I. Introduction
This chapter reviews both philosophic and psychological approaches to mental imagery. Its purpose is to survey the intellectual history of imagery and to introduce the background information and distinctions necessary to understand contemporary issues. This chapter does not purport to be an exhaustive or unbiased history of the topic of mental images. It includes topics and issues that were selected with an eye toward generating a positive account of mental images (chapter 3) and providing a basis for criticism of other contemporary views (chapters 4 and 5).
The history of philosophy does not provide much in the way of treating mental images as a separate topic. Discussions of imagery generally appear in the context of theories of memory, imagination, and thought. I first review some of the influential philosophic theories on these topics, focusing on how these specific kinds of mental images have been conceived. This review allows us to understand the historical roots of two schools of thought that run throughout the history of mental imagery. One school, the imagist, allows a large role for mental images in the operation of the mind; the other school, the descriptivist, understands mental operations primarily in terms of language, allowing mental images a much smaller role in explanations of mental processes. The imagist/descriptivist distinction is crucial to our inquiry and appears repeatedly in subsequent chapters.
In psychology, we find a more direct treatment of the topic
of mental images, since the definition of mental images is
crucial to understanding the nature and goals of psychology
itself. The history of imagery in psychology is part of a
century-long debate about how mental images are to be understood
as objects of the science of psychology. A knowledge of this
history is crucial to understanding the approach of contemporary
cognitive psychology examined in chapters 4 and 5.
II. Imagism
A. Memory Imagery
A review of traditional theories of memory reveals three
important features about the philosophy of memory images. The
first feature is the distinction between conscious memory states
and the explanation of them in terms of the underlying states of
the body and/or soul. This distinction has been maintained, at
least to some degree, by nearly all philosophers, but has often
resulted in confusion between what memory is and how to account
for it. The second feature is the wide, nearly universal,
acceptance of the principles of association as the paradigm for
explaining the structure of memory and the process of recall.
Finally, traditional accounts of memory nearly always invoke the
picture metaphor, an analogy that has caused endless trouble in
philosophy. It resulted, especially after Descartes, in the
homunculus problem, perhaps the most infamous of all
philosophical problems relating to imagery.
1. Aristotle
Aristotle defines memory as "an affectation of perception or conception conditioned by lapse of time" (On Memory and Reminiscence, 449b24). By this he means that memory involves some alteration (affection) of an original perception or conception that is not coincident with the original moment: "the moment of the original experience and the moment of memory of it are never identical" (On Memory and Reminiscence, 451a31). Aristotle emphasizes that this definition has the important consequence that memory cannot be said to be "implanted concurrently with the continuous implantation of the [original] sensory experience" (On Memory and Reminiscence, 451a25). Memory itself Aristotle understands as a conscious state in which the reconstruction of past experience occurs.
For remembering, as we conceived it, essentially implies
consciousness of itself. (On Memory and Reminiscence,
452b28)
For whenever one exercises the faculty of remembering, he must say within himself, 'I formerly heard (or otherwise perceived) this,' or 'I formerly had this thought.' (On Memory and Reminiscence, 449b23)
It is clear that Aristotle means to defend his view against those who confuse the process of implanting, or the material conditions that account for memory, with conscious memory itself.
Aristotle holds that the processes occurring during perception as well as those that account for the permanence of memory are physical in nature. In what is undoubtedly the most famous passage from Aristotle regarding the origination of memory images, he compares the process of original implantation to pressing an image into wax:
The process of movement [sensory stimulation] stamps in, as it were, a sort of impression of the percept, just as persons do who make a wax seal. (On Memory and Reminiscence, 450a30)
Aristotle observes that the permanence of memory appears to be a function of the quality of the matter in the body that is to receive the impression. Various organic states of this matter, such as growth, decay, becoming too soft or too hard, will directly affect its ability to receive and retain an impression (On Memory and Reminiscence, 450b1-12).
Aristotle does not counterpose the physical and the mental in the contemporary (post-Cartesian) sense. Actualized memory, a conscious activity of the soul, does not, in Aristotle's view, involve some different (non-physical) part of the soul. Aristotle assumes that the soul and the body are unified. The relation between the physical basis for memory and its conscious manifestation is to be understood not in terms of separate physical and mental substances, but in terms of the potential and actual states of a single part or "organ" of the soul. All the conscious faculties of the mind are due to a power or potential that resides within some physical organ of the soul.(1) When this potential becomes actualized, the soul achieves a different state, activity, or mode of being, which we call a conscious state.
The question naturally arises which material organ or part of the soul is activated when memory occurs. Aristotle answers it is the same part as is activated in perception, because memory, just as perception and conception, involves a presentation in consciousness (On Memory and Reminiscence, 450a22). This is an important observation, for it shows that from the earliest times there has been some speculation that mental images involve the reactivation of a faculty or organ involved with perception.
Memory, according to Aristotle, is a low-level function of a living soul. All animals with sense perception and awareness of the lapse of time are capable of memory. In man, the capacity of memory is augmented by a higher power, which Aristotle calls recollection. Recollection involves an act of the will which initiates a search for what one wants to remember. Aristotle describes the act a number of ways: a kind of inference, a deliberative act, or an act following a series of previous memories linked by certain relations. The latter description has had profound historical impact, both because it describes the process in physical terms (as "movements" of the soul) and because it defines the fundamental relations among the components of memory as similarity, dissimilarity, and contiguity:
Whenever, therefore, we are recollecting, we are experiencing certain of the antecedent movements, until finally we experience the one after which customarily comes that which we seek. This explains why we hunt up the series [of movements], having started in thought either from a present intuition or some other, and from something either similar, or contrary, to what we seek, or else from that which is contiguous with it. (On Memory and Reminiscence, 451b17)
The discovery of the associative principles mentioned in this passage deserve some additional comment. The associative principles used in recollection have often been cited as one of the first and most important contributions to the science of psychology. Aristotle's actual observation was not, however, as historically important as the uses found for it. We shall see how the British empiricists reinterpreted this observation to support a new philosophic vision of the basic character of mental operations.
There is one other feature of Aristotle's philosophy that has proven to be extremely influential. Although, as we have shown, Aristotle distinguished between the conscious state of memory, or the consciously directed techniques for retrieving memory, and the physical impressions underlying these capacities, he also referred to these physical impressions as "pictures." Just prior to his introduction of the analogy of the wax seal Aristotle says,
[I]t is clear that we must conceive that which is generated through sense-perception in the sentient soul, and in the part of the body which is its seat -- viz. that affection the state whereof we call memory -- to be some such thing as a picture. (On Memory and Reminiscence, 450a27)
The same occurs when he describes the physical aspect of the
process of recollection. He describes recollection as "a
searching for an 'image' in a corporeal substrate" (On Memory and
Reminiscence, 453a5). These references to mental images or
pictures as if they were separate entities in the mind that could
be literally seen have had many unfortunate consequences in the
history of philosophy, especially after Descartes. As a result,
Aristotle is usually blamed for being the originator of the
"picture in the head" view of mental images.
