ABSTRACT
There are situations where individuals encounter completely novel problems for which no analogical problem and solution information may be known. The nine-dot problem (illustrated in Figure 3.1) presents such an example. Until recently, accounts of human performance have generally proposed that problem solvers impose additional and inappropriate constraints that preclude the discovery of a solution. Once these inappropriate constraints are removed, so the story goes, the problem then becomes easy to solve. The traditional Gestalt account is that the nine-dot array is unavoidably processed perceptually as a square whose boundaries should not be violated. Weisberg and Alba (1981), amongst others, point out that providing instructions to include lines that go beyond the boundary of the square does not reliably lead to solution. Instead, Weisberg and Alba propose that individuals impose an inappropriate constraint that lines should extend only between dots, a constraint derived from prior experience of dot-to-dot drawing. Curiously, they fail to notice that their critique of the Gestalt position applies equally to them.
