Simple Remembering

Abstract

Dretske has provided very influential arguments that there is a difference between our sensory awareness of objects and our awareness of facts about these objects—that there is a difference, for example, between seeing x and seeing that x is F. This distinction between simple and epistemic seeing is a staple of the philosophy of perception. Memory is often usefully compared to perception, and in this spirit I argue for the conditional claim that if Dretske’s arguments succeed in motivating the posit of simple seeing, then parallel arguments should equally motivate a posit I call simple remembering. Simple remembering would be a conscious form of memory about an object or event which is prior to and independent of any beliefs the subject may or may not form about the object or event simply remembered.

1 Introduction

I often have the experience while running of hearing my breath: inhale, exhale inhale, and so on. My breath is “there” in my experience—I hear it droning on, and in fact it is rather loud—but I do not usually form any conscious beliefs about it, like, ‘ I am really breathing hard now,’ or, ‘My exhales seem longer than my inhales.’¹ Of course not. My auditory awareness of my breath does not usually take the form of conscious beliefs about my breath. My mind is usually elsewhere: on the tasks that await me at home, on the approaching hill, on the car that has just passed by. Nor does hearing my breath on a run imply that I shall come to form some conscious beliefs about my breath later on. I usually have better things to do than to reflect back on such matters. This is perfectly consistent with the fact that if my breath becomes relevant, either during the run or after—if I find that can’t “catch” my breath, for example, or if I learn later that I have run through a patch of dangerous air-pollution, or if I find myself writing a paper for which hearing one’s breath makes a good example—I am able to quickly and easily form beliefs, perform inferences, and deliver judgments on the topic. The point I am driving at is that I do not have to form such conscious beliefs or judgments just because and whenever I hear something. We can call this sort of sensory experience that is itself independent of and prior to belief, simple awareness, although it has been called by other names in the literature such as objectual awareness, and non-epistemic awareness. Objects and events that are in the “background” of our experience often command only this simple sort of awareness from us.

This simple sort of awareness of objects and events seems to occur in all of the sense modalities: seeing, smelling, touching, and so on. In regard to the visual sense, for example, I gather that there is a practice in Japanese Zen meditation in which the eyes are kept open, gazing softly at the floor in front of one (Kapleau, 1970, p. 30). Presumably, it is inevitable that conscious beliefs occasionally arise, even for seasoned meditators. Nevertheless, the aim of “non-judgmental awareness,” intent attention on a focal point without conscious beliefs, judgments, or inner commentary, is also quite achievable. One’s eyes may be open, one’s mind alert, one may see the floor, without coming to consciously believe facts about the floor. One sees the floor simply.

Fred Dretske (1969, 1978, 1979, 1981, 1993, 1995) has probably done more than anyone else in philosophy to show that we can see, and hear, and touch, and smell, and taste things without forming beliefs about them.² Memory is often usefully compared to modes of sense experience. In this spirit, I want to argue that the sorts of reasons Dretske gave for thinking that there exist simple forms of sensory experience should move us equally to posit a simple form of remembering. A bit more explicitly, I shall defend the conditional claim that if Dretske’s arguments for simple seeing succeed, then parallel arguments ought to succeed in motivating simple remembering. Many have thought that Dretske’s arguments for simple seeing do succeed. After all, the distinction between simple and epistemic seeing is now a staple of the philosophy of perception. I am partial to these arguments too, but in this paper, my primary concern will be to show how we can transpose the arguments to the case of memory, rather than to defend them from all comers.

This project may be of interest for a number of reasons. One reason is basically historical in that some major figures seem to have held that remembering an object or event implies forming a belief about it.³ The discovery of simple remembering would presumably show them wrong, (although I leave it for more expert interpreters to say). A second source of interest may lie in the surprising nature of simple remembering. Simple seeing may be fairly intuitive—we see more objects and events than we form beliefs about. In our everyday experience of the world, multitudes of objects exist and events play-out whether we make judgements about them or not. It’s much less intuitive that we consciously remember more objects and events than we form conscious beliefs about. Isn’t remembering an object or event just a matter of retrieving a belief about it from memory? A third source of interest in simple remembering may derive from its potential uses as a tool of theory. Dretske’s distinction between simple and epistemic awareness yields different “levels” of content (Pacherie, 2000). If different mental states (or events) may possess different levels of content, we can tell a more progressive story of how sensation, caused by an object or event, can give knowledge about it; via simple awareness. Perhaps the posit of simple remembering can allow for similar epistemic progressions from retrieved information to full-formed knowing. The first step, my circumscribed task in this paper, is to transpose Dretske’s arguments, making adjustments where relevant.

The plan for the paper is as follows. Section ii explicates Dretske’s distinction between simple and epistemic states. Section 3 introduces Dretske’s argument from consistency for simple seeing. I adapt it to the case of memory in Sect. 4. Section 5 presents Dretske’s continuity argument for simple seeing, which I adapt in Sect. 6. Section 7 presents arguments based on seeing without noticing and remembering without noticing for simple seeing and remembering (respectively). Section viii presents Dretske’s argument for simple seeing based on processing limits of the subject. I submit that this argument arrives at simple remembering as an intermediary conclusion. The essay concludes in section ix.

2 Dretske’s Distinction: the simple, and the epistemic

We find the distinction between epistemic and simple seeing throughout Dretske’s work, from his first book to his last, as well as in many papers.⁴ Quassim Cassam (2008) provides an excellent short description of the distinction, which I shall use as a framework for our discussion.

