Vision Science

IN SEEING THINGS AS THEY ARE, John Searle turns his attention to perception — visual perception, to be precise. Perception is both the basic way that minds connect with configurations of objects and attributes in a local environment, and an epicenter for sensory feeling and experience. That is, perception is a site of both representation and phenomenology. And since the capacities for representation and phenomenology have long been taken by philosophers to be characteristic marks of the mental, philosophical questions about perception provide a window into philosophical questions about minds more generally.

When it comes to the long tradition of thinking and writing about perception, Searle takes the situation to be rather bleak. He believes that the entirety of philosophical work on perception since Descartes has been bewitched by what he calls “the Bad Argument” and, as a consequence, is unnecessary and incoherent. Yet Searle wants to not just bury philosophical theories of perception but also praise them. In particular, he believes that once the bad argument is identified and diagnosed, nothing will prevent us from endorsing a form of direct realism about perception, of the sort Searle himself developed in his 1983 classic Intentionality. According to this form of direct realism, we do not perceive external objects by way of first perceiving intermediate ideas, impressions, or sense-data; instead, perception serves to provide us with immediate presentations of external objects and attributes themselves. In short, our perceptual capacities enable us to see things as they are in the local environments in which we find ourselves, and this fact should serve as the backbone, rather than an optional add-on, to philosophical reflection on minds and their epistemological condition.

This may seem fairly straightforward, and, in some ways, it is. There is an external world, and it is full of things: tables, crocodiles, textures, etc. These things and this world exist whether I like it or not: their existence is independent of my beliefs, opinions, or preferences, and hence we say that such an existence — or, to use the technical term, such an ontology — is objective. There is also a subjective world, and it consists of internal states of mind. Such states are not ontologically objective, but subjective: they depend for their existence on the person who has them. Moreover, there is generally something that it feels like to be in or occupy a state of mind: we all know what it is like to be mad or tired, and we similarly know (although this case is more complicated) that believing something feels different from not believing it. The central claim of direct realism is that perception puts the external world into contact with the subjective one. Thanks to the rise of modern vision science, we have a basic sense of how this story might go: an internal causal process is initiated by arrays of light moving from entities in the external work to sensory receptors in our retinas; these arrays of light are then processed by a module in ours head that constructs an output perceptual representation on the basis of proprietary perceptual principles. This perceptual representation has content: it encodes conditions of satisfaction that are either accurate or inaccurate, depending on the extent to which the representation corresponds with the scene that initiated the causal process. While beliefs, desires, and other mental states are also associated with representational contents — beliefs can be evaluated as true or false, desires can be fulfilled or unfilled, and so forth — perceptual representations are special in that they cannot be detached from, or entertained independently of, the scenes that prompt them. In the more contemporary philosophical jargon, perception provides a non-conceptual means by which minded creatures get in touch with an objective world.

There seems to be a lot to recommend this basic picture of perception. How could we plausibly deny that perception plays a central role in connecting us to, and helping us acquire knowledge about, the empirical world? According to Searle, however, the picture has not only been denied; it has been denied by “just about every famous philosopher who writes on this subject.” “Indeed […],” Searle observes, “I do not know of any Great Philosopher who even accepted […] Direct Realism.” If you’re wondering who the “Great Philosophers” might be, he is referring to Bacon, Descartes, Leibniz, Spinoza, Locke, Berkeley, Hume, and Kant, and maybe Mill and Hegel, too.

Each of these great philosophers falls prey to what Searle calls the Bad Argument. The Bad Argument turns on the truism that perception, like all other mental faculties, does not provide us with infallible access to the world. We sometimes get things wrong, from a perceptual point of view. Consider the environment you’re in right now: there is probably a computer display in the foreground, maybe some furniture or a few other people in the background. Now take a snapshot of that scene. This snapshot has representational content; it represents the world as being a certain way. But your snapshot could be inaccurate in matters of detail, as is the case in illusions — for example, the color of the edge of the screen that you perceive as being white could, in fact, be gray. Indeed, your snapshot could be inaccurate in matters of substance, as is the case in hallucinations — for example, your perceptual experience of there being any material bodies at all in front of you now could be the unfortunate result of a badly malfunctioning brain or a Matrix-style computer simulation. In all such cases, we are left with an unhappy result, namely that your perceptual system has produced an inaccurate representation of the external world.

So what? Why should these truisms about the possibility of perceptual error bear on direct realism? The answer, it turns out, is that there are good reasons to think that two perceptual states that differ only in that one of them involves an accurate perception of the world and the other involves an inaccurate perception of the world are psychologically of a kind. In other words, the type of perceptual state you’re in does not vary if you’re hallucinating or compos mentis. After all, the happy and unhappy cases of perception are phenomenologically indistinguishable: they feel exactly the same, even though one is accurate and the other is not. In addition, the point jibes with what vision science seems to suggest about the ways in which our visual system constructs perceptual representations from incoming arrays of light. So long as the proximal stimulation on the retina is the same, even if the distal causes of those proximal effects vary, the perceptual system will generate the same kind of representational content.