2. Descartes and the British Empiricists
Most philosophers and psychologists through the beginning of this century generally accepted Aristotle's view that memory proper is consciousness of prior events and memory storage is due to physical impressions or changes within the body, as well as his views that the principles of association apparently structure memory. The differences among subsequent philosophers' descriptions of memory result not so much from disagreements about what it is, but how to account for it in a way consistent with alternate metaphysical views about the relation of the mind and body.
Descartes broke with the scholastic view that soul or psyche is a principle distributed in levels of perfection throughout living organisms. Descartes regards animals other than man as mere mechanisms, devoid of soul. Further, he does not accept (in the Meditations) the scholastic distinction between the active (immortal) and the passive (bodily and mortal) intellects. He regards the human mind as a unitary, immortal substance whose function includes not only abstract thinking, but also the functions previously assigned to parts of the soul that were thought clearly to be seated in physical organs, such as memory and sensing.
Descartes faced a problem in explaining memory. On the one hand, he could not abandon the tradition of Aristotle (and common sense as well) that memory is due to physical alterations of matter in the body. On the other hand, his metaphysics of the soul required that consciousness could only exist through an immaterial agency. A connection therefore had to be established between the material site of the original physical impression responsible for memory and the immaterial consciousness. Descartes's solution was to describe the interaction of the mind and body as occurring at a specific site within the brain, namely, at the pineal gland located in the center of the brain. Memory retrieval is described as follows:
Thus when the soul desires to recollect something, this desire causes the [pineal] gland, by inclining successively to different sides, to thrust the spirits towards different parts of the brain until they come across that part where the traces left there by the object we wish to recollect are found. (Passions of the Soul, Bk. I, Art. XLII).
Descartes explains that traces of memory are pores in the brain enlarged by the animal spirits(2) when the object was originally perceived. Once the memory traces were found, there had to be some means of restoring the memory to consciousness. It could not be, as in Aristotle's system, that the reactivation of the material substrate was itself a reactivation of the soul resulting in a conscious state, for that would be to embed consciousness in matter. Descartes's solution was to propose that the pores "excite a special movement in the [pineal] gland which represents the same object to the soul, and causes it to know that it is this which it desired to remember" (Passions of the Soul, Bk. I, Art. XLII). Thus, there is a backward movement from the traces to the seat of consciousness, but the movements themselves do not result in anything until the mind becomes conscious.
There is a similarity to Aristotle, in that the account distinguishes between physical changes and the conscious awareness of memory. But for Descartes, the "seeing" of the image occurs not in the body but in the immaterial soul. Descartes's description, unless it is to be understood entirely as a vague analogy having no explanatory value at all, invokes two suppositions. First, it assumes there is literally some sort of internal image generated at the pineal gland. Second, it invokes the notion of some sort of inner, immaterial "eye" that sees the image present in the pineal gland. Unfortunately, there is little room for a metaphoric or analogical interpretation of this inward seeing, and Descartes's explanations of related phenomena do nothing to help. The literal nature of this inward "seeing" applies, for instance, not just to memory, but to ordinary vision as well. Descartes suggested the spirits had to combine the separate images from each eye in the pineal gland (Passions of the Soul, Arts. XXXII, XXXV) and then invert the image so that the soul could "see" it right-side-up.
The British empiricists sought a different explanation of memory images, one that would avoid Descartes's mind/body interaction problem and would appeal to mechanical principles whenever possible. These philosophers sought a greater continuity between explanations of sensory states and higher, or conscious, states. As a result, they often tended to equate memory with the physical changes responsible for it.
Hobbes, for example, accepted the metaphysical principle that motion "produceth nothing but motion" (Hobbes 1651/1968, p. 86). Since no other principle is available to us empirically, Hobbes reasoned motion must be capable of providing a complete account of the mind. Sensation was evidently caused by physical processes alone, and since no new, non-material results could be introduced into the causal chain following sensation, memory and other apparently non-material conscious states must be nothing more than motion. As an after-image leaves an impression on the senses that is retained after the object has disappeared, so these same motions continue in the body, becoming memory. Hume had a similar theory. He assumed that sensory input in the form of an impression was automatically copied by the mind at the time of the original sensation. Since they were analogous to physical motions, they decayed in force over time, and these less-forceful copies, subsequently revived by the mind, became memory.
Although this approach tended to equate our ideas with physical motions or impressions, the empiricists also betrayed a knowledge that this form of explanation was not entirely adequate to the phenomenological facts. Hobbes recognized that although what he called the "reall effect" of sensory motion was "nothing but Motion," we did not experience these effects as motion but as conscious states such as "Delight, or Trouble of Mind" (Hobbes, 1651/1985, p. 121). Hume found that there were states of mind that he could not trace back to an original impression, such as an imagined shade of blue. In his first edition of the Essay, Locke defined memory as
the power to revive again in our minds those ideas which, after imprinting, have disappeared, or have been as it were laid out of sight. (Book II, Chapter X, para. 1)
In the second edition Locke modified this because he realized it could be understood to imply that conscious states (having an idea) retained their character as ideas even when the mind was not conscious of them. Locke added the following explanation:
But our ideas being nothing but actual perceptions in the mind, which cease to be anything when there is no perception of them, this laying up of our ideas in the repository of memory, signifies no more but this,-- that the mind has a power, in many cases, to revive perceptions which it once had, with this additional perception annexed to them,-- that it has had them before. And in this sense it is that our ideas are said to be in our memories, when indeed they are actually nowhere; -- but only there is an ability in the mind, when it will, to revive them again, and, as it were, paint them anew on itself... (Book II, Chapter X, para. 2)
Locke's account, then, reverts to Aristotle's in both the definition of memory and its explanation in terms of potentials of the mind. Locke, like the other empiricists, realized, even if they did not address, the gap in their explanation of how conscious states of memory relate to prior physical inputs and the physically-based storage of impressions.
I mentioned that the empiricists reinterpreted Aristotle's
principles of association. If memory could be explained by
associative principles and as analogous to a mechanical system,
perhaps the other, so-called higher powers of the mind could be
explained in similar fashion. The reduction of the number of
principles involved would greatly simplify explanations of mental
activity. The empiricists therefore adopted the principles of
association, invoking them as explanations not only of memory but
of the flow of connected ideas in thought itself. This, of
course, was a radical departure from the received wisdom of the
scholastics, who attributed thought processes to rational or
spiritual powers such as "abstraction" or "formation of the
phantasm," rather than mechanical or proto-mechanical principles
such as association.