The distinction between seeing a barn and seeing that there is a barn nearby maps on to Dretske’s distinction between non-epistemic or ‘simple’ seeing and epistemic seeing. In epistemic seeing one sees that something is the case. The main characteristics of this kind of seeing are that it is propositional, factive, and has epistemic implications: it implies something about what the perceiver knows and about his conceptual resources. If S sees that there is a barn nearby then there is a barn in the nearby. If S sees that there is a barn nearby then he must have the concept barn. Finally, if S sees that there is a barn nearby then he knows that there is a barn nearby. Simple seeing, in contrast, is non-propositional, non-factive and lacks the epistemic implications of epistemic seeing. Simple seeing is ‘the seeing of objects and things – not facts about these things’ (Dretske 2000, p. 98), and so is not constrained by one’s conceptual resources. (Cassam, 2008, p. 217, original emphasis)

So we have here at least three main points of contrast, although we will tease out what Cassam calls epistemic implications a bit further. Let’s go through it all more slowly, beginning with epistemic seeing.

Suppose you see that Ava is waving, (that you see it in the epistemic manner of seeing). If you see that Ava is waving, then you know that Ava is waving. The implication of knowledge is one feature of epistemic seeing.⁵ Supposing one knows that Ava is waving, then it must really be the case that Ava is waving. For it is generally recognized that only truths are knowable. Therefore, a second feature of epistemic seeing is factivity, that its content is true. A third feature of epistemic seeing is that its content is propositional. If in seeing something epistemically, one entertains a content that is true, this content must have the sort of structure which it makes sense to evaluate as true or false. In the simplest case, (an atomic case, one might say,) the content needs to have a structure that establishes something in the role of the subject of the thought, (in this case Ava,) and presents it as instantiating a distinct property or feature, (in this case the property of waving). Epistemic seeing is thus propositional in the sense that its content has this logical structure.⁶ A fourth feature of epistemic seeing is that it implies that the subject believes what they see. This connection is entirely familiar, since propositional knowledge is always taken to imply belief. But if the fact that S sees that Ava is waving implies that S believes that Ava is waving, then S must have and apply the concept of waving.⁷ So a fifth implication of epistemic seeing is that it is conceptual in the sense that the subject must have and apply the concepts needed to specify the proposition they see, believe, and know.⁸ To sum up then, epistemic seeing is (1) knowledge implying; (2) factive; (3) propositional; (4) belief implying; and (5) conceptual. Let us now turn to simple seeing.

A number of reasons, (some we will now preview,) recommend that a subject might be able to see something but fail to see that-, believe that-, know that- any particular fact obtains about the object or event that is being seen. For example, although one might see what happens to be Ava waving, and although one might know Ava and what it is to wave, one might still fail to bring the event under these concepts. One might fail, for example, to recognize Ava as Ava or her waving as waving. Nothing in the content of seeing Ava wave compels a particular belief about the event one sees (consult discussion of the consistency argument, Sect. 3 and 4). Moreover, it is clear that human infants and some non-human animals can see Ava waving as well as you can, their eyesight is good, but infants and animals presumably do not believe that Ava is waving, having no conception of what it is to wave (see discussion of the continuity argument, Sects. 5 and 6). Quite in line with these thoughts, Dretske frequently introduced simple seeing as a way of seeing that satisfies a negative criterion.

Negative criterion: “for any proposition, P, the statement ‘S sees D’ does not logically entail the statement ‘S believes P’,” (1969, p. 6).

In later work, Dretske gave a negative criterion that seems to apply to simple modes of awareness in general.

Negative criterion*: For all things x and properties F, S is conscious of x does not logically imply S is conscious that x is F (1993, p. 266).⁹

The negative criteria dictate that a given instance of seeing an object or event simply, does not compel the subject to form any particular belief about the object or event being seen.

The negative criterion may appear somewhat too broad as it is originally stated, too broad to precisely describe the clearly intended outcome of some of Dretske’s arguments for simple seeing. For without supplementation, the negative criterion seems consistent with the proposal that seeing something simply does indeed imply that one form some belief or other about the object or event seen. On this reading, the point of the negative criterion is simply that that there is no one particular belief that is compelled by the simple way in which one sees the object or event.¹⁰ Dretske explicitly rules out this interpretation in later work.

“We have, on one hand, the claim that:

(1) Simply seeing X is compatible with no beliefs about X.

On the other hand we have such claims as:

(2a) Simply seeing X is incompatible with beliefs about X.

and

(2b) Simply seeing X occurs only if, as a matter of fact, the seer has no beliefs about X.

It is (1) that gives expression to the relevant view about simple seeing. At least it gives expression to the only view I have ever propounded and thought worthy of defense – despite persistent efforts to interpret me otherwise. (1979, p. 3).

Dretske is concerned here to distinguish simple seeing from some other possible views about forms of seeing that in some sense or other require the absence of belief. For present purposes, however, the important point that this passage makes is that simple seeing is supposed to be possible in the absence of beliefs about the object or event simply seen. Moreover, despite the initial negative criterion’s openness to misinterpretation, Dretske is, of course, correct in claiming that the view expressed by (1) is the view he defended in Seeing and Knowing. Although Dretske’s arguments are subject to interpretation, it seems clear that claim (1) is the intended outcome at least of arguments based on seeing without noticing (see Sect. 7). Briefly, it is apparently quite possible for one to see, for example, all of the books on a shelf, without noticing and hence without forming conscious beliefs about each book one sees (Cf. 1969, p. 11). Moreover, it also seems to be quite possible to walk down the street in such a preoccupied state of mind that while one sees objects and events, most of what one sees passes by without notice, belief, or judgment (Ibid., p. 13). Moreover, it may even be possible for one to look for something, one’s cufflink in a drawer, for example, and to see it among other things, without noticing it, without seeing that any fact obtains about it, and sadly without finding it (Ibid., 19).¹¹ In later work, Dretske argued that subjects may see objects under processing limitations that prevent the formation of conscious beliefs about each object one sees (see discussion in section viii). These thoughts suggest that it is possible to see an object or event, for it to be presented within one’s visual experience, without one’s experience being so structured to establish it, in its own right, as either a subject, or as feature predicated of a subject, in a propositional content one entertains, believes, and knows.

