2.1.2 Characterizing Problems

Given the strong emphasis on problem‑solving in this context and in the general literature on thinking, the question arises "what is a problem?" The definition offered by the Gestalt psychologist Karl Dunker is still serviceable. He wrote that "a problem arises when a living organism has a goal but does not know how this goal is to be reached".

This is a useful initial formulation that signals a number of points. First, that a "task" set by an experimenter is not necessarily a problem for a given individual. Whether it is a problem or not, depends on the subject's knowledge and on his ability to locate relevant knowledge, should he have it? Second, a problem may vanish or be dissolved if the person changes his goals. A third point is that a problem does not effectively exist until the person detects some discrepancy between his goals and the situation he finds himself in.

Most psychological studies of problem‑solving (especially, as we shall see, those within the information processing framework) have dealt with well defined problems. If we accept Rittman's useful proposal that problems in general can be viewed as having 3 components (viz. a starting state, a goal state and a set of processes that may be used to reach the goal from the starting state) then a problem is well defined if all 3 components are completely specified. Problems in mathematics, in logic and in various board games tend to be well defined. Although well defined, such problems can be very difficult and the psychologist is faced with the task of explaining how we humans, with our various limitations, manage to solve geometry, chess and similar scale problems in reasonable time. Of course, it will be still more difficult to explain how we tackle those ill‑defined problems that are more typical of real life than the well‑defined variety.

Undefined problems leave one or more components of the problem statement vague and unspecified. Problems can vary in degree of defineness an animal.

It seems a reasonable strategy for psychologists to start with people's ways of handling apparently well‑defined problems and then move on to consider ill‑defined tasks. Perhaps people tackle ill‑defined tasks by seeking a well‑defined version of the problem, which they then work within until the problem is solved or a new definition is tried. If this is so, then studies with well‑defined problems will be relevant to part of the process of solving ill‑defined problems. Indeed, processes of defining, or interpreting, the problem are also important in well‑defined tasks and some attention has recently been given to task interpretation processes that must play a role in both well‑ and ill‑defined tasks.