3. Summary Remarks
Descartes's ontology coupled with the language of the inner picture derived from Aristotle, results in the homunculus problem. This problem is quintessential to the topic of mental imagery, and appears over and over in historical philosophic accounts.
The problem begins in Aristotle's analogy of imprinting a picture in a wax seal. If images are impressed into a material substratum, as a shape is impressed in a mold, or as a picture is printed on a page, how can this picture be recalled unless it can literally be seen? Aristotle's own theory of inward seeing (unlike Descartes's) does not imply any literal inner eye, only the reactivation of the organs involved in vision. To say or imply, then, that what is impressed in the matter of the body responsible for memory is in any sense a picture, must be wrong. It cannot be a true description of the means by which this information was stored in matter.
Descartes's suggestion that there are memory "traces" in the brain (a term still in use today) is perhaps somewhat better in that it does not directly connote that these traces are pictures. But Descartes then reinserts the image into the story of memory recall by having the image reconstituted on an inner viewing area in the center of the brain. The empiricists did not escape the suggestive "picture" metaphor laid down by Aristotle or the inner spectator model of Descartes. If anything, they made the problem worse. Locke's emendation notwithstanding, the overall effect of empiricists' philosophy of memory images was to solidify the spectator model of the mind. It results in the notion that images in the mind are thing-like entities that are retained "as-is" in the mind whether or not we are conscious of them. The conscious mind is then conceived of as a spectator that can inwardly rummage around, searching for the mental images already residing in the mind, as if walking through a picture gallery. If carried to its logical extreme, this spectator in the picture gallery scenario leads to total skepticism. Inner and outer perception become entirely equivalent because the subject stands in the same relation to internal objects as to external objects --- that is, in a condition that allows for the subject to doubt whether appearances are indicative of anything at all. In this case, there would be no reason to believe that any appearances were privileged indicators of an independent reality; the subject would be in a permanent state of doubt.
Because of the homunculus problem, traditional philosophy is
frequently understood to have been a total failure in treating
mental images. It does not advance much upon the common man's
vocabulary of images and how we seem to experience them and store
them as objects, and the attempt to explain how physical inputs
could have psychic results simply results in skepticism and more
confusion.
........... [text omitted]
C. Thought Imagery
If the philosophers of the imagist school and the common man agree, despite the fine points, that memory and imagination images are functional elements of mental life, their views begin to diverge when it comes to the topic of thought imagery. The common man is apt to claim that one thinks in images -- that mental images constitute the very fabric of thought itself. The philosopher, even the ardent imagist, is not as quick to accede.
This is evident if one considers three issues of central historical importance regarding the role of images in thought. These are
1. Are images necessary to thought? Are they sufficient for thought?
2. Do the mental images involved in thought differ in nature from those in memory or imagination?
3. Do the features of thought images limit or constrain thought itself?
It turns out that even within imagism, philosophers are diametrically opposed; there is no consensus view on these questions. There is a convenient way to map out some possibilities, however. If all of these questions were answered affirmatively, this would be called Strong Imagism, or the view that ideas are mental images and that thinking consists entirely of operations using them.(3) The common man, at least prior to reflection, is likely to believe in Strong Imagism. This is not, as Price pointed out, a view any philosopher actually seems to have held, though many philosophers have held some portion of it and some moderns (e.g., Berkeley and Hume) have been accused of endorsing it completely.
I shall divide the discussion of thought imagery into three
parts. The first reviews the historical answers to the three
questions above. The second reviews the essentials of Price's
theory. The third is a summary evaluation of the problems faced
by an imagist perspective on thought imagery.
1. Historical Perspectives on Thought Imagery
Aristotle has the most unequivocal answer to the question of the necessity of images in thought. He states
To the thinking soul images serve as if they were contents of perception... That is why the soul never thinks without an image. (On the Soul, 430b14-17)
The rationale for this view derives from Aristotle's need to explain our knowledge of universals, and universals, in Aristotle's view, are precisely what cannot be represented directly. As we know, Aristotle described the subordinate powers of the soul (sense, memory, and imagination) as primarily passive in operation. These powers are limited to what has been provided by the senses, namely, images of sensible particulars. In order to account for knowledge of non-particulars, Aristotle needed a new power, so he posits the existence of the active intellect, an immaterial agency that would be capable of knowing non-material, abstract objects and producing appropriate images of them. This agency is able to strip away the inessential elements of the images of prior sensible particulars in order to produce an intelligible image, or representation of a universal.
Although the text is somewhat unclear, and this interpretation might be debated, it appears that we have no actual consciousness of the intelligible image. Consciousness of universals occurs because the active mind impresses upon the passive (embodied) mind a token particular that represents the intelligible species.
Aristotle uses mathematical examples to illustrate his meaning. When we imagine a triangle, we imagine it as something specific or determinate in shape, yet we understand that this image is not the object of our thought, which is something abstract, but merely a token. Similar considerations apply to eternity, God, etc., which are thought of by means of images that remain finite (On Memory and Reminiscence, 450a1-15).
What Aristotle actually means, I believe, in the above passage is that the mode of our thinking is due to our material constitution. The organs of presentation, which are the same as, or part of, the organs of sense, enable us only to represent determinate particulars. What distinguishes thought images is not any of their presentational qualities, but simply their causal history. Those which are due to the agent intellect allow us to think, rather than simply remember or imagine. The presence of the image does not complete the thought, since judgment or truth assessment is also necessary (see passage quoted above on imagination). Since we have knowledge of universals, it is this knowledge that it ultimately operative in thought and the image is merely "incidental" (On Memory and Reminiscence, 450a1-15).
Descartes held precisely the opposite view about the necessity of thought images. He argued that imagery is not necessary for understanding. Not surprisingly, this follows from his supposition that the body is not necessary for conscious thought. We already know that Descartes believed memory and imagination images result from bodily influences, and as Descartes wanted to purify thought from the body he claimed that knowledge and thinking could exist without these influences. A piece of wax may appear to us in various shapes, sizes, and even as a solid or a liquid, yet still remain the same substance. No specific image of the wax can correctly represent our understanding of it as a substance that remains the same through multiple physical incarnations. Descartes concluded that it is quite possible to understand what something is without being able to form an image of it.(4)
If one rejects the notion that either immediate rational insight or images planted in consciousness by immaterial powers somehow result in knowledge of universals, one needs a new form of explanation for this kind of knowledge. One possibility is that particular images come to have universal application by lacking certain features -- in being general, vague, or indeterminate in a way that makes them suitable as stand-ins for universals. This, after all, is what is implied by Aristotle's notion that abstraction strips away particular features and creates a set of universal features true of any object of that type. If this feature of vagueness or universality were to be maintained in conscious appearances, the properties of images used in thought would be better matched to their function. Moreover, one would not need to propose that such images were generated by any unknown agency; they could be readily explained as results of automatic mechanical processes such as the superposition and recombination of images already stored in the imagination.