Where does this leave the notion of simple seeing? Because simply seeing X is compatible with no beliefs about X, simple seeing clearly is not belief implying (with respect to the object or event simply seen). Moreover, since knowledge about X requires belief about X, which simple seeing does not guarantee, simple seeing is also not knowledge implying (with respect to the object or event simply seen). Moreover, since it is apparently necessary to notice something to consciously entertain a proposition about it, and simply seeing X does not imply noticing X, simple seeing is non-propositional (with respect to X). Moreover, since only mental states with propositional content can be factive, simple seeing is non-factive.¹² Lastly, it seems that if a content is conceptual, then it is propositional, since to apply a concept is to apply a concept to something, and this is what yields propositional, function-argument structure. So, if simple seeing is meant to be non-propositional, as indeed it seems, then simple seeing is non-conceptual. It bears repeating that all these five features of simple remembering are relativized to the object or event simply seen.¹³

We are now in a position to articulate what the analogous notion of simple remembering would amount to. Like simple seeing and other simple forms of perception, simple remembering would be a conscious form of awareness, in this case, an awareness of an object or event previously experienced. It would be not knowledge implying, or belief implying, not factive, propositional, or conceptual. It would be, we may suppose, a form of remembering that satisfies a negative criterion that is modeled on Dretske’s.

Negative criterion for simple remembering: It is possible for a subject to remember an object or event simply, and to fail to have a conscious belief about the object or event.¹⁴

In other words, no conscious belief whatsoever about an object or event is mandatory for a subject to remember that object or event simply. Like other forms of simple sensory awareness, simple remembering would be compatible with having or forming conscious beliefs about the object or event in question. And like simple mental events in general, simple remembering would also be compatible with the absence of such beliefs. The point to emphasize is that simple remembering would not itself be, as regards its content, propositional or conceptual, and as regards its attitude, it would not amount to a believing or knowing of such.

Dretske also characterizes simple seeing positively, (that is, not merely as a way of seeing which lacks the five features of epistemic seeing which we have discussed). In Seeing and Knowing, the positive characterization of simple seeing holds that a subject simply sees an object or event, D, just when it is visually differentiated from its immediate environment by the subject (1969, p. 20). Visual differentiation is itself a term of art which denotes a state of affairs “constituted by D’s looking some way to S and, moreover, looking different than its immediate environment,” (Ibid., original emphasis). An object or event looking a certain way to a subject consists in a differentiation in the subject’s experience, as Dretske often describes it.

The fact that simply seeing D does not bring to mind a conceptual propositional presentation of D does not indicate that simple seeing, and other simple forms of sensory awareness, lack epistemic importance, far from it.¹⁵Simply seeing an object or event forms the basis in Dretske’s epistemology for seeing facts about the relevant object or event. Remaining mindful of space, I can only sketch Dretske’s picture. In Seeing and Knowing, the proposal was that when one sees a fact about an object, that x is F, by seeing the x, and not by, for example, seeing its reflection on the water, its image on the television, a report on it in the newspaper, its measurement by an instrument, an X-ray, a fuel-gauge, etc., one must see x simply, it must be true that x is F, it must be true that x would not look L to one if it were not F, and that believing this, one must take x to be F (Ibid., 78–93). Whatever the defects of this account—some have charged, for example, that it is too demanding to suppose it necessary that x would not look L to one unless it were F (Jackson, 1977, p. 161)—it does seem to be a live option that one can form a belief about the look of something, without the look already having to be conceptual before one forms the belief about it. Dretske’s teleosemantic theory portrays the transition between a simple and an epistemic mental state somewhat differently, as being effected through a subject’s selection, abstraction, and exploitation of a portion of the total content of the simple state. Pacherie (2000) describes this nicely.

In later writings (1981, 1988), Dretske offers a treatment of his distinction between simple perception and cognitive perception in information-theoretical terms. The key notions introduces by Dretske in Knowledge and the Flow of Information (1981) are the analog versus digital modes of coding information. According to Dretske’s use of the terms, a signal (structure, event, or state) will be said to carry the information that s is F in digital form if and only if the signal carries no additional information about S, no information that is not already nested in s’s being F. Otherwise, the signal will be said to carry information in analog form…Simple perception is the process by means of which information is delivered within a richer matrix of information (hence in analog form) to the cognitive centers for their selective use. Cognitive activity on the other hand, is the conceptual mobilization of incoming information, and this conceptual treatment is fundamentally a matter of ignoring differences, of abstracting, classifying, generalizing, hence a matter of analog to digital conversion,” (Pacherie, p. 239)

Dretske gives us a picture of how the more determinate content of conscious simple perception of an object or event can be abstracted into belief which ignores some of this fine-grained detail, and proposes to treat an object as the same as others which possess the conceptualized property in question. Simple awareness of an object is meant to be, of course, a consciousness of the object, and as such, I see no need for simple awareness to be already belief-like in content to play such epistemic roles as we have just discussed.

