Verplanck, W.S. (1984) The Egg Revealed. The Behavioral and Brain Sciences, 7, (4) 605-606. Comment on B.F. Skinner's article, "Problem Solving," in same issue. Copyright Cambridge University Press

"I'm afraid you've got a bad egg, Mr.Jones."
"Oh no, my Lord, I assure you! Parts of it are excellent!"

"Problem Solving" needs either a brief and summary appraisal or an extended one--too long for the space likely to be allotted. Written lucidly and persuasively, it demonstrates that Dr. Skinner can define and analyze a broad set of behaviors using a l imited set of behavioral terms and concepts. Believers will find it convincing, even brilliant. Behaviorists who have worked on problem solving, who have observed their subjects carefully, and who have collected data analytically--especially those who h ave also labored to clarify the technical terminology of behaviorism operationally--will find it a good deal less than persuasive. Glib, superficial and misleading are more appropriate terms. Closely argued and exemplified points are followed by one lin ers, such as: "But to speak of the purpose of an act is simply to refer to its characteristic consequences." That cannot survive critical examination.

The difficulties arise from two sources: First, the definitions of "stimulus," "reinforcement," and "contingencies" are so broad, so flexible, and so loosely used that they can be, and are, applied to anything whatsoever. More precisely defined, more unequivocally applicable behavioral concepts are required if they are to be used fruitfully. We are past the stage of talking about Stoff. We can distinguish among solid, liquid, and gaseous states. We need to start thinking at the level of ele ments and their properties in isolation and in compounds with others.

The second difficulty lies in the absence of a substrate of observational data that would require Skinner to write and think more realistically about his subject matter. When you know little of your topic but have thought about it a great deal without actually looking, then it is easy to analyze the area with broad beguiling strokes.

In the author's terminology, this paper is an excellent example of rule-governed behavior, but the rules are those of logic and grammar, applied to India-rubber concepts. They are not rules derived from the contingencies of reinforcement and of observ ation that function when a behaviorist carries out research on people faced with, and solving, problems.

Skinner is to be congratulated on his belated discovery that subjects must "construct" or find discriminative stimuli if they are to solve problems, and for his rediscovery of what I term "notates," "notants," and "monents" (Verplanck, 1962). The dist inction between "contingency-shaped" behaviors (conformances) and "rule-governed behaviors" (compliances) is appropriately drawn, although the former are inadequately described and would seem to exclude imitating (contingency polished, perhaps, but not sh aped), which often appears in problem solving, and cannot be excluded from either consideration or research by labeling it "innate" or "instinctive." Missing, too, is molding, or being "put through," which problem solvers often request. But perhaps thes e, too, are covered by the "contingency" tent.

The analysis, then, is incomplete, vague to the point of total imprecision and unfalsifiability. Hence it is not likely to discomfit "cognitive" psychologists, much less persuade them.

In sum, this is an exercise in the use of language. It is not a scientific contribution but an interpretive one, contributing little that is new (except to Skinner's own thinking); nor will it stimulate the research that will be needed if a less liter ary and more searching behavioral analysis of problem solving is to be done.