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.

* * *

Piotr Kozak’s Thinking in Images is an insightful book that deals with some of the most fundamental questions regarding imagistic thinking, including: a) what is thinking?; b) what are images?; and c) what, specifically, is thinking with images?

Kozak’s main point is that images instantiate thinking, not accompany it. The answers he provides to the abovementioned questions are the following: a) thinking is a kind of operation (p. 71); b) images are measurement devices (pp. 75, 137) – just like rulers; and c) to think with images is to measure, or to carry out an operation, with the help of images. More specifically, thinking with images is “a rule-governed manipulation of construction rules revealing how something can be perceived in a measurement set-up (…). Thinking with images reveals the ways the world can be perceived and measured” (p. 14). Kozak calls his view a measurement-theoretic account of images.

In this short commentary, I will not question the success of Kozak’s proposal in giving us an understanding of how we can learn about knots or black holes from images, or better understand perceptual illusions and impossible images. It seems plausible to me that the measurement-approach to images could be useful here, as argued by the author. What I would like to focus on instead is asking two questions of Kozak, in the hope that his answers will help us understand his project even better: 1) Who or what is engaging in the measuring operation? And: 2) What is being measured when thinking with an image of, e.g., a loved one? Let me unpack these below.

(1) Who or what is engaging in the measuring operation?

This is a clarification question about the explanandum. Kozak says that the explanandum is the “role of images in the operation of thinking” (pp. 72-73), and thinking is an operation that “locates an item in some logical or physical space, (…) that enables us to localize and orient the item in relation to some parameters, such as volume, pressure, temperature, colour and shape, with the help of which we can recognize the item” (p. 14). But, who is doing the said operation?

This might seem like a moot question, since Kozak is interested in the nature of the imagistic thinking, which targets “the conditions that have to be met if such [thinking] operations are to be possible” (p. 74). He might say that he is not concerned with “localization” of the said process. He divorces the question of the nature of the operation, from engaging in or mastering that operation, and claims that “to understand what thinking with images is, one has to understand what kind of operations images can perform” (p. 75, emphasis added).

Yet I wonder if the target explanandum is set in the right way. Is it correct to focus on the nature of operations, without considering the thinking/operating subject? Looking back at the master question – “what is thinking with images?” – I initially interpreted it as “what is it, for us, to be thinking with images?” My assumption was that I am the one doing the thinking with images. But Kozak does not seem to make that assumption. In fact, in the quoted passage above, it almost looks as if the images are doing the thinking, for they are performing the operations, which constitute thinking. But do images perform the thinking operation? That sounds like a mereological fallacy (Bennett & Hacker, 2007).

Ultimately, I’m not sure whether Kozak endorses internalism about imagistic thinking. Some ideas in the book point to it, such as the exposition of how the imagistic operations work in the two-dimensional model of iconic reference. So, on the one hand, he seems to be talking about operations done at the subpersonal level, in some cognitive architecture – even though he never explicitly attributes the operations to a specific cognitive mechanism or the brain. On the other hand, at times he speaks of a person’s engagement with those operations: “the images we think with are primarily something we do and not something we possess” (p. 70, emphasis added); “[i]mages are the measurement devices we use to inform ourselves about the world” (p. 156).

In the end, since there are no operations in themselves, someone or something must be doing them. The ruler does not measure a square drawn on a piece of paper. I am the one measuring that square with a ruler.

If it is a “cognitive mechanism” that is operating with images, a worry arises that Kozak is not explaining the nature of imagistic thinking but only presupposing this kind of thinking taking place, when framing it as constitutively operational. It’s a re-description; there is no explaining of the thinking going on, or at least, not in an interesting sense. The interesting sense (to me) would be to explain how a person is thinking with images understood as measurement devices.

