Verified by 
Many of us are comfortable with the belief that analogy plays an important role in the creativity process. But the ongoing debate as regards the existence of mental imagery has some of us feeling uncomfortable in relation to the mental images — the visions and simulations — we see and use during acts of creation. Some of us may simply need the assurance of the ‘experts’, who presumably can tell us whether or not what we are seeing is some kind of ‘epiphenomena’. Some of us hope the pragmatist within us, who often makes good use of mental imagery, will somehow convince our inner grumbling realist that what is useful might actually be real.
Much of our (dis)comfort and grumbling might be resolved by reading John Clement’s outstanding analysis of the problem solving and model construction activities of scientists and students. Clement’s analysis suggests that mental imagery — which can itself emerge in the form of an analogy – is often very useful, and possibly even real. In fact, a close reading of John Clement’s wonderful book, Creative Model Construction in Scientists and Students: The role of imagery, analogy, and mental simulation, may allow some of us to finally admit that, yes, we too have dynamic visions – we too can generate analogies, restructure problems and run simulations in our minds eye.
The reality is that many of the problems scientists and students face are novel problems that require novel solutions and iterative, enquiry based learning. Clement makes good use of novel physics problems in his research. He then uses the protocols of scientists and students who respond to these problems as the basis for his analysis of the nonformal and formal reasoning processes that people bring to bear on a novel problem.
Importantly, Clement’s analysis points to a largely under-analysed and partially hidden world of qualitative and nonformal, but powerful reasoning processes grounded in imagery and concrete, runnable, perceptual motor schemas that are very often used by scientists and students when they face novel problems. This contrasts with the view of scientists as abstract thinkers who use only formal logic and mathematics to solve problems. A further insight from Clement’s analysis is that these nonformal, concrete and imagery-based problem solving processes often precede or work iteratively with more formal, abstract, mathematical processes.
For instance, in a fascinating analysis of how students and scientists went about solving a novel physics problem using springs (i.e., which of two springs, a narrow or a wide spring, would stretch more when an identical force was applied to it), Clement details five stages and multiple mental and physical acts that facilitated the building of a model as to the workings of springs. Some scientists began by attempting to describe mathematical workings based on known formulas, which ultimately failed to distinguish in particular the specific workings of narrow and broad springs. This led many to construct analogies and mental imagery that helped them to build a better understanding of the problem (e.g., comparing the stretch in a narrow and wide spring to the bend in a short and a long piece of wire).
While many of these analogies and mental images were found to have faults (e.g., the bending wire analogy does not help one to achieve the insight that springs work via torsion), the conflict generated as a result of failed thought experiments or failed mental simulations were often catalysts for the developed of further, more refined analogies and mental imagery, which helped some scientists and students to move from simple qualitative descriptions (Stage 1) to tentative explanatory models (Stage 2), and from here to fully imageable and spatiotemporally integrated models incorporating torsion effects (Stage 3).
For the rare thinker who invested their time and energy and intelligence further – Clement himself spent months on the problem – these Stage 3 integrated, imageable models could be translated into models with increasingly geometrical levels of precision (Stage 4). Ultimately, through an iterative process of generative abduction, model evaluation, schema alignment, and mathematization, some scientists (in this case Clement himself) worked to develop a quantitative model of the spring problem that was built upon increasingly adequate qualitative models (Stage 5).
Consistent with the reasoning of Shepard (Shepard, 1984) and Kosslyn (Kosslyn, 1994, Kosslyn et al., 2006), Clement considers imagery a natural way to represent perceptual properties such as shape, relative position, surface texture, geometrical structure, motion trajectories, and certain types of causal relationships (e.g., spatiotemporal constraints on any system of objects). Notably, schema driven imagistic simulation is central to Clement’s theory of imagery based creative and constructive thinking in science. Schema driven imagistic simulation can participate in reasoning operations and in thought experiments, which often extend schemas to cases outside their normal domain of application and tap implicit knowledge in the schema (e.g., spatiotemporal knowledge that emerges in the context of imagistic simulation).
Flexible perceptual activation of perceptual/motor schemas may also contribute to flexible analogy generation or access to new ideas via association; and divergent ideas can also be modulated by imagistic transformations. Dual simulations can be run simultaneously to compare and contrast models, and imagery can be used to represent multiple model constrains efficiently. For example, similar to the processes described by Watson (1968) with regard to his discovery of the double helix, Clement considers how a chemist in trying to imagine what ion could be reacting with a certain crystal surface might first imagine the configurations of atoms on the surface of the substrate; then imagine the size of an ion that could react with it; then choose a candidate molecule of a known shape; then rotate the image of the molecule in several directions to examine if the molecule could fit into the shape of the substrate.
Thus, according to Clement, imagistic simulation is a powerful tool that scientists and students can make good use of: it can be used to generate, evaluate, and modify models in a very flexible and efficient way. Furthermore, while it is often assumed that scientists and students think and act in different ways when solving problems, deeper insights into the ways in which imagery, analogy, and mental simulation are used by scientists and students can help to inform teachers as to the common ground and the common tools that bolster creativity, reasoning, and explanatory model construction in scientists and students alike.
These insights may also have implications for teaching practice. For instance, when teaching science in the classroom, instead of aiming only for students to acquire static symbolic structures, the goal of instruction might be the development of dynamically runnable mental models of both observable phenomena and mechanisms operating “beneath” the observable phenomena being studied. Also, because many of the scientists and students in Clement’s research used hand motions while reporting that they were running imagistic simulations, it may be that depictive hand motions, as an indicator of mental simulation processes, are an important undervalued means of communication that can contribute to both the content and process goals of teachers. In a similar vein, Clement notes how drawings, an obvious means of communication that may enhance students’ use of imagery, are generally underused in the classroom. Overall, Clement believes that imagery enhancement techniques may be a powerful instructional tool but he also recognizes the need for further research in the area.
Clement also recognises the limitations in his method of analysis and he is aware that he infers certain imagery processes from his protocols that remain somewhat speculative and open to critical reanalysis. He highlights many processes that are still poorly understood, including the detailed nature of imagery representations, imagery simulations and transformations, and how these are connected to the navigation and hand-eye manipulation systems; how different types of analogy, imagery, and mental simulation are sequenced and coordinated and controlled by mind and brain; how these mental operations are coordinated with the design and implementation of real experiments (e.g., how thought experiments inform real experiments and vice versa); how these processes can be aided, impeded, or complemented by those from social interactions, and by external representations; how emotions can aid in or interfere with the creative process; etc.
Although Clement argues that a certain amount of volatility at the level of elemental simulations and transformations may percolate upward and support more creative nonformal reasoning, model construction, and model application processes, experimental work is needed to clarify and explain how this process might work. Clement’s qualitative analysis of protocols was certainly generative — he generated a large set of candidate cognitive processes that may be critical for our understanding of creative model construction — and he does an excellent job of working with this large set of cognitive processes to generate hypotheses and a preliminary theory as to the workings of imagery, analogy, and mental simulation in the creative process. However, he also recognises that further experimental work is needed to hone in on individual hypotheses. A truly wonderful book — a wonderful act of creation.
References
KOSSLYN, S. M. 1994. Image and brain : the resolution of the imagery debate, Cambridge, Mass ; London, MIT Press.
KOSSLYN, S. M., THOMPSON, W. L. & GANIS, G. 2006. The case for mental imagery, Oxford ; New York, Oxford University Press.
SHEPARD, R. N. 1984. Ecological Constraints on Internal Representation - Resonant Kinematics of Perceiving, Imagining, Thinking, and Dreaming. Psychological Review, 91, 417-447.
Find Michael Hogan on Twitter