This is the thesis of "general ideas," usually attributed to Locke. The locus classicus of Locke's reputed invention of the indeterminate image is the following:
For example, does it not require some pains and skill to form the general idea of a triangle (which is yet none of the most abstract, comprehensive, and difficult); for it must be neither oblique nor rectangle, neither equilateral, equicrural, nor scalenon; but all and none of these at once. (Locke, 1690, Bk. IV, Ch. 7, para. 9)
There is some doubt that Locke actually believed there were general ideas of the kind that incorporated mutually incompatible features, but the passage is famous because Berkeley launched an attack on it. Locke's name has been associated with the thesis of general ideas ever since. In any event, the thesis of general ideas has proven resilient, in both philosophy and psychology, because of its intuitive simplicity: thought images are different from other images and they need to be different because form follows function. Thought images need to represent general terms, therefore they are general in form.
The thesis of general ideas does not place particular restrictions on the manner of our thinking. It contrasts sharply with another tradition within imagism that does. Suppose we accept it as a principle that since all images originate in perception as particulars, the mind has only these to work with, and no new natures can be introduced. Both Hobbes and Hume attempted to follow the implications of this thesis. Both identified thought with a sequence of "ideas" (both sounds and images) in the mind. All these were determinate particulars, differing only from the original perception in degree of intensity and the combinatorial possibilities supplied by the laws of association.
Hume argues, for example, that we have no original
perceptual image of causality, only the regular succession of one
kind of events followed by another kind of events. Therefore we
have no authentic conception of causality. We have no direct
experience of an image of the soul or of God, therefore the
existence of these as independent substances is a dubious
proposition at best. Another example: consider a mathematical
point. Hume asks, what image can we form of it? Hume's
introspective evidence is that a point may be imagined as a small
dot. He argues such an image has no proper size at all. A small
point can be imagined as easily as a large point, and no matter
what size the image of a point is understood to be, the image
does not admit of division and still remain as an image of a
point. Since our image of a point does not admit of division,
Hume reasons, there is no such thing as an infinitely divisible
magnitude. The properties of the image of the smallest
imaginable magnitude do not allow this conception.
Mathematicians and philosophers are therefore wrong when they
appeal to our understanding of our innate conceptions of
mathematical realities, for this is sheer contrivance.
2. A Contemporary Theory of Thought Imagery: Price
The primary mistake of the Strong View and the mistake made by many traditional imagist accounts, Price argued, was to imply that mental images are concepts or that they resemble concepts. On the other hand, Price defended the fundamental insight of the imagists that thought requires a symbol. The result was a theory that is an amalgam of insights from Aristotle, Locke, and Hume with a revised analysis of the relation between concepts and images.
Like Aristotle, Price rejects the notion that universals exist as extra-mental entities independent of the mind. Since universals do not exist, it is not possible for the mind to "see" or come into direct contact with them. The theory that intellection consists of coming into contact with abstract objects and their relations directly is rationalism in the tradition of Plato and Descartes. Price calls this the Classical Theory of Thinking.
Price defends the Symbolist Theory of Thinking against the Classical. The Symbolist Theory asserts that there must be something that embodies thought, serving as its vehicle. Like Aristotle, Price holds that the only candidates for thought-vehicles are occurrent particulars in consciousness. We observe that there are no occurrent concepts; we find no introspectible mental entities that we can identify as concepts. What, then, are concepts? Price says having a concept is best described as having a set of mental and/or physical dispositions. Having the concept "dog" means that I am disposed to recognize dogs when they are present. It also means that I can employ the concept when they are not present. But in neither case, is my capacity to employ the concept the result of my mental inspection of an occurrent particular that literally is identical with the concept of "dog." This is not to say there cannot be occurrent images that are present when we employ a concept, just that the mere presence of an image is not to be confused with the concept itself. Price describes this nicely in the following passage:
It is dogs, and not the concept Dog, that we recognize; and it is dogs, not the concept Dog, that we think of in absence. The concept is that which enables us to recognize them or to think of them, but neither in recognizing nor in thinking does it present itself to the mind as an inspectable or an introspectible entity. If we say, as I have myself, that it is 'brought to mind' on such occasions, this must not be taken to mean that we have a look at it, so to speak, but rather that it is operative in our minds (indeed psycho-physically also), being aroused from the state of latency in which it is at other times. What does present itself to the mind, and is inspectable, is in recognition a perceived particular. In thinking, even the particulars which we think of do not present themselves to the mind, still less does the concept itself. What do present themselves to the mind are the non-instantiative particulars which we think with (or 'in'), the particulars which we have been calling Symbols: either completely non-instantiative particulars such as words, or quasi-instantiative particulars, whether mental images or other replicas. (Price, 1953, p. 277)
We can therefore understand concepts, Price suggests, as an "image-producing disposition" (Price, 1953, p. 280). Or, stating the relation between concepts and images in terms of what images are, we can say that images are occurrent mental particulars that manifest a concept. It is these mental particulars in and with which we think.
Price's use of the term "non-instantiative particulars" in the above passage needs further explanation. Price understands "instances" of a concept to be the actual objects denoted by the concept. Thus, the concept "dog" is instantiated by the millions of actual dogs in the world. The word "dog" is obviously not an instance of the concept Dog. A word, as a printed item, an utterance, or as inner speech is nothing like an actual dog. It is, however, something we can use in thinking about dogs. It is therefore a non-instantiative symbolic particular. But a mental image of a dog, according to Price, does bear a resemblance to an actual dog. For this reason, Price categorizes images as quasi-instances of concepts. Mental images, or quasi-instances, can be treated as if they were the actual things they are images of.
The reason mental images can be treated as instances is that mental images resemble not concepts, but the actual instances of concepts, as illustrated in Figure 2-1 below. Quasi-instances, or mental images, are symbols by virtue of natural resemblance to natural objects. Both have a similar effect on the mind. "[A] dog-like image or model or picture will bring the concept to mind 'straight off' without the need of any further process of learning or habituation" (Price, p. 265).
Now, in Price's theory, although we do not inspect the mental image, the image is not just an inert appearance having nothing at all to do with the concept. As an instrument of thought, a symbol, it is also an expression of the concept, a tool not just used incidentally, but through which we think. Following the thesis of general ideas, Price believed that the expression or symbol required an appropriate form in order to be able to perform its function. In his view, the question of general ideas is a metaphysical demand. An entity of the right kind is necessary to perform this task if the Symbolist ontology is correct, and the right kind of symbolic particular for concepts is a particular whose features are general or vague enough to serve as a quasi-instance of a set of closely-related items.