A further point might be noted which Dretske argues at length (1969, pp. 62–77)—the account is not a version of the sense-datum theory. On that latter theory, the objects of perception, what seeing refers to immediately or directly, are some sort of mental phenomena, often taken to be phenomena one cannot be mistaken about. Seeing then refers only mediately or indirectly to the external world objects, if any, which the mental phenomena, sense data, sensa have been caused by. By contrast, on the current account, one simply sees objects and events, not mental phenomena. If we think of seeing like receiving a message, who or what we are seeing is who or what composed the message, not the media of transmission (Ibid., pp. 73–74).

A full theory of the nature of simple remembering’s content, and of how this content may feed into downstream processes of belief and inference, is not our goal here. To pursue the analogy with simple seeing, however, we may safely suppose that if one simply remembers an object or event, then this object or event will be differentiated within one’s conscious memory experience. Most of the time, the objects and events one simply remembers may even be described with some justice as “visually” differentiated within one’s experience, so long as the visual need not be confined to whatever the work of one’s eyes themselves may be. At least, the sort of look which differentiates objects within one’s visual awareness seems to have some close relationship to the sensory profile by which such objects would be differentiated in conscious memory experience. This commonality, match with prior experience, what is sometimes called “authenticity” in the philosophy of memory, seems to comprise one component of memorial accuracy (see, e.g., Bernecker, 2010).

It is time to see what support can be given for simple forms of awareness, and in turn, for simple remembering. We will begin by introducing Dretske’s consistency argument, and then the continuity argument. One nice interpretation of these two sorts of arguments is that they aim, in the first instance, to establish that forms of awareness have non-conceptual content. We then turn to arguments based on seeing without noticing, and seeing under processing limits. I prefer to interpret these arguments as aiming primarily to establish the non-propositionality of simple forms of awareness. It seems that if arguments for these two conclusions can be made to work, we would be most of the way toward establishing that simple remembering indeed has all the criterial features of the more familiar simple forms of awareness, seeing, hearing, and so on.

3 The argument from consistency

It will be generally agreed that it is possible to see something and make one of an indefinitely large range of mistakes as to what it is one is seeing. A person can see his aunt, for example, but not believe that she is his aunt, or a woman, or even a person, mistaking her for mannequin, a hologram, a phantasm produced by an evil demon, etc., etc., etc. (Ibid., p. 7). One need not even believe that one sees anything in order for one to see something. For example, a person in a Perky-effect experiment (Perky, 1958) may perfectly well see colored shapes projected on a wall but believe, because of how the experiment has been set-up, that she sees nothing, believing that the shapes she is in fact seeing, are imaginary. No particular belief about an object or event appears to be commanded of one simply because one has seen it. As Dretske puts it,

This way of seeing a teapot or a tiger is consistent with one’s believing that it is a visual hallucination, a mirage, a reflection, a part of one’s own brain, a phenomenal gloss over an underlying reality, a mental image, or congeries of such images. It is consistent, in other words, with any false belief one may care to mention about the generic character of what one is seeing. And since it seems obvious that it is also consistent with any true belief, it is logically independent of such beliefs. (Dretske, 1969, pp. 8–9, my emphasis)

The idea, I take it, is this. If we suppose that simply seeing an object logically implies that one believes or comes to be believe a proposition about the object, then it certainly seems to follow that there would be some other belief one could have which would be inconsistent with one’s simple seeing (and its doxastic retinue). Naturally, any belief is inconsistent with a belief of its contradictory—the belief that P and the belief that not-P being inconsistent if anything is. There seems, however, to be no further belief one could form about the object or event one sees which would be inconsistent with one’s simple seeing of it, the way it appears to one visually. If one simply sees a teapot, its appearing some way to one does not open the possibility that, if one only formed a particular belief about it, one would simultaneously hold two contradictory beliefs. One can form a belief about the object or event one sees simply, but this belief, no matter what belief it is, does not seem liable to contradict our visual experience itself. Like Dretske says, this way of seeing an object or event seems consistent with any false belief and with any true belief about it. Events of simple sensory awareness do not play with belief, logically and conceptually, as other beliefs would. What, if we like Dretske’s line of argument, should we take it to show?

Dretske’s line of argument seems to show that simply seeing an object or event does not logically compel the formation of any belief about it. But belief is usually taken to be a composite of an attitude, an attitude of acceptance, affirmation, or endorsement, taken toward a propositional content. So, for Dretske to make any claim about a difference between the types of contents involved in simple seeing and belief, namely that content of the latter type is both propositional and conceptual while content of the former type is neither, he would need to motivate the idea that the difference between simple seeing and belief is not merely a difference in the types of attitudes they involve.¹⁶ In other places, Dretske makes the relevant sort of assumption explicitly.

I will follow the practice of supposing that our awareness of facts takes the form of a belief. Thus, to smell that the toast is burning is to be aware that the toast is burning is to believe that the toast is burning. It is conventional in epistemology to assume that when perceptual verbs take factive nominals as complements, what is being described is not just belief but knowledge. Seeing or smelling that the toast is burning is a way of coming to know (or, at least, verifying the knowledge) that the toast is burning. It will be enough for present purposes if we operate with a weaker claim: that perceptual awareness of facts is a mental state or attitude that involves the possession and use of concepts, the sort of cognitive or intellectual capacity involved in thought and belief. I will, for convenience, take belief (that P) as the normal realization of an awareness that P. (1993, p. 266)

If the assumption is reasonable that awareness of a fact, awareness with propositional conceptual content, takes the form of belief of that fact, then it would seem to follow that if what I have been calling the argument from consistency works, then the content of simple seeing would not be propositional or conceptual. Dretske’s assumption seems pretty reasonable. Normally, when I am aware of a fact—when I see that the tree is a magnolia, for example, or I hear that someone is playing a guitar, or I taste that the chardonnay has been aged in oak—to be aware of the fact is to believe it. Of course, not all situations are normal in this sense. I might have been warned not to trust my eyes, ears, nose, or whatever. But perhaps such abnormal cases do not matter very much. For if it is at least sometimes the case that awareness of a fact takes the form of a belief, then, if there is a manner of seeing an object or event which does not open the possibility of holding contradictory beliefs, then there would seem to be a difference in the types of contents involved in seeing objects and events in this manner and in forming beliefs about them. We must leave the matter here. Our central interest is to merely to establish the conditional claim that if Dretske’s arguments work for simple seeing, they should work for simple remembering.