The philosophical orthodoxy — as Searle puts it, more or less from Descartes onward — has taken all this to spell bad news for the realist picture of perception. After all, if I’m hallucinating that there is a computer in front of me now, then the object of my perception is not an external object but some sort of internal object — an idea, an impression, or a sense-datum — created by my mind. And if such hallucinatory cases of perception are psychologically of a kind with accurate cases of perception, then it would seem to follow that the objects I see even when I’m not hallucinating are likewise internal, mind-dependent, objects. In particular, it would seem to follow that we never directly perceive external objects. Bad news indeed.

Many philosophers have accepted this argument as stated, and have developed what I’ll call anti-realist theories of perception. According to one version of anti-realism, associated with the 18th-century philosopher Bishop Berkeley (yes, he’s among Searle’s list of Greats), the so-called objective world is in fact constructed out of mental phenomena, perhaps ideas in the mind of God or of some social collective, but ideas nonetheless. A more common version of anti-realism accepts that there is, in fact, such a thing as a world independent of our minds, but that we nonetheless can only perceive that world by first perceiving mind-dependent sense data.

Searle has little patience for anti-realism in any of its forms. For Searle, being an anti-realist about perception is the first step on a slippery slope to the dubious pronouncement that we cannot have any empirical knowledge whatsoever. He also argues that it leads to the position — often called solipsism — that there is no publicly accessible world about which we can share beliefs and languages, and coordinate our actions. There are other problems, too, including that the claims anti-realism provides offer no coherent account of the features that initiate an episode of perception (what’s the light reflecting off of?), but most importantly, the anti-realist account “just seems to be inadequate to our own experiences.” “It is just a fact,” Searle writes, “about our experiences that they reach right out to independently existing objects and states of affairs in the world.” In short, Searle believes that anti-realism about perception is one of the greatest intellectual failures of the Western philosophical tradition.

Perhaps sharing some of Searle’s concerns about anti-realism, a number of philosophers — including, as Searle likes to point out, those in his own department at UC Berkeley — have recently developed a more critical response to the argument. These philosophers reject the central premise of the argument above concerning the relationship between the happy and unhappy cases. This style of response leads to a position called disjunctivism: the thesis that there is no explanatorily important kind of perceptual state that is both specific to, and shared between, an accurate perception and its inaccurate counterpart. At its core, disjunctivism is motivated by a desire to preserve the claim that perception involves direct relations that hold between a subject and a mind-independent world. Indeed, disjunctivists often insist that perceptual states are literally built up out of the external objects and attributes that the world provides. For the disjunctivist, realism requires recognizing that there is a fundamental divide between happy and unhappy cases of perception. When we’re hallucinating, we aren’t really perceiving in the proper sense of the term; rather, we are in a degenerate kind of mental state that has exactly the same phenomenology as real perception, but is nevertheless of an entirely different kind.

Disjunctivism provides a way to block the argument from a perceptual error, and thereby, a way to save direct realism. And yet Searle has no more time for disjunctivism than he has for anti-realism. The disjunctivist hypothesis, he argues, is simply at odds with the science of perception as it has been revealed by both cognitive psychology and neurobiology. More to the point, disjunctivism fails to offer a coherent account of the causal processes that trigger the degenerate kind of perception, or of how the degenerate kind of perception relates to perception in the proper sense of the term. But even when it comes to happy cases of accurate perception, Searle maintains that disjunctivism is flawed: for disjunctivism tries to build external objects into what it is to have accurate episodes of perception, but (he argues) it just makes no sense to try to locate an external object in an internal state of the brain.

Searle’s objections to both anti-realism and disjunctivism are not without force, though it is unclear to me how persuasive they will be to those who don’t already share his philosophical commitments. After all, many anti-realists are quite willing to endorse the skeptical and solipsistic implications of their views, and versions of the so-called “extended-mind” thesis, according to which aspects of the environment can be proper parts of a subject’s psychological states, are cheerfully acknowledged by disjunctivists. More useful, to my mind, is Searle’s claim that all these arguments about illusion, hallucination, and perceptual error — in sum, the Bad Argument — make the same elementary error. The error is one of equivocation, for the Bad Argument turns on an equivocation between claims about the contents of perceptual states and claims about the objects of perceptual states. The contents of perception pertain to how a subject represents the world as being, whereas the objects of perception pertain to what in the world a subject’s perceptual states happen to be about. All perceptual states do indeed have contents, but, Searle suggests, not all perceptual states have objects. Hallucinations provide an excellent case in point: they are perceptions that represent the world as being a certain way (i.e., they have representational contents), but for such perceptions there is no corresponding state of the world that actually is that way. In other words, such perceptions have no real associated objects associated with them. It is thus a mistake to assume that because happy cases and unhappy cases are psychologically of a kind at the level of perceptual content that they must thereby be alike at the level of perceptual object; no such inference follows.