In order to further support his position, Price insists that we do have introspective evidence that general ideas exist. Mental images, he notes, can be fleeting or vague as we speak or think. Such vague or indeterminate images take
CONCEPT (DOG)
/ | \
/ | \
/ | \
Actual Actual Mental Image of Dog
Instance 1 Instance 2 Quasi-instance 3
Dog 1 Dog 2 |
|_________________|
resembles
Figure 2-1. Concepts and Instances. An Actual
Instance of Dog can resemble a Quasi-instance
(mental image) of Dog.
place or form in the background of thought, occurring simultaneously with other mental activities in the foreground. As long as we do not focus on them, they remain inchoate instances. If we stop to "examine" them, we stop thought.
Following Hume, Price also holds that any single image,
general or not, is not adequate to a concept, and this, he feels
is yet another indication of the correctness of the Symbolist
position. There are many cases where individual images, whether
determinate or not, would be inadequate as symbols of a class.
We must be able to use other, related images as symbols or our
concept would be limited to a particular quasi-instance. We know
or feel that at any moment other images of the same type as the
one presently serving as a symbol, examples falling under the
same concept, are "present in power," as Hume said. A single,
occurrent image could not have the meaning adequate to a concept;
only the mind's readiness to produce others of the same type
indicates that one is in full possession of a concept.
3. Summary Remarks
In discussing the role of mental images in thought, the philosophers we have reviewed have generally placed images at the service of metaphysical theories, rather than letting image experiences speak for themselves. Aristotle believes the body is necessary, but not sufficient, for thought, so images are necessary but not sufficient for thought. Aristotle concedes that there are many thoughts for which there are no images, but instead of explaining this through some other factor, he employs "images" again -- this time, as rarefied intelligible essences -- that are the result of mysterious processes. Descartes assumes the body is not necessary for thought. Therefore, images are not necessary. These accounts have the advantage of liberating thought from the restraints of the image, but then make the presence of images in consciousness merely a consequence -- and a mysterious one at that -- of a metaphysical theory about the mind/body relation. When the Moderns attempted to bring an earthy realism to images and explain thought in terms of them, they often substituted the properties of images for what were presumed to be the properties of concepts. Since concepts must have had their origin in perception, they made mental images occurring during conception fit the requirements of their theory.
Price's theory attempts to avoid these errors, but he ends by accepting too many of the very problems and equivocations he sought to avoid. According to the Symbolist Theory, thought must be instantiated through occurrent symbolic particulars. Let this be granted. It is an additional assumption, however, that these symbols need to be in a specific form. Price's argument does not show the necessity of images, but just of expressions in general. Perhaps linguistic expressions can do all the work. Price admits (anticipating the descriptivist view) that many of the most important symbols necessary for thought (e.g., logical expressions) are entirely non-imagistic in character. Thus, Price also places the role of mental images in the service of a presupposition about how they must function.
Price is also equivocal about the inspectability of images
and their independence from concepts. He claims images are not
inspected, but says a mental image will "bring the concept to
mind" (see above, and Price, p. 265), as if the image itself
operated as perceptual stimulus. This, of course, is the
homunculus error, and it implies a visual inspection or sudden
recognition of a mental image. Price, like the other imagists,
is unable to avoid phrases that evoke the notion of inner
inspection. There is also an equivocation in holding that the
nature of mental images is that they are occurrent particulars
that express or embody concepts and holding that more images are
present in power. If an occurrent particular does express or
embody a concept, why are others necessary? Why is there an
insufficiency in a single image if the Symbolic Theory is
correct? Hume noted this problem himself when he stated that
particulars serving as stand-ins for general concepts served in a
capacity "beyond their nature" (Hume, 1978/1740, para. 61).
D. The Legacy of Imagism
A basic presupposition of traditional imagism is that percepts are integrated visual wholes that can be stored in the mind. Although it is clear that philosophers have distinguished between what is stored and our conscious internal images, the overall impact of this tradition leaves a rather simple story of how images work in the mind. Original visual impressions are taken in, modified as necessary, and are brought back again as mental images. These forms of internal modification range in degree according to the necessities of each metaphysical account. The modifications range from near zero (memory), to minimal modification through recombination with other images (imagination), to gross modification, such as stripping away of the inessential features (thought). Yet, in each case, the restored image maintains the property of being an image, retaining its resemblance to an original and fulfilling some function similar to the original. Since the images are like the originals our thoughts about them are similar to our thoughts about the originals.
There are fundamental problems with this basic scenario, the foremost of which is that it lends itself to a literal interpretation of inner pictures, and consequently the homunculus objection. The homunculus problem admits of only a few solutions. Among them, hopefully, is the strategy I shall use: 1) identify inner seeing as strictly a metaphor; 2) propose no ontological separation between mind and body, so that, following Aristotle, inner and outer seeing are actualizations of the same thing; and 3) do not identify what is stored as a picture. This view will be supported in chapter 3. An alternate attack on the homunculus problem, proposed by the descriptivists, is discussed in this chapter, below.
A second problem for imagism is that it fails to correctly sort out phenomenological and conceptual distinctions. Hume and Hobbes suggested that the only difference between memory and imagination images is their intensity. They add, almost as an afterthought, that imagination also recombines images, yet as we observed, they provide no plausible explanation of how visual wholes can be combined. This kind of description simply does not address the facts of the matter. It is both a phenomenological and conceptual truth that we distinguish sharply between memory images and imagination images. Intensity has nothing to do with the phenomenological differentiation. To presume so is to rewrite the description of mental phenomena in terms of a forced analogy to physical momentum. Both memory and imagination images can be vague or sharp, intense or faint. Neither are they conceptually alike, as Aristotle wisely observed. Memory comes with the recognition that this is of some prior experience. Imagination images do not. Imagination images come with the awareness that they are only possible representations of reality or that they are decidedly false. Imagination images are active, consciously constructed creations of the will, while memory images are for the most part automatically retrieved with no hint of creative origin. The simple ontology of the British empiricists is inadequate. Some other description of the differences among these types of imagery is required.
Finally, imagism falters when it attempts to explain thought in terms of images. It is obvious that a stream of images alone cannot constitute thought. If that were the case, then thinking would be equivalent to watching a movie that did not suggest thought about anything else, but rather formed the entire substance of thought itself. While imagist philosophers have not embraced this gross error in every particular, neither have they entirely avoided the implication that they thought something like this to be the case. Aristotle's dictum that images are necessary (though not sufficient), Locke's idea that conceptual information is conveyed by means of vague images, and Hume's idea that thought contains primarily the resurrected images of prior sensory experience all imply that mental images are the principal means through which thought is accomplished. The Strong View, then, gives over to the Moderate View, as it must. Once this happens, either we need a more complex symbol account incorporating words or another, non-symbolic (rationalist or conceptualist) view of thought. In either case, some reassessment of images is necessary.
