Monika Dunin-Kozicka (née Chylinska) is a lecturer in philosophy and cognitive science at the John Paul II Catholic University of Lublin, Poland. Her areas of research include pretense, creativity, imagination and counterfactuals.
This week at The Junkyard we’re hosting a symposium on Piotr Kozak’s recent book: Thinking in Images: Imagistic Cognition and Non-propositional Content. On Monday, we began with an introduction from Piotr Kozak. Commentaries follow Tuesday through Thursday.
* * *
Can we be creative in using rulers and thinking in images?
And why even ask such an odd question? First, any time when we use rulers we recreate an old simple procedure — we apply the ruler to the thing to be measured and read the standardized measurement results. We are substantially uncreative then. Second, when we think in images our chances to come up with something new are good — for not only can we operate with images in many ways (e.g. rotating them, combining them, seeing them from a different perspective) but we can also arrive at a new image as a result of such operations. Indeed, we can be original when thinking in images. Why even put rulers and images together in this question then?
For those already familiar with Piotr Kozak’s Thinking in Images, collating a ruler (or a balance) with an image should not be such an eccentricity, for the author of the book — a proponent of the so-called measurement-theoretic account of thinking with images — directly says: “…thinking with images is a skill in using construction rules, comparable to using rulers” (p. 14), and also claims that “the main idea is that images are not like measurement devices. They are measurement devices” (p. 137; author’s emphasis). That is, according to Kozak, what images (including mental images) do is represent the ways of constructing and recognizing objects. Just like a ruler sets up the criteria for identifying the searched distance and producing a line segment, an image of John indicates the means of recognizing John and arriving at John. Specifically, if you know the construction rules that determine possible transformations of the constructed object, and know the construction invariants — so the parameters that are preserved during these transformations (e.g. ruler scale or John’s baldness) — you can be successful in both recognition and production. Then, through the measurement-theoretic lens, thinking in images seems to be mainly about following the rules and sticking with what is already fixed in order to achieve the goals of recognizing and recreating things. If the referent determined by the content matches the target, then an image construction or recognition is correct. The more we are successful at such matching, the better our skills of thinking in images are.
If thinking with images is about matching, recognizing, preserving, reconstructing, knowing the rules and following them, then there seems not to be much room for novelty — which is, however, contradictory to our common belief about the creative prospects of thinking in images. After all, it seems that thinking in images allows us to do more things than using a ruler for it not only lets us recognize an object (as a ruler lets us recognize a length of a line) and arrive at the object (as a ruler allows you to arrive at a line of a certain length) but also to arrive at many different objects based on various operations such as combinations, supplementations, subtractions, replacements, shifts, augmentations, etc. When aiming at such creative or productive goals, we do not typically focus on the construction invariants or on what has to be preserved but rather we search for the possible variations of some parameters to arrive at something new. We may be also willing to unfollow or change some rules so that we can achieve something different from what we were previously offered. For example, we can create something like a collection of images, each of which represents only certain parts of John (e.g. his nose, bald head, glasses, fingers) and each of which is painted on a large canvas, making John unrecognizable when seeing the images one at a time and only recognizable when one sees all the images at once from a far distance. In such a case some standard rules of constructing images are altered to introduce a new variation of depiction that makes it possible to recognize its targets only from a certain optical perspective. It seems that in this case — at least for the author of this style in depiction — thinking with images relies more on searching for some variants than invariants of an image construction. And it seems that in many other creative endeavors the situation is similar.
Should we then abandon the measurement-theoretic account as an account that fails to properly capture our creative uses of images? In fact, I will argue here that we should not, and the reason is that — upon careful examination of Kozak’s proposal — it can be unraveled that, insofar as this account seems inadequate to explain our creative ways of thinking in images, it is more a matter of having different focal points than a matter of not giving a sufficient description. In other words, the measurement-theoretic account has it all; it is only that Thinking in Images focuses on such elements of the account that, at first glance, make it not very pleasing for people like me who seek a good theory of creative thinking with images. Ultimately, however, the measurement-theoretic account appeals to me quite a lot. Why is it so?
Even though in Thinking in Images it is repeatedly stated (e.g. p. 101, 111-13, 116-18, 120-21, 130) that thinking in images relies on identifying those parameters of some space that remain unchanged across different constructions — so the construction invariants — there are obviously other parameters too, that is, the ones that can be changed without collapsing the whole structure. Let’s call them the construction variants. Both invariants and variants are obviously to be spotted in any construction. For example, if we want to draw a cat, we can make any color variations, or leg-length variations, or cat-thickness variations, and we would still arrive at a cat on the condition that we will respect some of the cat-invariants, e.g. the triangular shape of a cat’s ears and a triangular cat’s nose (but whiskers could be of a value too). Thus, identifying the relevant invariants may be useful for our creative thinking and doing because knowing what cannot be changed in the construction informs us simultaneously what can be changed, so that our outcome will be something new but also something comprehensible for others, or reasonable, or useful, etc. After all, if we want to draw an original image of a cat, we want the cat to be recognized in the image. We have to depict some cat-invariants then. But establishing these invariants we will also know what we can alter in the construction: what we can add, cut out, replace, shift, exaggerate, reduce, enlarge, combine with something different, etc. Knowing invariants could be then a key for our creative variations while thinking in images.