4 An argument from consistency for simple remembering

A very famous example in the philosophy of memory concerns a painter, who we infer, must be remembering the scene which he paints, judging from how accurately he has painted it, but who does not believe that he remembers the scene, or that he ever saw the scene, or even that the scene is, or ever was, real, taking what he paints to be purely imaginary. Martin and Deutscher (1966) devised the example to undermine the assumption they ascribe to Locke, Hume, Russell, and Malcolm, that remembering an event requires that the subject believes (of what is being remembered) that ‘this occurred.’ The example may support an argument from consistency in the style of Dretske.

Suppose that someone asks a painter to paint an imaginary scene [… The painter’s] parents then recognize the picture as a very accurate representation of a scene which the painter saw just once in his childhood. […] Although the painter sincerely believes that his work is purely imaginary, and represents no real scene, the amazed observers have all the evidence needed to establish that in fact he is remembering a scene from childhood. (Martin and Deutscher, p. 168)

The case does seem effective in undermining the claim that, “it is not possible for someone to remember something unless he believes that it happened,” (Ibid., p. 168). The painter does not believe that it happened, that he ever saw the scene. But the painter does very much seem to remember the scene. How else would he contrive to paint it with what is, (we are free to suppose,) such exquisite accuracy? That the painter remembers the scene seems to be by far the best explanation.¹⁷ What it less clear, at least at this point of the discussion, is whether the argument shows that remembering an object or event does not require a belief about what it is that is being remembered.

As it is described, this case is not free of belief. Martin and Deutscher tell us that, “the painter sincerely believes that his work is purely imaginary, and represents no real scene,” (Ibid.) This should not be allowed to distract us. There is little plausibility in the assumption that the way the painter has the scene before his “mind’s eye” requires that he form this belief. It is quite possible for the scene to be remembered thus, to appear before his mind in the same way, and for the painter not to form this mistaken belief about the scene. Simple mental events, seeing, hearing, and so on, are always compatible with beliefs. The point is that there is no belief one just has to form about what one is simply aware of purely in virtue of being in a state of simple awareness of it. To continue the inference, let us note that it seems that there is no conscious belief the painter could form which would be inconsistent with his remembering of it—with the way the scene appears before his mind—in just the fashion that contradictory beliefs can conflict. Rather, the painter’s remembering of the scene seems consistent with any false belief he might form about the scene, and with any true belief he could form about the scene. But this seems not to be a matter of remembering and believing being the sorts of attitudes that are constitutionally incapable of conflicting, clashing, or mutually-contradicting. Rather, if one remembers a fact, for example, that a person first traveled to space in 1961, but one believes that it is not the case that a person first traveled to space in 1961, then there is a clear sense in which one’s mental states conflict. So, if in remembering the scene, the painter is free to form any beliefs whatever about it, and without the possibility of this sort of strict clash of mental states, then the painter would seem to remember the scene in the simple way at issue.

5 The continuity argument

It is a likely thought that the perceptual world we humans enjoy coincides with that of some non-human animals, cats, cows, crows, etc.—that we have a way of seeing (or hearing, etc.) in common with some of these sorts of animals. Dretske introduces the notion of simple seeing by making that sort of point. Simple seeing, he describes as “a primitive visual ability which is common to a great variety of sentient beings, an ability which we, as human beings, share with our cocker spaniel and pet cat,” (Dretske, 1969, p. 4). Although we may share a way of seeing, or hearing, etc. with non-human animals, it would be unlikely in the extreme that such animals share our human concepts. So if objects or events can look, or sound, or taste, etc. the same ways to some of them as they do to us, the experiences we share cannot require the concepts we do not share. But beliefs, judgments, knowledge, (propositional knowledge anyway,) does require such concepts. So the ways of experiencing the world we share do not require beliefs. In Naturalizing the Mind (1995), Dretske sketches the point like this:

I can see Paul playing the piano and believe he is playing the piano, but the visual experience represents the piano playing in much different ways than does the belief. These are different kinds of mental representation … Experiences of piano playing do not require the concept of a piano (at least not in the same way as a belief or judgment requires it). They require no understanding of what a piano is or what it sounds like. Even mice can see and hear pianos being played. Believing is something else. It requires the concept of a piano, some understanding of what a piano is. Mice who hear pianos being playing do not believe pianos are being played. Their understanding is, I assume, too feeble to believe this even though their hearing is good enough to hear it. (Dretske, 1995, p. 9)

Let us delve deeper into this line of thought. We are, very plausibly, endowed with a way of seeing the world in common with a variety of animals. Lassy can see the yellow frisbee much as you or I can, and seeing it she can catch it in mid-flight. Bruce the cat can hear Beethoven’s 9^th blaring on the radio. Quite probably, objects can look to animals as they look to us, events can sound to animals as they sound to us, and experiences can be common. But in what manner should we understand the proposal that such non-human animals lack our concepts?