These failings are enough to motivate the search for another
account of our cognitive capacities that eliminates the problems
caused by the initial assumption that images are the primary
elements in perception and cognition. We shall explore how this
possibility is expressed through contemporary descriptivism in
the next section.
III. Descriptivism
"Descriptivism" is a broad term I shall use to designate a family of views in contemporary philosophy of mind. These views derive from linguistic/analytic philosophy since the turn of the century and have been articulated by such thinkers as Ryle, Dennett, and, to a certain extent, Wittgenstein. Historically, there was a close association between the development of descriptivism and the development of behaviorism in psychology. Currently, versions of philosophic descriptivism have been adopted by cognitive psychologists (see chapter 4). The descriptivist view discussed in this section is an amalgam of descriptivist arguments drawn from many sources. Consequently, with the possible exception of Dennett, it does not represent the view of any single philosopher. Despite being drawn from many sources, I believe it is a faithful representation of an important contemporary perspective that has developed as a response to traditional imagism.
Descriptivism seeks to replace images as explanatory tools in the philosophy of mind with linguistic elements, linguistic structures, and linguistic practices. The goal is to show how these, rather than mental images, can be used to explain our cognitive capacities. At the minimum, descriptivism asserts that the primary work of the mind is carried out through linguistic elements; images, though they may be involved in mental functions, are entirely secondary. At the extreme, descriptivism asserts that all the capacities of the mind, from perception through abstract thought, can be explained through the work of linguistic, or proto-linguistic, elements and their interrelations.
Descriptivism relies on a set of more or less standard objections to mental imagery. Among the claims common in descriptivist accounts:
1. Visual mental imagery experiences (i.e., as the traditionalist conceives of them, as mere pictorial presentations) are epiphenomena. They are the causally-ineffective results of other cognitive processes.
2. All (or nearly all) the causally-effective cognitive processes involve descriptions. Conscious imagery processes may "feel" visual and may seem to be involved in our cognitive processes, but that is just an illusion sustained by the imprecise way we have of talking about our experiences. Descriptive processes, some of them automatic, actually underlie imagery experiences.
3. Mental images, conceived of as the traditionalist's "pictures" in the mind, cannot exist. There is nothing that has the properties of real pictures in the mind.
What I call strong descriptivism argues for all these claims.
Taken together, these claims amount to the view that in a
perspicuous explanatory language of causally-effective cognitive
processes, references to mental images could be eliminated. We
shall examine this extreme descriptivist view, since it provides
the most interesting contrast with the views we have already been
exposed to. We begin our account with a review of four often-used descriptivist arguments.
A. Argument 1: No Images for Essential Concepts
A fundamental descriptivist objection is that there
are many important classes of ideas for which there are evidently no images at all. This applies, first of all, to the logical connectives and quantifiers (and, or, not, some, all). Clearly, these concepts are essential to thought. Since there are no images of them, images are inessential to formulation of sentence meanings that rely on the meaning of connectives and quantifiers. By extension, if such individual logical sentence meanings cannot be accounted for, no series of meaningful logical steps can be accounted for in terms of meanings. The result is that images will be irrelevant in logically connected reasoning processes.
Secondly, there are many perfectly meaningful words and sentences that do not seem to have mental images associated with them. Words such as "epiphany," "betrayal," or "obfuscate," although they may evoke images, certainly could not be expected to evoke even remotely similar images among different individuals. Clearly, any images they may evoke are irrelevant to understanding the meaning of the terms. A similar analysis applies to a sentence like, "the average tax rate has changed disproportionately to consumer spending." Since such sentences form a considerable part of all discourses in abstract reasoning, true thought images are evidently very much in the background, rather than the foreground of thought.
The assumption that images are necessary, or ought to be
present in consciousness, descriptivists argue, is just an ad hoc
assumption made in order to bolster the idea that something,
derived with a minimum of alteration from raw sensory input, is
responsible for our ability to reason using universal terms. If
we simply dispense with the assumption that the form of our
thinking must follow a picture-like derivative of sensation, we
can view the matter another way. We can look for clues to the
substance of thought in linguistic practices and logic rather
than visual perception.
B. Argument 2: No Content
What is the meaning or content of an image? Is it obvious or "given," something we may know without learning or the necessity of language? The assumption that images have intrinsic or pre-linguistic meaning seems to lie behind the uses to which images have been put in imagist ontologies of the mental. Price and Hume, for example, seem to assume that if we have a mental image of a chair, we know the image means "chair," rather than "example of furniture."
Descriptivists argue against the notion that images have any intrinsic or "obvious" content. Consider the mental image I would have of Tim or Tom, the identical twins. My images of the two are identical, but when I think of one rather than the other, I surely intend that my image, if it has any function at all for me, be understood as representing only one of them. If the images qua visual or imaginary presentations do not differ, what makes the difference in meaning possible? We must conclude that something other than visual properties of an image determines whatever meaning it has. At root, descriptivists point out, we are forced to admit every sort of image is fundamentally ambiguous, for in and of itself an image depicts or designates nothing specific at all.
Only some context can bring meaning to an image.
Descriptivists argue that this context is supplied by linguistic
elements and the patterns of use that grow up around them. At a
minimum, an image needs to be verbally labeled, as it were, in
order to signify anything at all. Without such a designation,
images are meaningless. As Wittgenstein remarked, not even God
could tell the meaning of a mental image.
C. Argument 3: The Language Trap
This objection holds that we do not in fact refer to any mental object when we speak of mental images. Instead, we refer to concrete, observable patterns of human speech and action. Descriptivists of this school claim that an analysis of language will show that we refer to mental images as a way of indicating that we can pretend, suppose, hypothesize or imagine that something is the case. All of these occur in ways that are publicly manifested in actions and habits of speech.
What we are really doing when we use terms that imply the literal existence of mental images is to employ certain modes of intersubjective communication that indicate our knowledge of human capacities. Although we seem to refer to mental images, this is in fact a replacement or metaphor for more complicated sentences that would explain our expectations and interpretations of the current state of affairs in other terms. For example, to say the artist has a mental image of David in the stone, is a replacement or metaphor for our belief that the artist is capable of carving the stone in the shape of David. These capabilities can ultimately be made directly observable. Such references do not actually refer to inwardly seen and completely private mental objects. To actually refer to private mental entities would be pointless in a world where the meanings of words are established by observable human practices.