But, if it is in fact the case that the imaging operation is something that a person is doing, then a problem arises with Kozak’s assurance that “the operational view is not reducible to the dispositional view, and thinking with images is not reducible (…) to a certain skill (Bennett and Hacker, 2003; Ryle, 1949)” (p. 69). I think that this door should be left open. If the operation of imagistic thinking is something that I am doing, then it could be meaningfully thought of as a skill. Skills belong to me, after all, and not to my cognitive architecture. I could be better or worse at performing the operation. There is even an indication that Kozak appreciates skills, when he endorses that “thinking with images is a skill in using construction rules” (p. 14), or that mental imagery “is a skill of perceiving based on procedural knowledge of construction rules. It involves the skill of identifying the parameters of perceptual space in order to localize an object or event” (p. 178).

In short, while we could endorse the claim that “images should be taken as measurement devices comparable to rulers and balances” (p. 75), I would say that they are measurement devices provided to carry out certain operations for us. Answering the question “who or what is doing the measuring with the image?” is not a matter of locating the process, but a matter of fixing the explanandum.

(2) If images are essentially measuring devices, then what are we measuring with an image of a loved one?

According to Kozak, the job of mental images is to show “how a measured object would look in some measurement set-up by representing the ways of identifying this object” (p. 137), or to “localize some state of the world and help us find our way in the world” (p. 140). But it seems to be the case that there are images that are not (or do not need to be) measuring anything. Suppose that I form a mental image of my beloved mother, just for the sake of it. I know where she is, and I know what she looks like. So what does the image inform me of, that I don’t already know? What if I do not need to localize or identify her? For what purpose, then, would I (or my cognitive system) construct a mental image? Can the measurement-theoretic account of imagistic thinking explain such everyday imaging practices, when we, i.e., recall or reminisce or ponder about a loved one? This is the question of the scope of the proposal.

I would say that there can be images that do not measure anything. But for Kozak, that does not seem to be an option. Consider again the idea that images are like rulers, for they are “ways to localize the properties of the world” (p. 138). When I do use a ruler, I do so for a specific purpose – I want to draw a triangle. With a protractor, I want to measure its angles. With a measuring tape, I want to know the size of my waist to buy the right size jeans online. With a stadiometer, my pediatrician wants to measure my child’s height. But when I form a mental image of my mother, or an olfactory image of hot chocolate, or an auditory image of birds chirping in the garden – what exactly am I measuring, and for what purpose? Those images, it seems to me, could have no other purpose than a phenomenal one; they are also done for their own sake. But “[a] measurement for its own sake is a fruitless endeavour”, Kozak writes (p. 156).

Kozak applies his view mainly to understand images like inflation diagrams, maps of London, impossible triangles, or black holes. He does briefly consider Cezanne’s paintings, and a portrait of John, arguing that the latter might be used to localize and identify John (p. 140), whereas “Cezanne’s painting exemplifies a measure we use to recognize this colour [a particular shade of green] in experience” (p. 157). But speaking of localizing colours in experience is not very clear (does it just mean: to experience them?). And does the image of my mother, by the same token, localize my mother in my experience? How to make sense of that, phenomenologically speaking?[1] Perhaps the reach of the measurement-theoretic account of images to mental images we create in everyday context (where it is not so clear that we are measuring anything) could be the topic of Kozak’s next book. 

[1] Maybe it is worth looking at Sartre’s phenomenology here, for inspiration. To Sartre, like to Kozak, the image is not a thing in the head (Sartre 2012, p. 90). According to Sartre, “The image of my friend Pierre is not a vague phosphorescence, a wake left in my consciousness by the perception of Pierre. It is a form of organized consciousness that relates, in its manner, to my friend Pierre. It is one of the possible ways of aiming at the real being, Pierre” (2012, p. 91). The idea here is similar to Kozak’s idea of a picture of bald John, whose “[d]epicted properties represent the way we see John, not the way John is. … [R]epresenting John as bald identifies the ways of arriving at John” (p. 159).

