Why Abstract Representations Improve Creativity

The last post discussed the dual pathway to creativity, which explains that creative ideas emerge either as a result of detail-oriented thinking and persistence (which gives rises to original ideas within narrower categories) or as a result of cognitive flexibility and mental-set breaking.

The flexibility pathway is very dependent on broad mental horizons and abstract processing. This post describes why abstract processing may help creative thinking and why it may hurt it; the next post will briefly describe the basic practical tools for increasing it.

Generally, the same thing (whether an object, an event, or a person) can be represented more or less abstractly. For example, you could think about eating an orange in more concrete terms, such as taking the fruit, peeling the skin and so on. Yet, you could also think about it more abstractly, for example, as taking care of your nutrition so you could live a healthy and satisfying life. How abstractly you think about something depends on many things, such as psychological distance or the abstractness of the language used to define a problem.

Research shows that more abstract representations normally enhance creative thinking. To illustrate why abstract representations are often helpful, consider the following classic insight problem of a prisoner in a tower:

A prisoner was attempting to escape from a tower. He found a rope in his cell that was half as long enough to permit him to reach the ground safely. He divided the rope in half, tied the two parts together, and escaped. How could he have done this?

To find the solution, you could develop a rather concrete representation of this problem, for example, by focusing on the length of the rope. Alternatively, you could represent it more abstractly, for example, as a problem of safely getting down by using the available materials.

Why Abstract Representations Improve Generation of Creative Ideas

Psychologist Thomas Ward of University of Alabama offers three plausible explanations why abstract representations improve creative insight and generation of ideas.[1] For one, abstraction helps to focus on essential properties. For example, an essential property of the rope is not its length, as different ropes have different lengths and even them same rope can be made longer or shorter; it is also not the shape or the width, as different ropes have different shapes or different widths. Instead, it is the thinner fibers, strings, or wires that make up the rope. Focusing on such essential properties—whether physical or functional— makes it easier to think about recombining them and using in different ways, such as coming up with the best known solution to this problem of unraveling the rope lengthwise and tying the strands together.

Another plausible explanation is that abstract representations may help to restructure the problem, especially by clarifying goals, constraints, and by making underlying assumptions more explicit. All of this makes it easier to challenge the original structure and underlying assumptions.

The third explanation is about the so-called retrieval blocking. Normally, when one specific instance becomes more accessible in the mind, it makes competing traces less accessible (i.e. it blocks their retrieval). So, for example, thinking about one aspect of the rope, such as its length, blocks the accessibility of other aspects, such as its flexible fibers. Abstract representations may be useful precisely because they reduce the dominance of any single specific example, thus allowing for better retrieval of alternative examples.

When Abstract Processing May Hurt Creative Thinking

Despite its advantages, abstract processing sometimes may hurt more than help. As the previous post mentioned, creative generation of ideas sometimes works better with analytical as opposed to intuitive problem solving. So a lot of real-world innovation depends heavily on detail-oriented thinking, which is necessary, for example, in the innovation process to break down a product or service into its parts and then creatively manipulate them by rearranging them, eliminating some, adding others, and so on. Such essentially analytical thinking can generate really creative solutions, but its detail-oriented nature requires concrete representations.

Alternating Use of Abstract Representations

The best practical takeaway for the real-world creativity is to continuously alternate between abstract and concrete representations. In fact, this also applies to many typical problems used by psychologists in laboratory research. Even the generic-parts technique, mentioned in the last post as an example of analytically driven creative problem solving, can be very successfully used to tackle classic insight problems like the prisoner-in-the-tower. Still, even with the generic-parts technique abstract thinking might be very useful. For example, with the 2 steel-ring problem (also mentioned in the last post) abstract processing might be useful for seeing the wick as a simple string. It is the same with all challenging problems—successful problem solving usually requires alternating between more abstract and more concrete representations.

Of course, some types of problems or some stages of problem solving may be inherently more suitable to one or the other. For example, abstract and holistic thinking is very useful in the initial stages of problem solving, particularly in problem identification, and especially when dealing with more difficult and strategic problems, where overwhelming complexity is endemic. Concrete and analytical thinking, on the other hand, is often invaluable for deconstructing the problem and better understanding its specific constraints.

Still, the inescapable conclusion seems to be that the key to creativity lies neither in faithfully abstract thinking nor in devotedly concrete thinking. Instead, the answer appears to be in the alternating use of the both.


[1] Thomas B. Ward, What’s Old about New Ideas? In Steven M. Smith, Thomas B. Ward, and Ronald A. Finke (eds.), The Creative Cognition Approach (Cambridge, MA: MIT Press, 1995) pp. 172-173.


%d bloggers like this: