A ‘heuristic’ is a strategy that helps make decisions, or discover solutions to problems. What is special about heuristics, though, is that they are a fast, informal - intuitive - way of doing so. In making decisions, we rarely have all the facts at our disposal, and even if we did, our minds would not have the capacity to weigh up potentially infinite possible options. Human rationality is not unbounded. It is ‘bounded’ (Simon 1956), and evolution has left us with economical ‘rules-of-thumb’ that enable us to make the most of our finite cognitive capacity.
Much current work in evolutionary psychology conceives of the mind as an ‘adaptive toolbox’, as a set of dedicated cognitive mechanisms - or ‘modules’ - likely to have evolved in small incremental steps (Barkow et al. 1995; Sperber 2001b). One way in which such mechanisms might improve overall cognitive efficiency is by providing what Gigerenzer et al. (1999) call ‘fast and frugal heuristics’, which apply to a particular domain, and yield reliable conclusions when applied to input from this domain. As Gigerenzer and Selten (2002: 7) put it: ‘heuristics are middle-ranged, that is, they work in a class of situations … What we call the adaptive toolbox contains a number of these “middle-range” tools, not a single hammer for all purposes.’
On the face of it, having to trust heuristics may seem disadvantageous, especially since they are not foolproof. Gigerenzer and Selten (2002) suggest not. They ask us to consider a thought experiment in which two teams are set the task of designing a robot that can catch a ball (no such robot exists). One team adopts the unbounded ‘omniscientific’ approach and programmes a robot with all the necessary knowledge of the projected parabolas a ball might follow, along with myriad other instruments to perform the calculations that will get the robot to the right place to wait and catch the ball. The other team study what cricketers or baseball players actually do (the first team dismisses this idea because, since sportsmen aren't conscious of the measurements and calculations they are using when they catch a ball, interviewing them or watching them would serve no purpose). On the basis of this approach, they programme the robot to follow what has been called the gaze heuristic. The robot does not move immediately the ball is airborne. Instead, it makes a rough estimate of whether the ball is going to land in front of it or behind it. The robot then starts running in the appropriate direction, whilst fixing its gaze on the ball and adjusting its running speed so that the angle between the eye and the ball remains the same. Using this method, the robot does not need to calculate where the ball will land. Provided it can move quickly enough, it will catch the ball whilst it is running.
Many such heuristics have been identified. They include the recognition heuristic, by which we tend to assign higher value to objects with which we are familiar and the contagion heuristic, by which we tend to avoid contact with objects that have come into contact with other objects we regard as contaminated. Emotions may be heuristics. Faced with a dangerous animal, for example, fear puts our body into the state it needs to be in to either fight or run away: we don't need to reason ourselves into feeling frightened (though we sometimes try to reason our way out of it).
There are clear implications for pragmatics. After all, the central question in pragmatics is how it is that hearers accurately and seemingly effortlessly infer speaker meaning. Heuristics would seem to be an appropriate choice. Levinson (2000) makes use of what he calls an ‘I-heuristic’, which yields default inferences in the form of conclusions that are automatically drawn but may be overruled by contextual information. Relevance theory's cognitive approach (Sperber and Wilson 1995) sees cognition and communication as relying heavily on fast and frugal heuristics, which make it possible to pick out potentially relevant inputs to cognitive processes and process them in a way that enhances their relevance. Both these approaches diverge from more traditional, Gricean accounts of intentional communication (Grice 1989) - indeed, from philosophical characterizations of utterance comprehension generally - which rationally reconstruct the comprehension process in the form of conscious and reflective inferences about the mental states of others.
Having said that, it's worth noticing that in more recently published work, Grice (2001b: 17) describes a view of inference in which inferential processes did not always have to be conscious and explicit: ‘we have … a “hard way” of making inferential moves; [a] laborious, step-by-step procedure [which] consumes time and energy … A substitute for the hard way, the quick way, … made possible by habituation and intention, is [also] available to us.’ Perhaps a heuristic is what Grice had in mind.
See also: Artificial intelligence; cognitive pragmatics; generalized conversational implicature, theory of; Grice, H.P.; modularity of mind thesis; modular pragmatics; neo-Gricean pragmatics; philosophy of language; philosophy of mind; post-Gricean pragmatics; rationality; reasoning; relevance theory; social cognition; utterance interpretation
Related Credo Articles
—Georg Polya, How to Solve It Polya 1954: 68. Anything is right that leads to the right idea. So far in this book it has been argued that you can
Abstract Heuristics are approximate strategies for decision making. They do not process all the available information and therefore can yield a reas
The term heuristic is of Greek origin and means serving to assist in finding out or discovering something. Imagine that you feel sick and...