Get complete information on Computer Programs and Flow Charts

Before examining the role of the computer as a tool for studying cognitive processes it will be useful to describe briefly the basic features of a computer program. The computer cannot figure out how to solve a problem by itself; it is helpless until it has been given a detailed set of instructions.

These instructions make up what is called the computer program. To write a program it is first necessary to analyze the problem to be solved and to break it into its component parts. One of the best ways to do this is to construct a flow chart, which resembles the play diagram a football coach might draw on a blackboard.

A flow chart shows component parts of the problem just as a coach’s diagram indicates each player’s assignment. The flow chart also shows how the various parts are to be fitted together.

Once the problem has been mapped out, each part of the chart must be further broken into detailed instructions on how to handle each operation. One part of a flow chart might require a hundred or more individual steps in the program. Then the material can be fed into the computers.

Figure shows a flow chart of a very simple information-processing problem. The chart tells you the steps involved in finding the distribution of word lengths for a passage of English text.

The input to the program is the text material, which has been punched on cards with numerical codes for the letters and space with all other punctuation ignored. The boxes in the flow chart represent work to be done, and the diamonds, decisions to be made.

This flow chart by itself is not very impressive, but when many of them are cascaded one onto another into a hierarchy of operations the computer can indeed perform in a truly intelligent fashion.

The flow chart of a computer program is a convenient way of picturing the flow of information through a system. The human organism can be conceptualized as an information-processing system, and it is quite natural sometimes to use flow charts as models of psychological processes.

Models based on flow charts have come to be called information-processing models. Many psychologists feel that information-processing models are well suited for theorizing about psychological phenomena, particularly about complex cognitive processes.

Simulation Models:

Computer programs used to mirror the cognitive activity of human beings are called simulation models. The first significant attempt to simulate complex cognitive processes was made by Newell and Simon (1956), who developed an information-processing, model to prove theorems in symbolic logic.

Their program, which was dubbed the Logic Theorist, did not try to prove theorems by searching all possible sequences of logical operations until one was found that yielded a proof, Rather, the approach taken in the Logic Theorist was to incorporate heuristic methods of the type used by human beings for proving theorems.

A heuristic is a strategy, trick, simplification, gimmick, or any other procedure that drastically limits the search for solution in difficult problems. Heuristics do not guarantee that a solution will be found, but when they do work, they greatly reduce the search time required to obtain a solution. Human thinking obviously makes use of heuristic procedures.

Good chess players solve their problems heuristically, since they cannot possibly foresee the consequences of every possible move (see Figure). In solving a geometrical problem we often add a line here or there, hoping that by forming a new diagram we may perceive relations that were previously not evident. This new construction may help, but there are no guarantees implied.

The use of heuristic procedure by humans to not well understood, but there can be little doubt that they play an important role in most problem-solving tasks. Several examples of heuristics particularly relevant to human thinking are worth describing.

One of these is the heuristic of “working backwards”: we begin with the result to be proved and then attempt to work backwards step by step to that which is initially given.

Another is the “make-a-plan” heuristic: we think of a problem similar to the one we are trying to solve but to which the solution is already known; this method of solution is then used as a plan for solving the more difficult problem.

A third heuristic is the “means-end” procedure: here we compare the current state of affairs with that which we wish to obtain, find a difference between the two states, and seek an operation that will reduce the difference, and repeat the procedure until we obtain the desired effect.

The logic Theorist, with its various heuristic methods, is an extremely impressive computer model of human thinking.

For example, it has been used to derive the 52 theorems in the second chapter of Whitehead and Russell’s famous treatise, Principia mathematical (1925); whenever a theorem was proved, it was stored in memory and was available, together with the original axioms, for use in proving adequate proofs for 38 of the theorems, and some of the proofs were more elegant than those originally offered by Whitehead and Russell. Or course, the Logic Theorist was not programmed to provide more rapid or “better” proofs than a human would, but rather to simulate human behavior is an actual problem-solving task.

When a computer can be programmed to perform such a task and behaves in a way that is very much like a human, then indeed progress is being made in understanding thought processes.

Since the development of the Logic Theorist many investigators have formulated simulation models for an array of psychological processes. There are models for concept formation, language comprehension, attitude change, music composition, chess playing, and even for neurotic personality processes-to name a few.

These exciting developments are of great importance in unraveling the problems of human thinking. But the development of these models is clearly a two-way street, for one has to know something about the nature of creative problem solving in order to program it on a computer. Because the computer will do only what it is instructed to do, the steps have to be clearly and completely specified in the computer program.

If the psychologist who charts the computer program has made incorrect assumptions in interpreting the steps involved in problem solving, the program will not succeed or at least will not display outputs that accurately match the behavior of human subjects.

The computer serves to check the adequacy of the theoretical assumptions that the psychologist believes will account for the psychological process understudy. The chief advantages of the computer are its speed and its attention to all details of what it is programmed to do.

The development of information-processing models for psychological phenomena is still in an early stage, but the results have been encouraging. Newell and Simon (1972) have argued that there is already substantial evidence for explaining much of human thinking and problem solving in terms of a few basic processes, arranged into an appropriate hierarchy yielding outputs that appear incredibly complex.

They suggest that the human information-processing system is primarily serial in its operation. It can process only a few symbols at a time, and the symbols being processed must be held in a limited short-term memory, the content of which can be rapidly reordered and changed.

The major limitations on the subject’s capacities to employ efficient strategies arise from the very small capacity of the short-term memory and from the relatively long time needed to transfer information from short- term to long-term memory.