The reason that we have lost sight of the metaphorical
aspects of talk about mental imagery is that we have been fooled
by our habits of language. We become victims of the casual use
of language when we begin to believe that the words we use to
communicate actually pick out incorrigibly private items that
exist in some mental space. Since linguistic forms of
communication ultimately rely on common agreement of objective
reality, there can be no place for any meanings that are attached
to distinctively private mental objects. If we actually assumed
that all such mental states were intrinsically private, with no
possible manifestation of public behavior, language would be
impossible. The fact that we have a public language about the
mental realm, then, does not prove its existence as a separate
and inaccessible place filled with special sorts of objects such
as mental images. It actually proves the opposite, that all
suppositions of a private mental world and its content can
ultimately be dispensed with.
D. Argument 4: Wrong Properties
This argument asserts that mental images lack the proper characteristics to be accepted into the class of true, proper, or actual images. Actual images are located in space and time, have specifiable parts, specific relations among those parts, and determinate sensible features such as size, color, and shape. This applies to all actual images, whether schematic line drawings or minutely detailed full-color photographs. Since mental images have none of these properties, they do not count as images at all.
Dennett has made extensive use of this form of objection, often combining it with related objections. It will be useful to review some of his arguments here, since together they result in a comprehensive, strong eliminativist view.
Dennett (1969) makes a distinction between the sub-personal or scientific account of images and the personal (conscious or phenomenological) account of images. In neither of these accounts, Dennett claims, do proper images intractably appear.
From the sub-personal perspective, Dennett argues, mental images are obviously not what he calls "proper," or actual, images at all. There are images on the retina of the eye, but this is the last point in the neurophysiological process which transforms light impulses into vision at which there can be said to be an image. The subsequent neural firings along the optic nerve, continuing to other portions of the brain which are responsible for seeing or imagining, have nothing in common with proper images.
A variation of this objection is Dennett's version of the homunculus argument. Dennett argues that in order for something to function as an image, it must be perceived. Images cannot, therefore, be the product of the process of perception; they can only be things that exist independently of us -- things we can literally see. If the perceptual process literally produced images, these would be images of things in the mind, and it would be those images that we would inwardly see and not the external things themselves. And, so the argument goes, since the internal images need to be literally seen in order to be recognized, there would have to be an internal agent, a man-analogue or homunculus, that could look at the images in order for recognition to occur. But this man-analogue could not perceive any images unless it too had an internal homunculus that could perceive internal images. We have the beginning of an infinite regress. The point of the argument is that we cannot introduce a new, intermediary entity into perception and call it (the intermediary) the object of perception without losing the essential idea of what it means to perceive: for the subject to see, directly, the object of perception.
The defender of images will reply that if proper images cannot be said to exist in the mind from the sub-personal, or scientific standpoint, there is surely an intractable sense in which they must be granted to exist at the personal, or phenomenological, level. So much of our personal experience is structured by images, it seems ludicrous not to grant them existence. Dennett dispatches images from this arena by arguing for what he calls the "descriptional view of awareness." He argues that seeing and imaging actually involve descriptive rather than imagery-based processes. This, he claims, can be demonstrated by examples. If we imagine a man, we have no need to imagine any specific details about the man. A mental image may specify particular features, such as hair color, but it need not. Mental images are not like actual pictures that must, in virtue of their being pictures, show such details. Mental images are more like descriptions in that we can specify as many details as we want, but the failure to do so does not change their basic character. Imagining, then, is actually the process of providing ourselves with descriptions, rather than images.
The illusion that there are mental images to be inwardly seen or inspected occurs because the process of creating these descriptions is often automatic. As quickly as we imagine something, we provide an inward description, automatically, that provides that actual content to our imagining. The so-called "image" does nothing at all, because it is actually a form of inner description. The same account of automatic descriptions extends to perception. Dennett says, "seeing is rather like reading a novel at breakneck speed" (Dennett, 1969, p. 57). There is an illusion of visual content, but the actual content of perception is conceptually ordered in the manner of language. Dennett concludes that mental images we experience have none of the properties of actual pictures or images. The actual modes of our thinking and perceiving at the personal level can be accounted for through descriptions.
As a final argument against the existence of images in the mind, Dennett gives his own view on the issue of whether mental images must consist of determinate parts in specific relations or can also be vague or indeterminate. (We have already seen that the same issue has been at stake at least since the time of Locke). Dennett argues that we know from direct experience that a mental image is vague and lacking in distinct, numerable parts. Dennett's example is the mental image of a tiger. How many stripes does it have? Dennett claims there is no answer to this question because a mental image of this type is inherently indefinite. No specific number of stripes is represented. Since this "mental image" fails to meet one of the criteria for being an image, it is not a proper image at all.
Dennett concludes that since mental images have none of the
required properties, they are not images in the proper sense at
all. Mental images, by this criterion, cannot be said to exist.
Dennett leaves open only one remote possibility for the
"existence" of mental images. If there are mental objects that
behave as if they had the properties of proper images, then there
may be said to be functional images in the mind. This is the
contemporary imagist view defended by Kosslyn, the subject of
another chapter.
E. The Legacy of Descriptivist Arguments
Objections 1 and 2 make valid points. As some imagists (e.g., Price) have themselves observed, there are indeed no images for the logical connectives necessary for deductive thought and the synthesis of concepts. It is also true that images, considered as mere visual presentations, have no intentional content.(5) Traditional imagists did, in fact, seem to assume that mental images could derive intentional content by means of visual similarity to perceptual objects that already have a self-evident meaning. Descriptivists are right to point out that there is no such thing as an image with a self-evident meaning, much less any string of images that could itself carry a meaningful thought.
Arguments 3 and 4 are far less convincing. Accounts based on these arguments place one at odds with the form of our conscious experiences. No one, I submit, experiences seeing as high-speed reading. If some analog to high-speed reading is what is actually occurring as an unconscious process, and this is invoked as an explanation of seeing, it places our whole relation to the external world on another footing, because language, not the operation of the senses, will ultimately structure our access to the world. In any event, the descriptivist account does not capture the nature
of the conscious awareness of seeing, and, as a result, it does not address the nature of either visual memory or visual imagination. Seeing is ineffable; the experiential quality of it cannot be explained or described to a blind person. Seeing means being conscious of a unified, non-verbal, presentational whole with visually inspectable features. If, as seems self evident, visual memory and imagination are similar forms of conscious awareness that involve an ineffable visual component, some further development of the characteristics of the images involved is necessary.
