“Mechanical Reasoning by Mental Simulation” by Hegarty Essay (Article Review)

Abstract

Mechanical reasoning is part of reasoning mental ability. The article to review is “Mechanical reasoning by mental simulation”, written by Mary Hegarty. The author has been publishing on this particular subject since the early 1990s and presented this review article based on personal experience and review of the work of other psychologists.

Introduction

Louis Thurston, 1938, revolutionized the old theory, which considered intelligence as a single general mental ability. Instead, Thurston suggested there are seven primary mental abilities. These primary mental abilities are verbal comprehension, reasoning, perceptual speed, numerical ability, word fluency, associative memory, and spatial visualization.

The aim of this work is to review the article “Mechanical reasoning by mental simulation by Mary Hegarty. The article is a review of the literature and published in Trends of Cognitive sciences volume 8-issue number (6) in 2004.

According to Hegarty, the reasoning is a cognitive process, in which a person infers new information from given information. As mechanics is the science of motion, Mechanical inference is a reasoning mental ability by which people drive information about the movement of objects.

Article Review

The research question

The author proposed two research questions to answer in this article. First, is the way by which individuals mentally represent mechanical reasoning problems? Second, what are the mental algorithms that work on these representations if anyone is to make a mechanical deduction (mental model)?

The hypothesis

Reviewing the literature, the author stated that according to current research, an individual uses mental stimulation for mechanical reasoning. The author added that current studies assume the basis of mental simulation is an internal spatial representation of the mechanical test example. In addition, new research on mental imagery and representation enabled psychologists to assess the nature of these algorithms. From this perspective, the author reviewed the literature on mental representation in mechanical reasoning.

The test of the hypothesis

Hegarty composed the literature review focusing on four main points. They are spatial representations in mechanical reasoning, analog processes in mechanical reasoning, mental stimulation versus visual imagery, and knowledge representation in mental simulation of mechanical systems.

Spatial representations in mechanical reasoning:

Since motion is a spatial character, the author explained, so in mechanical reasoning, the represented world (the referent) is also spatial and is a spatial array. This means that characteristics of mechanical reasoning are: A) it depends on visible spatial information (objects locations, objects shapes, and their connectivity). B) It includes representation of nonvisible characteristics (for example force) and C) the relations, in space, among objects match the spatial relations among the objects they represent.

Individual differences in spatial representation:

In previous experiments (on solving a pulley problem), Hegarty suggested that solving mechanical reasoning problems showed a strong association with measures of spatial abilities and a nonsignificant association with verbal abilities. Analysis of individual differences in spatial representations offers evidence that mechanical reasoning may depend on spatial representations.

Dual-task studies:

Based on the assumption there are separate buffers in working memory, whose task is to keep “visuospatial” representations from verbal representations, Hegarty performed earlier experiments. The author and colleagues, in an earlier publication, measured the interference between mechanical reasoning and maintenance of a “visuospatial” buffer (working memory load) versus a verbal buffer (working memory load).

Protocol studies:

In these studies, individuals are asked to think aloud while reasoning a mechanical inference question. They often use imitative body movements to tell the thought of how different units of a mechanical array move. These body movements come ahead of verbal descriptions of the units’ motion.

Analog processes in mechanical reasoning:

In this section, the author proposed two questions to answer; first, if answering mechanical reasoning questions is by mental algorithms, is there evidence that inference algorithms are similar to the physical course of action they simulate. Second, on reasoning by mental stimulation, do individuals have precise knowledge of the physical question they are reasoning?

Reaction times:

The author reviewed two earlier studies, the first tested whether mental rotation of objects is similar to real physical rotation. The second study used a system of two interlocking gears to test whether individuals’ response time is proportional to the angle of rotation when trying to infer the rate of rotation of the two systems.

Is there any evidence for the separation of physical representations from clear and obvious knowledge? It is important to show that people reasoning by mental simulation have a definite knowledge of the physical situation they are reasoning.

Mental simulation versus visual imagery:

The author asked three questions; is mental simulation a universal or holistic image of a physical situation? Second, is the basis pure visual information? Third, does mental simulation involve motor representations as well as visual representations?

4- About knowledge representation in mental simulation of mechanical systems, the author reviewed the literature about the following points: how might psychologists characterize the working memory representation on which mental simulation is based? As direct knowledge cannot explain mental simulations, can implicit knowledge explain them? Finally, what is the representational format?