There are different ways to go. Christopher Peacocke has been a major proponent of the continuity argument. In his estimation, it provides “the most fundamental reason for recognizing nonconceptual representational content in perception,” (2001, p. 609). In Peacocke’s telling, non-human animals lack our concepts because they lack concepts more generally.¹⁸ Dretske has been much less reluctant to attribute concepts to non-human animals. In his teleosemantical writings, (e.g. 1981, 1988, 1995,) Dretske identifies experiences with states whose representational properties derive from their systemic functions, and conceptual states with states whose representational properties derive from their acquired functions (1995, p. 15). This is relevant because Dretske gives examples of non-human animals tokening states whose representational properties derive from acquired functions—monkeys that learn to respond to the larger of two rectangles (1981, p. 152), dogs that learn to salivate when middle-C is played on clarinet (1995, p. 14), and others. Presumably then, Dretske intends the continuity argument to trade on the more liberal idea that non-human animals simply lack our concepts, having perhaps acquired different concepts suited to their own needs. Let us return to the example of Paul’s piano-playing.

Suppose one lives with Paul, and late at night one is awakened by his piano-playing. ‘That’s piano-playing!’ one judges after a moment’s confusion. Suppose further that a mouse, we may refer to as Mouse, has been located under one’s bed, and Mouse hears the piano-playing too. Maybe Mouse believes of the piano-playing that it is ‘#&@!,’ falling under, we may suppose, this mousey concept. Let us call the piano-playing, how it is auditorily presented to both Mouse and oneself, ℵ.

If it is true that the same auditory experience of the piano-playing is presented to both Mouse and oneself, it would seem to follow that this auditory experience is itself non-conceptual, non-conceptual in a manner that beliefs cannot be, and moreover, that this experience does not strictly require belief, (although beliefs, of course, may happen to arise, as they have in this case).

In the interest of plain dealing, one might note that similarity between human and non-human experiences may be a matter of extents, respects, and not a binary affair. Moreover, it seems clear that how similar human and non-human animal sensory experiences can be will make a big difference for what Dretske’s argument can show. I’ll sketch just two possibilities here. If we take it that a human and a non-human animal can have sensory experiences of a particular object which coincide rather fully (at least relative to a given sense modality), then it would follow that any concepts which the human and non-human animals come to bring the object under must, (because they are different concepts,) be applied in the context of subsequent beliefs, and cannot play a part in the initial experience of the object which is shared. If, on the other hand, we take it that the sensory experiences of an object or event enjoyed by human and non-human animals can only coincide to some minimal extent, this would presumably open up the possibility that the experiences in question are both somewhat conceptual and somewhat non-conceptual. The respects of shared experience would come out as non-conceptual, while the respects of the experiences that are not shared might be conceptual or might not. Dretske presumably intends the first possibility, for it is this which would secure the features of simple forms of awareness we have discussed. This possibility is relatively intuitive, but in any event we shall not debate it here.

6 A continuity argument for simple remembering

There may be a unique challenge in providing a continuity argument about memory. Of course, it is always difficult to ascribe mental contents to non-human animals, (as any continuity argument will require,) because such creatures cannot express themselves linguistically, (or at least not in a language we comprehend). The best we can do is to attribute mental states based on the evidence we do have, often behavioral in nature; namely, to attribute whatever patterns of mental states best explain or predict such behavior. The unique challenge in attributing states of remembering is that the causal chains between objects and memories of them is usually much longer than the causal chains between objects and perceptions of them. This can make it tricky to say what it is a creature might be remembering, if remembering at all. But the difficulties in attributing states of remembering to non-human animals seem unique more in degree than in kind. For even in the case of memory, there may be good empirical evidence pertaining to how non-human animals represent the world. We will do well to rely on such evidence.

Continuing the piano example, suppose that one has grown tired of the mouse under one’s bed, and one contrives to catch it in a cage. Suppose that at fair distance from home, one sets it free—three blocks up, and two blocks over perhaps. Remembering the location of the apartment, one returns home, but to one’s dismay, so does the mouse.

It is, by now, widely accepted that a great variety of sentient beings remember and navigate familiar spaces by means of cognitive maps. Edward Tolman first proposed the hypothesis in the 1940’s to account for the short-cutting behavior of rats which had been trained on a maze. This behavior could hardly be explained on the prevailing behaviorist principles, in terms of conditioned stimulus and response associations, for it involved taking a completely novel route to a reward. For example, in the study “Orientation and the shortcut,” Tolman et al. (1946) trained rats to take a particular path to a reward. After training the rats to a high degree of proficiency, the experimenters altered the maze, blocking the original path, adding 18 new paths, but not altering the reward’s location. When the trained rats were tested within the new maze, the greatest proportion took the shortest and most direct path to the reward, a path which could not have been familiar to them. The best explanation of these and related results is provided by the hypothesis that the rats which took the shortcut remembered the location of the reward, and more generally the model that rats store cognitive maps. Later researchers have made significant progress in discovering how such maps are represented in the brain, and how they are anchored to the perceived environment. Two Nobel prizes have been awarded for such work. While much of the early work on cognitive maps had focused on rodent models, more recent developments support the existence of very similar cognitive maps in humans (for a comparative review, see Epstein et al., 2017). Research on cognitive maps gives significant plausibility to the idea that in the piano example, home’s location would be remembered by both the mouse and oneself in virtue of the same sort of cognitive maps. This, in turn, licenses the presumption of overlap in content between the two memories of home’s location. But if we may rely, as before, on the premise that non-human animals lack human concepts, it would follow that a form of memory has content that does not require concepts in the way belief does, and hence does not require belief.

