.png){kind=icon}
Chapter Summary
This chapter examines how we think in order to solve problems. In classical approaches, Gestalt psychologists focused on insight problems, solutions to which required the problem solver to con-sider different angles, and were often characterized by Gestalt switches—a sudden change in how information is organized. Wertheimer, the founder of Gestalt psychology, described two ways of conceptually approaching a problem: productive thinking, which is more likely to allow Gestalt switches and insight, and structurally blind thinking, which is more likely to prevent these pro-cesses from occurring.
In his examination of how past experience influences our problem-solving, Duncker claims that our initial analysis of the situation often leads us to adopt functional fixedness. Functional fixed-ness is a barrier to effective problem solving and refers to an inability to see beyond the most common uses for an object. Hints, in some cases, help people overcome this phenomenon. Solutions to problems can involve insight or they can be solved without insight. In the latter case, participants are said to experience a feeling of “warmth” or a feeling of knowing. Such feelings reflect individuals’ metacognition. Insight solutions, on the other hand, are said to arrive suddenly, involuntarily, and in an all-or-none fashion.
Currently, there are two different approaches to the study of problem-solving using insight. First, the progress monitoring theory states that people use the simplest route to solve a problem, and only consider alternatives once the simplest route has failed, and consequently monitor their progress. Second, the representational change theory concentrates on the two processes (constraint relaxation and chunk decomposition) that lead to representational change. Neuroscience studies have shown that activity in the anterior cingulate cortex (ACC) and the hippocampus is associated with the insight process. Also, sleep is thought to promote insight.
Researchers have examined functional fixedness and found that it is a universal, developmentally acquired trait. German and Barrett (2005) showed that functional fixedness persisted amongst people that have little advanced technology living in the Shuar area of the Amazon in Ecuador. Fur-thermore, as demonstrated by Luchinses’ jar experiments, people can become so accustomed to a particular way of solving a problem that they can become blind to other strategies (Einstellung effect). In other words, they tend to show negative transfer, a strong but wrong routine. Neuroscience studies suggest that the dorsolateral prefrontal cortex is especially important in overcoming the Einstellung effect. Other approaches cast this effect as the mindless application of a solution to each new problem rather than acting in a mindful way by seeking alternatives.
Human problem-solving is also studied using computer simulations. This is known as an artificial intelligence approach. Artificial intelligence often makes use of heuristics and algorithms. Computer programs often require an evaluation function, whereby all possibilities are considered and evaluated. Computerized games have what is known as a problem space. Chess, for example, has a very complicated problem space because it has a large search tree. Like in many areas in psychology, a good way to collect data on problem-solving is to examine simplified problems that are not part of everyday life, such as toy problems. The “Tower of Hanoi problem” is a good example of a toy problem and can be worked out by the General Problem Solver (GPS) with the use of production rules. The process by which GPS actually solves a problem is called means–end analysis. GPS creates a goal stack, where subgoals are piled on top of the final goal. In order to perfect computer programs, researchers had participants think aloud as they solved problems. They then used the human thought processes to aid in programming.
Science has also provided a fruitful area for the study of problem-solving. Four complementary methods assist with scientific problem-solving: historical accounts, observation of ongoing scientific investigations, laboratory studies, and computational models. Nersessian came up with the term “cognitive history of science” which refers to the combination of case studies through-out history and of scientific investigations on cognition. Through this approach, phenomena like the Zeigarnick effect can be observed. The second method, observation of ongoing scientific investigations, involves in vivo research (versus in vitro). In such research, it has been found that unexpected findings occur quite often and are a crucial aspect of science, as is distributed reasoning. Unlike laboratory studies and computational models, historical accounts are considered to be face valid. Finally, computational models, like BACON, can also help us learn about problem-solving in science.
Chapter Objectives
- To describe the Gestalt approach to insight and problem-solving.
- To consider functional fixedness and how it can hinder problem solving.
- To examine artificial intelligence approaches to problem-solving and how they resemble the ways humans solve problems.
- To discuss the various approaches to the study of problem-solving in science.