Importantly, within our creative and constructive performances we can always set some new construction invariants too. After all, in one of the possible worlds cats can have dog-like ears, dog-like taste preferences, and can even bark like dogs. Now, having decided on such construction invariants, we can see what will necessarily change in this possible world as a result, and also what parameters can be optionally changed in there. For instance, cats will be eating bones, there may be no meow sounds (unless the dogs can do it) and there could be a general cat-dog confusion. And this is how one could actually design any fictional world for any creative purpose.
In fact, later in his book Kozak mentions the issue of the manipulation of the construction parameters that can be cognitively productive and indirectly sets the stage for a measurement theory of creative thinking with images. He writes: “Thinking with images is, using Dennett’s metaphor (2013), 'turning the knobs'. (…) It involves finding out what would change if the construction parameters were different. For instance, what would change if we spatially rotated an object in the mental rotation tasks or how would the world look if we reversed its colours” (p. 167). Now, what I would be much pleased with is a deeper examination of these issues for I believe that the measurement-theoretic account of thinking with images can give us some new good conceptual tools to talk about our creative doings.
One of such tool could be the concept of calibration, which appears in Thinking in Images and which is described as follows:
Calibration is an activity of modeling different processes and testing their consequences for mutual compatibility. (…) An image is an effect of a continuing and co-dependent process of searching for the best way to represent the target and trying to localize the target. It involves ongoing attempts to match a referent to a target by searching for the most precise way to identify the referent. Identifying the referent involves correcting our background assumptions regarding the target. This process is iterative and involves modifications of image construction and representation target. It respects existing iconic conventions but is not determined by them. Most importantly, it describes both the production and interpretation of images. (p. 141-42)
This could be taken as an apt description of a creative process, couldn’t it? The creative calibration would be just a taste of a new conceptual set-up yet to come.
Reply to Monika Dunin-Kozicka
In her perceptive comment, Monika decided to touch on the issue of creativity and imagistic thinking. She is correct that there is an intuitive link between our ability to think in images and being creative. She is also probably right that, although I did not explicitly write about creativity, the measurement-theoretic framework can explain it. However, the nature of the link between creativity and imagistic thinking does not seem clear to me.
On the one hand, in the cognitive science literature, there is no consensus regarding the role of images in reasoning and problem-solving. Some studies report that performance on some problem-solving tasks, such as those involving transitive inference, depends on the ease of imagining the premises of the reasoning (e.g., Clement and Falmagne, 1986; Shaver et al., 1975). However, numerous studies have failed to identify visualization as having any effect on reasoning (e.g., Byrne and Johnson-Laird, 1989; Knauff, 2013; Ragni and Knauff, 2013). Engaging imagery skills can even interfere with thinking.
On the other, there is no consensus regarding how images foster creativity. Most commonly, we point at the informative richness of images. In other words, we hold that images convey much more information than verbal descriptions, which is why they support creative problem solutions.
This line of reasoning, however, is inconclusive. For one thing, it is far from evident that images are informatively richer than verbal descriptions. The description: ‘the number between 0 and infinity’ is infinitely rich in content. For another, even if it was the case that images are informatively richer than verbal descriptions, then it does not imply that they support creative problem-solving. It is possible that since they convey too much irrelevant information, they impede problem-solving, which, in fact, is the case pointed out by psychological literature.
Let me put a bit more pressure on the last point. Creativity is not a novel but randomized activity. If a novice tries to repaint Mona Lisa, they get a novel result but probably not different from a doodle. The creative result should be valuable, too. It is common to explain this value by invoking external factors, such as the evaluation of society. However, one can ask, what is the internal factor of the result that determines such an evaluation. A plausible answer to this question is that creativity is not random but systematic. It discovers rules and patterns that are better than others. The informative richness cannot explain why some rules are better or worse and what they are.
Does this mean we should reject our intuition regarding the link between creativity and imagistic thinking? No, it does not! It means that we should rethink and clarify it. Much depends on how we think about images’ role in creative problem-solving. According to the psychological literature (Beveridge and Parkins, 1987), using an image to foster problem-solving is not enough. It is necessary to point at relevant aspects of the problem. Moreover, much depends on how the cognitive task is formulated. Images are way more helpful if they are used to convey operational knowledge-how than propositional knowledge (Lockhart et al., 1988).
Can the measurement-theoretic framework explain these results? I do not want to defend my approach here directly. Monika has done it much better than I could possibly have. Instead, I want to point out that any competing approaches to imagistic thinking face the same challenge. They must explain the link between the way we use images and creativity. Most importantly, they have to explain the systematic and rule-governed character of creativity fostered by images. I seriously doubt whether they can do that.
References:
Beveridge, M., Parkins, E. (1987). Visual representation in analogical problem solving. Memory & Cognition, 15, 230–237.
Byrne, R. M. J., Johnson-Laird, P. N. (1989). Spatial Reasoning. Journal of Memory and Language, 28, 564-575.
Clement, C. A., Falmagne, R. J. (1986). Logical reasoning, world knowledge, and mental imagery: Interconnections in cognitive processes. Memory & Cognition, 14, 299-307.
Knauff, M. (2013). Space to Reason: A Spatial Theory of Human Thought. Cambridge, London: MIT Press.
Lockhart, R. S., Lamon, M., Gick, M. L. (1988). Conceptual transfer in simple insight problems. Memory & Cognition, 16 (1), 36–44.
Ragni, M., Knauff, M. (2013). A theory and a computational model of spatial reasoning with preferred mental models. Psychological Review, 120 (3), 561-588.
Shaver, P., Pierson, L., Lang, S. (1975). Converging evidence for the functional significance of imagery in problem solving. Cognition, 3, 359-375.