This is a defense of philosophical imagism. Philosophical imagism attempts to hold the middle ground between traditional imagism and descriptivism. Traditional imagism stems from Aristotle and the British empiricists. Its general features are relatively well known. It accepts the common sense intuition that mental images are easily-identified elements in certain quasi-perceptual conscious experiences that we normally have. It appeals to the observation that mental images of many types appear to us, are generated by us, and are used by us as meaningful tools in our mental lives. Traditional imagism also tends to treat or speak of images according to the somewhat less obvious understanding that they are species of mental objects that are inwardly "seen" or apprehended.

Descriptivism is a relatively recent general philosophic movement, stemming from Ryle, Wittgenstein, Dennett and others who have been concerned to replace our ordinary intuitions about what is meaningful and how meanings are generated. Descriptivism has brought to the fore some of the following concerns: (1) Since there are no images for concepts essential to logic such as "exists" or "and," images can have only a limited role in the generation of logical thought. (2) Images, as a rule, have no apparent intrinsic meaning; what they mean must be provided by a linguistic context, which designates that the image is of some particular thing. (3) If we refer to mental objects as things that are inwardly seen, we invoke not only the concept that there is an inner homunculus that is doing the inner seeing, but that the inner objects themselves have the properties of ordinary visible objects, such as size shape and color.

Philosophical imagism seeks to avoid the errors traditional imagism is prone to. Primarily, it seeks to avoid the implications (1) that because images are inwardly "seen" there must exist an inner homunculus to see them; (2) that because we refer to mental images as objects, we must also mean they have the properties of ordinary objects, such as size, color, or shape. At the same time, philosophical imagism does not give up the common sense and traditionalist notion that mental images are mental objects. They have being and play a role in mental life in a way that is unique to mental, as opposed to physical, objects.

The positive defense of philosophical imagism takes up the first three chapters. In chapter 1, I lay out a common sense account of mental images and the problems it faces. In chapter 2, I describe the history of the imagist and descriptivist approaches in both philosophy and psychology. Chapter 3 is a discussion of the issues involved in developing a theory intermediate between descriptivism and imagism. It relies heavily on the insights of traditional psychology, Aristotle, and Brentano. It provides a metaphysical discussion of how mental images may be understood as mental objects. The metaphysical treatment is inconclusive on several points. It attempts primarily to map out the alternatives.

A negative, or contrasting, defense of philosophical imagism takes up the final two chapters. No new positive ideas are introduced to support philosophical imagism in these chapters. Their purpose is to show what some other views about mental imagery might involve if philosophical imagism is deemed inadequate. I examine two contemporary theories about the nature of mental images, one proposed by Pylyshyn and the other proposed by Kosslyn. These theories derive primarily from an attempt within psychology to develop new paradigms for understanding imagery. Pylyshyn has proposed a descriptivist-inspired theory and Kosslyn an imagist-inspired theory. Both are concerned to show how principles of computational systems using structures and processes can be applied to imagery phenomena in psychology. To some extent, they repeat, within the context of contemporary psychology, the concerns of the imagist/descriptivist opposition within philosophy. The vociferous opposition between these two psychological theories has become known in contemporary literature as the imagery debate.

In chapter 4 I seek to clarify, analyze, and discuss the difficulties with the theoretical side of the imagery debate in psychology. I present Pylyshyn's theoretical case for the epiphenomenalism of imagery in computational systems, discuss the difficulties with it, and show that Kosslyn's theory also results in epiphenomenalism for imagery, despite his claims to the contrary. In chapter 5, I discuss the experimental evidence and the dispute involved in its interpretation. In terms of resolving the imagery debate, the result is inconclusive. I argue that neither side has a clear advantage. I argue this is due to two factors. First, the strictures the disputants have themselves adopted regarding the rules of computational systems and the rules of empirical data gathering have made the debate very difficult to resolve. Second, their evidence reflects part of the ambiguous and elusive nature of images themselves. Like the philosophic descriptivists and imagists before them, these psychologists were each inspired by part of the truth about images but failed to see a middle ground.

I close the discussion with some reflections, based on the two example theories studied, about the nature of computational systems that appeal to or result in epiphenomenalism. I conclude that in theories of this sort, one is more likely than not to lose sight of the fact that images proper occur in conscious states and that both the visual and intentional contents of images can largely be apprehended by us directly without reverting to theories about unconscious computational processes.