Besides its objection to images, descriptivism is apparently also aimed at eliminating the notion that there is private, or exclusively mental, access to the elements of thought. Descriptivists argue we cannot make up, as it were, what the elements of thought will be. As Wittgenstein argued, one cannot inwardly say one word and inwardly mean another. I find that this argument also fails to do justice to the imagist perspective. There is, undoubtedly, a confusion in the imagist perspective between the elements of thought and the means of thought. These sometimes appear to be equated. The descriptivist argument applies against construing images as the elements of thought, but it does not address the sense in which images serve as a means. As a means to thought, images can be suggestive, rather than definitive. We can hold in mind one image that serves as a symbol of a meaning that itself has no definitive symbol. The image of a lion can be used in thinking of the class of lions, but also for the class of animals, for which there is no image.
The descriptivist argument even fails to do justice to the indefinite nature of linguistic symbols. Sentences like "I am going to the bank" have an indefinite meaning (river bank vs. financial institution). The mere occurrence of the symbol element "bank" does not define the meaning of the sentence, whether it is uttered in public or is an "inner utterance." In fact, according to the common man's view, failure in communication between interlocutors is often due to them not having the same mental image in mind when a given word is spoken. It would be the corresponding mental image of the intended kind of bank that would remove the ambiguity of such a sentence.
The descriptivist arguments provide some correctives to the imagist theory, but they hardly provide the basis for a convincing alternative theory based on common sense. The privacy of mental images is not a reason to suppose they do not exist, only that they cannot become the entire basis for shared conventions of meaningful communication. This is already a point conceded in the Moderate Imagist View. The descriptivist arguments assume that having descriptive information and having visual information amount to the same thing. This is surely false. Any list of facts about Fred and Sam is no substitute for seeing Fred and Sam. Descriptivists who equate seeing with high-speed description formation try to make experience conform to a theory. Finally, while imagism falters in explaining thought, descriptivism falters in explaining the presence of images in the lower powers of memory and imagination. If all the information taken in during perception were really descriptional, there would be no reason for it to "reappear" as an image in memory or imagination. Even if the need for the reappearance of the image could be established, the generative process of image construction would require still further explanation. The descriptivist would be forced to provide an elaborate set of rules by which descriptions can be transformed into images (a problem for Pylyshyn and Kosslyn, as discussed in chapters 4 and 5). Either, then, the descriptional view of perception is wrong, or descriptions become transformed into images in memory and imagination for no apparent reason.
IV. Psychology and Mental Images
A. The Questions for Psychology
In the nineteenth century, new developments in neurology, medicine, and physics laid the foundations for understanding psychology as a science separate from philosophy. Part of this new science was speculative, but it was also bound by a new paradigm of science that demanded the use of experimental data even in those matters that had traditionally been thought to be strictly metaphysical or religious in nature. The rapid development of the physical sciences during this period led many philosophers and psychologists to hope that the nature of mental processes could be illuminated by empirical, rather than speculative or metaphysical, methods. It is therefore reasonable to understand our guiding question about the nature of mental images for psychology to be a variation on the original. We want to know if psychology can answer this question:
Is there an empirical means for studying the nature and role of mental images?
This question cannot be answered by a simple "yes" or "no," for
it spawns innumerable other questions. Does the empirical method
include introspection or only the gross behavior of subjects?
Does the gross behavior of subjects include verbal reports of
their imagery experiences or only other sorts of bodily
movements? Finally, what is the scientific archetype for
psychology and how do mental images fit into this archetype? Is
psychology an offshoot of philosophy, a "soft" science
incorporating irretrievably metaphysical assumptions, or is it a
separate, "hard" empirical science like physics? How do these
scientific paradigms affect the way in which psychic phenomena
such as mental images are understood to be a part of empirical
psychology?
...............[text omitted]
V. Summary: Remaining Problems for Mental Imagery
We have examined proposed explanations to the puzzles encountered in historical accounts of mental imagery from both the imagist and the descriptivist perspectives. Imagist accounts of memory images as stored pictures can result in the homunculus problem. The imagist tradition of imagination and thought imagery does not explain the creative power of the imagination and creates a confusing picture of the relevance of mental images in thought. The Strong Imagist View, that thought is identical with having mental images, must be abandoned in favor of a Moderate Imagist View, which includes images among verbal symbols as vehicles of thought. The unsettled state of affairs with traditional imagism prompted us to examine the descriptivist school. Descriptivist arguments demonstrate that the images, as pure pictorial data, cannot be the carriers of meaning nor is it correct to speak of them as if they had the physical properties of actual images. Descriptivism solves the homunculus problem by insisting that "visual" mental images are not actually visual. There is no inner visual scrutiny of mental images as there would be if mental images were like actual external stimuli. While we agreed that meaning cannot be intrinsic to a picture, we found that descriptivism gives us a distorted picture of the world we experience. It solves the homunculus problem at the expense of understanding what it is like to see. Like imagism, descriptivism builds a story of mental contents from the ground up, starting in perception and ending in thought using a single metaphysical archetype.
The question of the place of images in the science of psychology at first appeared to be a different and ancillary question. Our review of the history of psychology from the standpoint of the fate of mental images revealed that psychology suffers from a similar problem in finding a definition for mental images -- swinging back and forth between common sense and eliminativist views with no resolution. Any contemporary philosophy of mental images, however, should not dismiss psychology for its failure to resolve the issue any more than it should dismiss its own history. It remains to be seen if psychology can offer meaningful perspectives and meaningful empirical data for philosophers.
The arguments we have examined so far suggest that mental
images are neither just pictures nor just descriptions. Perhaps
a compromise view in which mental images are understood to
incorporate features of both images and descriptions is possible.
I examine this possibility in the next chapter.
1. There is room for alternate interpretations of Aristotle's text on this point. See Modrak, 1987, for a comparison of Aristotle to contemporary functionalism. There is also an immaterial intellect in Aristotle's system, but it alone does not, apparently, have the power of conscious thought. The role of the immaterial (agent) intellect and its images is discussed in the section on thought imagery, below.
2. The animal spirits were also material particles, fine enough to travel through the nerves.
3. I owe the distinctions between Strong Imagism and Moderate Imagism and some of the associated points to Price, whose insightful book (1953) is probably the best contemporary attempt to understand traditional philosophic imagism.
4. This conclusion is true, but it is sometimes thought that Descartes also demonstrated the general uselessness of images in thinking. His famous example is comparing thinking of a triangle to thinking of a chilliagon. He noticed that we cannot form a detailed image of a chilliagon, but we understand it nonetheless. His example demonstrates the distinction between thought and imagination, but not that images must be utterly useless in every case. Less complex figures are more easily imagined, as Descartes pointed out. These might well be used to think about the properties of the figure, even though we still distinguish between just imagining a figure and thinking about it.
5. This is not to admit that descriptivists are correct in implying, as they often do, that mental images have no content. Mental images do present qualitative content, a point taken up in the following chapter.