A number of reasons stack up against the supposition that cognitive maps are somehow disconnected from experience. Unless conscious visual experience of a landmark allowed the positioning of oneself “within” a cognitive map, the purpose of having a cognitive map would be entirely obscure. Likewise, unless the content of a cognitive map were accessible, were capable of informing one of the lay of the land and of possible routes to take, the purpose of having a cognitive map would be obscure. Moreover, one sometimes has the experience of winding up in a location does not expect, of getting lost. Perhaps such an experience makes rational an update of one’s cognitive map (Cf. Heck, 2007), such rational relations usually indicating person-level dynamics. A last reason, more speculative than those preceding, centers on the role of cognitive maps in the so-called method of loci. The method of loci is a mnemonic technique of ancient origin, noted as far back as Cicero, although still in use by professional mnemonists today (Sutton, 2010). The technique consists in the assignment of a piece of information to a particular location within a known environment, a “memory palace.” The subject then visualizes taking a particular course through the environment, retrieving each piece of information upon visiting its assigned location. According to preeminent authorities, the technique works by storing content within a cognitive map. “On the cellular level, different rooms in the memory palace are represented by unique ensembles of place cells. Thus, the place cells ‘carry’ the memories and keep them separated,” (Bjerknes & Moser, 2019, p. 12). If May-Britt Moser is correct in arguing this, then cognitive map content is accessible and manipulable. Suffice it that this is a very respected empirical hypothesis.

So far, the argument of this section has kept close to Dretske’s continuity argument. In doing so, it rests on the assumption that non-human animals lack the concepts we possess. The continuity argument for simple remembering seems no worse off than Dretske’s in this respect. But I would briefly like to discuss an alternative route to the conclusion of simple remembering from the starting point of cognitive maps. This is the argument Richard Kimberly Heck (2007) has given that “[c]ognitive maps do not have conceptual content: their content is not structured in the way the contents of belief are,” (125) Of course, if Heck is right, and cognitive maps lack propositional structure, then we need not appeal to any continuity between the cognitive maps of mice and men.

Heck’s reasons for claiming that cognitive maps lack propositional structure are essentially two-fold. First, there seems to be no unique or privileged proposition that corresponds to the content of a cognitive map. Cognitive maps involve indeterminacy along dimensions structured propositions must specify. The only plausible proposition to correspond to a cognitive map would be an extremely long conjunction of all the relationships the map indicates. But what are the functions and what are the arguments of the conjuncts? What is the order of the conjuncts? What would specifying all the relationships a map indicates amount to? “There is no unique structured proposition that gives the content of a map because there is no such structure in the map; a map lacks the syntactic structure present in a verbal description of what it represents,” (Ibid., p. 126).

Heck’s second sort of reason derives from the fact that map contents do not behave and mutually interact like structured propositions. The structure of propositional contents guarantees they can be subject of logical operations, and related by logical operations to other propositions, while map contents apparently lack these features.

One cannot, for example, form arbitrary Boolean combinations of maps: there is no map that is the negation of my cognitive map of Boston; there is no map that is the disjunction of my map with my wife’s; and so forth. If the content of a cognitive map is a structured proposition, why shouldn’t there be maps with such contents? (Ibid., p. 126).¹⁹

None of this means, of course, that one cannot form beliefs on the basis of a map, be it cognitive or otherwise. But if Heck’s arguments stand, then we have an alternative basis for the conclusion that cognitive maps are not propositional, an alternative basis for simple remembering.

7 Seeing without noticing

A further motivation for simple forms of awareness exploits the connection linking conscious belief about something with the noticing of the object or event concerned. There is an impression that we see too many small events, too many countless objects, to form a belief about each and every one. Turn your head and there’s a new, vast panoply in sight. Most of the things one sees seem to pop in and out of visual experience without leaving a trace in one’s beliefs. They simply escape notice.

Dretske’s driving thought behind cases of seeing without noticing is that forming a belief about something, a conscious belief, requires singling it out in attention, and setting it up in the logical role of either a subject, or a property of something else. Whenever one sets something up as a subject or a property in a belief, one notices it. Therefore, if one is capable of seeing something without noticing it, attending to it, singling it out for special processing, the possibility of simple seeing would seem to follow. Suppose one visits the library of a friend, quite a large library, and one scans all the books on the shelf. It seems quite possible to see all the books, without forming a belief about each book. More than one book presumably escaped notice, judgment, belief, although fully differentiated from its fellows in one’s visual experience of the library. Sometimes, Dretske attempts a sort of informal demonstration of this phenomenon (Cf. Dretske, 1993). Try inspecting the following two figures.

There is little doubt that you saw each of the circles comprising Figs. 1 and 2, and so, on one way of speaking, you saw the difference between the figures. Namely, you saw the circle in the bottom left of Fig. 1 with no counterpart in Fig. 2. But while you saw the thing that is the difference between the figures, you may not have formed any belief about that circle, for example, the belief that it is the difference between the figures, since you may not have noticed that circle or identified it. “To occupy a belief state a system must somehow discriminate among the various pieces of information embodied in a physical structure and select one of these pieces for special treatment—as the content of that higher-order intentional state that is to be identified as the belief,” (Dretske, 1981, p. 174, original emphasis). On the other hand, simply seeing something, seems to presuppose nothing about whether the subject notices it, whether she takes it to be something in its own right, whether she exploits her visual experience of it in any manner (1969, p. 21).

