mammalian ancestors, this need not make it unsophisticated at learning and using information about probabilities and values or modeling the prospects and perils of a complex physical and social environment. Why wouldn't millions of generations of natural selection for optimal foraging have evolved and improved mechanisms that would be effective and efficient at evaluation and action-selection? And given what we know about the brain's power and speed, the rapidity of intuitive responses need not mean that they do not involve complex processing of information or calculation of expected value and risk (see Quartz, 2009). Intriguingly, humans who are given a chance to engage intuitively in implicit learning using representative samples and feedback seem not to make some of the errors with probability that have been well-documented in the "heuristics and biases" literature (Behrens et al., 2007; Pleskac & Hertwig, 2014). Perhaps the right moral to draw is that the "intuitive" system, like any system that projects on the basis of experience, needs good data to yield good results.
There are many limitations of probabilistic learning. For example, while such systems can yield new categories through learning, in solving life's practical, social, and theoretical problems we often must introduce new categories or hypotheses in a directed fashion, ahead of learning. Language and declarative thought are well-suited to provide this sort of cognitive scaffolding, especially since they enable us to share what we have learned and to deliberate or innovate together—even though they cannot do so effectively or efficiently without relying extensively upon the intuitive attitudes we have acquired through experience. We are brought back to Aristotle's observation in the Posterior Analytics (trans. 1941, 11.100b) that reasoning cannot supply its own premises. As Paul Slovic put it, "analytic reasoning cannot be effective unless it is guided by emotion and affect" (Slovic, Finucane, Peters, & MacGregor 2004).
What then would our experience be like if we actually were guided by prospection? It would be like the experience of our ice fisherman in Chapter 1 or like ourselves when faced with an unanticipated email invitation. It would be like the bad feeling of returning home to a hungry family empty-handed, or like the feeling that it just might be a good idea to make a second try at catching the fish. It would be like sensing that the fish might be startled by your shadow, and that you'd have a better chance if you snuck up from the other side of the ice hole. And it would be like the strong sense that turning down your coworker's invitation, even though you have mixed feelings about accepting it, is nonetheless not the thing to do.
Finally, it would be like your growing sense that this chapter has gone on long enough to make its point: human simulation and evaluation of possible futures—prospection—can take place implicitly as well as explicitly, and such implicit prospection could underlie intuition and explain why it is so central, and so often effective, in human life.