References

Bennett, M. R., & Hacker, P. M. S. (2007). Philosophical Foundations of Neuroscience. Oxford: Blackwell Publishing.

Sartre, J-P. (2012). The Imagination (Original: 1936. L’Imagination). Trans. K. Williford and D. Rudrauf. London: Routledge.

Reply to Zuzanna Rucińska

Zuzanna Rucinska, in her insightful commentary, poses two questions that touch the core of the measurement-theoretic account of imagistic thinking: ‘Who or what is engaging in the measuring operation?’; and ‘If images are essentially measuring devices, then what are we measuring with an image of a loved one?’ These questions are intriguing and well-conceived. However, I must confront them indirectly.

As for the first question, it is not obvious to me that when asking about the nature of thoughts and thinking, we should necessarily consider the nature of the thinking subject. Let me explain.

The intuition that thoughts are existentially connected to some subjects is based on the subject-property model, where properties are linked to some subject they identify. By the same token, thoughts identify the thinker. However, it does not imply that thoughts cannot exist without a thinker. This short remark can be cashed out in at least two ways.

On the one hand, thoughts do not have to be physically bound with the thinker as if there were necessarily some subject possessing a thought. To put it into a slogan, thoughts do not have to be in our heads. Moreover, they do not have to be anywhere. According to the operational approach defended in the book, thoughts are kinds of operations necessarily expressible (although not necessarily actually expressed) by some representations. Operations, such as calculation or construction, do not have to be conducted by some subject to exist (although they have to be conductible). Thoughts can be described in terms of the rules of operations, not some mental objects possessed by a subject.

On the other, thoughts do not have to be logically bound with the thinker as if the existence of thoughts necessarily required the use of thoughts by a subject. Granted, every act of thinking, including measurement operations, requires someone or something to think. However, it does not imply that there are no thoughts outside of using these thoughts or grasping thoughts, as Frege (1984) would say. If we can say that we share the same thoughts, then they are logically independent of the subjects who exercise them. To compare it to a measurement, we can use a ruler differently, but the way the ruler identifies the length is fixed outside the context of its use.

And yet, it does not imply that the issue of grasping thoughts is uninteresting. However, it is not the issue I directly cover in the book.

As for the second question, indeed, we do use images for cognitive purposes. However, I am not holding that every act of using images in thinking is cognitive. That would be a too far-fetched statement. Sometimes we are exercising thoughts just for the mere pleasure of thinking. If I think about my wife, I am not learning anything new about her. I am just evoking her mental image.

Now, every compelling theory of imagistic thinking should be able to explain cognitive and non-cognitive acts of using images. In the book, I am focusing mostly on the cognitive uses of images, but I have never said that they are the only ones. Moreover, I explicitly hold that every theory of imagistic thinking has to address the irreplaceable role of images, e.g., why I am evoking a mental image of my wife and not recalling an extended verbal description of her.

Can the measurement-theoretic framework include non-cognitive acts of using images in thinking? I see no reason why not. According to the 2-dimensional model of iconic reference, images offer us a direct link to the represented object. When I form a mental image of my wife, I directly refer to her. In the Wollheimian sense, I see-in my wife in the image. In contrast to Wollheim (1987), however, I am not holding that the seeing-in relation is primitive and analysable. In the book, I argue that this relation can be understood in terms of recognition-based identification. I claim that recognition-based identification is a kind of reference, different from demonstratives and descriptions, that enables us to localize the objects of our beliefs. In a sense, my wife is present in my mental image of her, just like temperature is present in the number 20 on a thermometer. Having my wife present in my mind seems valuable even though not cognitively fruitful.

References:

Frege, G. (1984). Thoughts, translated by P. Geach and R. Stoothoff. In B. McGuinness (Eds.), Collected Papers on Mathematics, Logic, and Philosophy (pp. 351–372). Oxford: Basil Blackwell.

Wollheim, R. (1987). Painting as an Art. New Jersey: Princeton University Press.