Fig. 1

…

Fig. 2

…

A misunderstanding of the point should be warded off. According to Michael Tye (2009, 2010a, 2010b), Dretske is inferring from the fact that a subject sees all the members of a group to the conclusion that she sees each member. But this sort of inference may not always be good. One sees an object, Tye supposes, only if one is visually conscious of it, and one is visually conscious of an object just in case one can “mentally point to it on the basis of experience … [and ask] ‘What’s that?’,” (2009, p. 259). However, it is possible to see a whole group of things, without being able to single out each group member and ask, regarding each, ‘What’s that?’. In Tye’s example (Fig. 3), it should not be possible, while fixating on the plus sign, to single-out the fifth bar to the right.

Fig. 3

Adapted from Tye (2009)

Tye’s point, to reverse the adage, is that sometimes one may see the forest, but not the trees, or at least not each of them.

This objection misses the point. Presumably, Tye is right that sometimes an object is presented in visual experience in an indistinct fashion that does not allow one to attend to it in its own right, but only insofar as it is a member of a group. But the case from seeing without noticing in no way depends on the supposition that objects simply seen are only indistinctly presented in visual experience. One might be sure to cast one’s eyes over each element in a panoply, have each in visual experience as distinctly as anything ever is, and still fail to subject each to a belief. Visual experience can put one in a position to single something out without one actually singling it out, noticing it, forming a belief about it.

If Dretske’s premise holds good that forming a belief about something involves selecting, or taking, or noticing it, then this idea can service an equally good argument for simple remembering. Suppose, for example, that one sets about recalling the route from the department to a restaurant in order to give directions to a colleague. In the process, suppose one remembers shops, intersections, a tree-lined park, a café, and much else besides. It is plausible to suppose that in recalling the route, one does not select, notice, or take in any particular way a particular one of the several trees that appear in one’s memory near the corner of the park, setting it up within a belief. If so, one’s memory of the tree is simple.

8 Simple awareness under processing limits

Another powerful consideration in favor of simple seeing is that subjects can see various objects and events under such processing limitations that they are unable to recognize, identify, judge of what objects they see. Exhibit A in this sort of argument is George Sperling’s classic experiments on what later became known as iconic memory.²⁰ Dretske appeals to this work in an underdiscussed paper of 1978 (see also, 1981, pp. 149–150). In the Sperling paradigm, an array of letters is flashed on the screen for a fraction of a section (Fig. 4). After a short delay, a tone cues subjects to recall a particular row. A high-pitched tone cues the top row, a middle-pitched tone cues the middle row, and a low-pitched tone cues the bottom row. It has been found very consistently that subjects in such experiments can accurately recall any one cued row (three or four letters), but not much more than that. They report a quickly fading image of the array, one which disappears too quickly for them to be able to read off letters in un-cued rows. Dretske’s main claim is that the truth of the counterfactual, that subjects could have recalled any letter had its row been cued, strongly suggests that subjects saw each letter. How else could subjects recall each if prompted? On the other hand, subjects do not, and in a sense, cannot form beliefs about each letter, at least not purely in virtue of seeing each letter.²¹

Fig. 4

Example array from Sperling (1960)

Dretske’s argument seems to reach the conclusion that subjects remember all of the items in the array simply, as an interim point along the way to simple seeing. Dretske argues that since test subjects could have reported any of the letters (given cuing), they presumably must have seen each of the letters. But seeing each of the letters would presumably be no help unless subjects also briefly recall each of the letters. For the array has already disappeared by the time a tone cues any items to be recalled. Therefore, that subjects can report any letter (given cuing) suggests they must remember each letter, however briefly. Moreover, that subjects cannot read-off all of the letters suggests that they do not remember all the letters for long enough to bring each under conceptualization in explicit belief. Therefore, subjects remember some of the letters, (those in un-cued rows,) simply.

9 Conclusion

For Dretske, the structural difference between seeing a fact about an object and seeing the object simply lies in that the visual content one exploits in judging the fact is already embedded, contained, or nested, as an element, in a richer, more unified and specific complex of visual content about the object which one entertains in simply seeing it. When one believes, for example, that ‘the book is on the table,’ the visual information one uses in forming this belief was already contained in the more specific and determinate content one entertains in seeing the book, the size of its print, the color of its binding, the texture of the paper, the orientation, color, shape, size, of the book, its distance from one, and so on (1981, p. 137).²² There is no seeing the book on the table without entertaining much other content than that it is on the table. But one can surely believe that the book is on the table without entertaining any content about, for example, its color. How the transition is effected from simple to epistemic seeing, from seeing to believing to knowing, is a complex story of selection and abstraction, and sometimes inference, a story which will have to wait for another day.

The interesting point I have endeavored to show is that Dretske’s distinction between simple and epistemic mental events may extend more widely than has previously been acknowledged, and specifically to the case of memory. Considerations parallel to those Dretske used to support simple seeing seem to do just as well in supporting simple remembering. This makes a degree of sense when we consider the commonalities of seeing and remembering more generally. Both seem to present imagistic panoplies we can allocate our attention around, onto objects and events we have not yet formed explicit beliefs about prior to our noticing them. Both seeing and remembering anchor our abilities to form and understand singular thoughts about objects and events in our experience. Moreover, the fact that seeing and remembering may bring objects and events sensorily to mind we have not yet gained a conceptual grip upon may be important to how we, as children and as novices, acquire new concepts for them (Dretske, 1981; Martin, 1992; Roskies, 2008).

Of necessity, this has been a preliminary foray into the topic of simple remembering. Many more questions have been raised than answered. For example, if simple remembering exists, how might it fit within or alongside the scientific taxonomy of memory? If simple remembering exists, how can we specify its content? If simple remembering exists, how might it inform downstream processes of categorization, judgment, and inference? These are interesting questions. But the prior question is whether simple remembering exists at all.²³