Constraint Satisfaction
The concept of constraint satisfaction is crucial for understanding and improving human reasoning and decision making. A "constraint" is a condition that must be taken into account when solving a problem or making a decision, and "constraint satisfaction" is the process of meeting the relevant constraints. The key idea is that often there are only a few ways to satisfy a full set of constraints simultaneously.
For example, when moving into a new house, my wife and I had to decide how to arrange the furniture in the bedroom. We had an old headboard, which was so rickety that it had to be leaned against a wall. This requirement was a constraint on the positioning of the headboard. The other pieces of furniture also had requirements (constraints) on where they could be placed. Specifically, we had two small end tables that had to be next to either side of the headboard; a chair that needed to be somewhere in the room; a reading lamp that needed to be next to the chair; and an old sofa that was missing one of its rear legs, and hence rested on a couple of books — and we wanted to position it so that people couldn't see the books. Here was the remarkable fact about our exercises in interior design: Virtually always, as soon as we selected the wall for the headboard, bang! The entire configuration of the room was determined. There was only one other wall large enough for the sofa, which in turn left only one space for the chair and lamp.
In general, the more constraints, the fewer the possible ways of satisfying them simultaneously. And this is especially the case when there are many "strong" constraints. A strong constraint is like the locations of the end tables: there are very few ways to satisfy them. In contrast, a "weak" constraint, such as the location of the headboard, can be satisfied in many ways (many positions along different walls would work).
What happens when some constraints are incompatible with others? For instance, say that you live far from a gas station and so you want to buy an electric automobile — but you don't have enough money to buy one. Not all constraints are equal in importance, and as long as the most important ones are satisfied "well enough," you may have reached a satisfactory solution. For example, although an optimal solution to your transportation needs might have been an electric car, a hybrid that gets excellent gas mileage might be good enough.
In addition, once you begin the constraint satisfaction process, you can make it more effective by seeking out additional constraints. For example, when you are deciding what car to buy, you might start with the constraints of (a) your budget and (b) your desire to avoid going to a filling station. You then might consider the size of car needed for your purposes, length of the warrantee, and styling. You may be willing to make tradeoffs, for example, by satisfying some constraints very well (such as mileage) but just barely satisfying others (e.g., styling). Even so, the mere fact of including additional constraints at all could be the deciding factor.
Constraint satisfaction is pervasive. For example:
• This is how detectives — from Sherlock Holmes to the Mentalist — crack their cases, treating each clue as a constraint and looking for a solution that satisfies them all.
• This is what dating services strive to do — find the clients' constraints, identify which constraints are most important to him or her, and then see which of the available candidates best satisfies the constraints.
• This is what you go through when finding a new place to live, weighing the relative importance of constraints such as the size, price, location, and type of neighborhood.
• And this is what you are do when you get dressed in the morning: you choose clothes that "go with each other" (both in color and style).
Constraint satisfaction is pervasive in part because it does not require "perfect" solutions. It's up to you to decide what the most important constraints are, and just how many of the constraints in general must be satisfied (and how well they must be satisfied). Moreover, constraint satisfaction need not be linear: You can appreciate the entire set of constraints at the same time, throwing them into your "mental stewpot" and letting them simmer. And this process need not be conscious. "Mulling it over" seems to consist of engaging in unconscious constraint satisfaction.
Finally, much creativity emerges from constraint satisfaction. Many new recipes were created when chefs discovered that only specific ingredients were available — and they thus were either forced to substitute different ingredients or to come up with a new "solution" (dish) to be satisfied. Creativity can also emerge when you decide to change, exclude, or add a constraint. For example, Einstein had one of his major breakthroughs when he realized that time need not pass at a constant rate. Perhaps paradoxically, adding constraints can actually enhance creativity — if a task is too open or unstructured, it may be so unconstrained that it is difficult to devise any solution.