Findings

1. Analysis of individual differences in spatial representations gives evidence that mechanical reasoning involves the changeover of spatial representations and is less contingent on verbal representations.
2. Reviewing dual-task studies suggest that mechanical reasoning depends on representations in the “visuospatial” buffer.
3. A thorough look at previous studies showed that in protocol studies, body movement occurs before verbal description verbal descriptions of motion.
4. For analog processes in mechanical reasoning, analysis of the first study showed the time to rotate an object mentally is proportional to the angle of rotation. The second study suggested that response time is proportional to the angle of rotation. Further testing of individuals trained to use an analog imagery strategy to solve the gear question, results were not significantly different from those of individuals in no training conditions. Further studies analysis showed that on answering from explicit knowledge, performance was poor. This suggested the evidence for a dissociation of analog processes from explicit knowledge.
5. Reviewing the literature, the author reached the following results about mental simulation versus visual imagery. First, people mentally simulate the way mechanical systems behave piecemeal but not holistically. Second, the basis of mental simulation is solely visual information. It also includes information about non-visible characters as force and density. Finally, mental stimulation leading to mechanical assumption may involve motor representations besides visual representations.
6. About knowledge representation in mental simulation of mechanical systems, Hegarty suggested that it is likely that mental simulations in mechanical reasoning depend on transformations of spatial images, which in turn depend on visual mental images experience. Second, for someone who reached a correct answer during a mental simulation of a mechanical reasoning question, such an individual has an inherent understood (implicit) knowledge of that correct answer. The author stated that representational format, at some stage, is neural transmitted information. However, the conscious experience of mental simulation is different from the hat of applying inference rules.

Interpretation of results

1. The author deduced the suggestion that mechanical reasoning may depend on spatial representations from analysis of individual differences in previous experiments on a pulley system. The experiment is based on the following question: In a pulley system; if the free end of a rope is pulled, will the lower pulley turn clockwise?
2. In dual-task studies, psychologists measure the interference between a primary task (an example is mechanical reasoning) and other secondary tasks depending on working memory buffers. Analysis of the author’s previous experiments showed the “visuospatial” working memory load interfered with mechanical reasoning more than verbal working memory load did. The reverse was also true; mechanical reasoning interfered with the “visuospatial” memory load more than it did with the verbal memory load. Thus, the dual-task method provides another evidence of spatial representations in mechanical reasoning.
3. The results of protocol studies suggest the internal representation is spatial rather than verbal. Thus, protocol studies provide another evidence for spatial representation in mechanical reasoning.
4. Analysis of studies of analog processes in mechanical reasoning provided evidence for the analog course of actions in mechanical inference. In the second study, when individuals trained to use mental images (imagery) of physical representation (analog) and those who are not to do so, suggested that mental imaging is the default strategy to solve the question. Studies used two experimental models to test the dissociation of analog processes; there was no systematic relationship between their answers in both conditions.
5. In reasoning a system of three pulleys, Hegarty found that individuals take more time to conclude the motion of a lower pulley than immediately above. Thus, she assumed that individuals infer the motion piecemeal rather than holistically. The author reviewed another article where the performance of individuals in reasoning a water-pouring question was assessed. Individuals made the correct reasoning when sitting up but not when lying down. Furthermore, when they were told to consider that glasses are full of molasses instead of water, they tilted the glasses further. This leads the author to assume that, participants could neither ignore the effect of gravity nor liquid density. In another experiment, researchers asked the test participants to close their eyes and tilt a glass until each one imagines that water is about to pour out. The researchers repeated the experiment with the participant’s eyes open. Almost all participants made adjustments indicating that a visual image of the extent of the glass tilt accompanied mental simulation.
6. Hegarty reviewed studies that made a distinction between spatial and visual mental images. The studies showed that spatial imagery represents the spatial relationship of the parts of the test system, the location of objects in space, and the way they move. Spatial imagery is not dependent only on visual perception, as one can have an auditory or touch the spatial image. Hegarty reviewed studies on the separation between theoretical and perceptual representation. These studies showed that there are effects, which can be dissociated from explicit knowledge learned by individuals such as the perceptual experience of when individuals see a moving object; they tend to remember the final position as the object displaced. This leads the author to suggest that representations underlying mental simulation have to be more like the experience of understanding a mechanical movement than is a descriptive representation. One potential reason for this is they involve some of the same neural activity working for perceiving spatial characteristics.

Personal assessment

Different studies provide evidence that mental stimulation for solving mechanical reasoning problems is expressed by body movements and maybe dissociated from reasoning by descriptive knowledge. It shows physical representation characteristics. It is not a process of inspecting universal visual image, but it builds in piecemeal. It can be used in association with non-mental images processes as a task break down to its basic parts. The study did not review the nature of mental representations. In addition, it did not review when mental representation is used in mechanical reasoning or what its limitations are as an inference process. However, this study reviewed many psychological experiments and was not limited to the author’s experience. It describes few experiments of testing mental stimulation for mechanical reasoning. It also provided questions for future projects, which may result in a common pool of ideas for future research. The article is important for psychologists researching mechanical reasoning as it provides essential basic ground for further research.

References

Thurston, L.L. (1938). Primary mental abilities. Chicago. University of Chicago press.

Hegarty, M. (2004). Mechanical reasoning by mental simulation. Trends in Cognitive Sciences, 8 (6), 280-285.