This distinction between content and object is not novel to Searle. It was made explicit in the late 19th and early 20th century by the students of Franz Brentano, among them Kazimierz Twardowski and Alexius Meinong, and it is now standard issue in the contemporary philosopher’s tool kit. Likewise, the idea of using the distinction between content and object to defuse the argument from illusion is also not new. Although this fact goes unnoted by Searle, it features prominently in a well-known paper by G. E. M. Anscombe from 1965. But what is apparently new in Searle is an appreciation for the philosophical insights that the distinction between content and object yields. I will highlight two such insights.

First, once the distinction between the content and the object of perception is appreciated, there will no longer be any barrier to our accepting that the objects of perception are, without exception, external, mind-independent entities, and that representational contents or other mind-dependent entities such as impressions or ideas are not things that we perceive; perceptual contents pertain to how the world is represented, not to what in the world is represented. In this sense, representations are not veils through which we perceive external objects; we perceive external objects directly, when we perceive anything at all. Second, the fact that our perceptual states sometimes have contents with no corresponding objects in the world does not imply that we always, or even often, find ourselves in such an unhappy predicament. More often than not, we find ourselves in happy cases. We are not brains in vats, and we are not in the Matrix. We are rarely on LSD, and many of us do not suffer from schizophrenia. A proper theory of the relationship between mind and world can allow us to recognize the possibility of an occasional hallucination, or perceptual error, without thereby giving up on the claim that perception enables us to see things as they are in the environments in which we find ourselves.

But now comes the hard part: the development of a theory of perceptual content, and a realist account of our knowledge of the external world purged of any contaminating effects of sense-data and their ilk. To his credit, Searle does attempt to undertake the first of these tasks. Indeed, the full title of the book is Seeing Things as They Are: A Theory of Perception. And a good portion of the book is devoted to articulating a theory of the nature of perceptual contents and how those contents come to be associated with particular states of mind.

We are told, correctly in my view, that perception is hierarchically structured around a set of basic features or attributes such as size, shape, color, and texture. These basic features or attributes are then enriched by higher-level features in the course of perceptual processing to provide a perceptual content that represents a full three-dimensional scene. According to Searle, both the basic and non-basic elements of perceptual contents come to be associated with the particular states that they are in virtue of background episodes of causal interactions — either in the individual or in the species as a whole — of causal interactions. Our perceptual states represent the objects and attributes that they do because, at least at one time, those objects and attributes have caused us to be in those types of states.

I am about as sympathetic a reader as Searle is likely to find. I am a professional philosopher by trade, working on issues in the philosophy of mind and language that intersect with contemporary cognitive science. As it so happens, I agree with Searle that the Bad Argument is properly so called. I also agree with Searle that anti-realist and disjunctivist approaches to perception should be rejected, and that something close to direct realism, as he understands it, should be accepted. But Searle’s thin development of his own positive theory — which is to say, the theory that is left when the critique of other theories has been made — left me cold. It is not just that I disagree with Searle about some of the details of his account; for example, his claim that spatial relations, such as depth, are not among the basic perceptual attributes, or that all well-functioning psychological states are consciously accessible by the person experiencing them, or that principles related to so-called perceptual constancies are dispensable. Nor is my dissatisfaction simply due to the fact that Searle fails to engage with data suggesting that perception itself, as opposed to beliefs formed on the basis of perception, is only indirectly affected by individuals’ prior expectations and preferences (perception is, to use the more technical term, largely encapsulated from other kinds of cognitive processes). It is rather that I firmly reject a central aspect of the methodology that Searle utilizes to guide the construction of his theory. For Searle develops his theory of perception from what we call the armchair, by which I mean that Searle develops the details of his positive account in almost complete isolation from the wealth of recent work in perceptual psychology. If perception is, as Searle insists, a natural kind like digestion or photosynthesis, then one cannot provide a theory of its operations or attempt to answer philosophical questions about its nature independently from empirical investigation. In short, Searle cannot have it both ways: he must either give up his naturalism or radically revise what he takes to constitute a theory of the relevant domain.

More profoundly, and more pressingly, I find myself out of alignment with Searle’s vision of the philosophical landscape. From Searle’s point of view, there are those philosophers of past and present generations who have given false or otherwise misleading accounts of perception, and then there is Searle and Searle alone. I have no principled objection to this degree of self-confidence — it may well have its place as a rudder to guide theory development or as a rhetorical device in philosophical argumentation. But one cannot help but be struck by the fact that the philosophers that Searle does cite and engage with are exclusively male. This is particularly worth noting given the ways in which the works of a number of female philosophers — again, G. E. M. Anscombe, but also Kathleen Akins, Heather Logue (mentioned briefly in her capacity as an editor, but not as an author), Susanna Schellenberg, Susanna Siegel, and many others — bear directly on the themes discussed in Seeing Things as They Are. It is well known that academic philosophy is now struggling, as it has for some time, with issues of diversity.[1] It is also well known that academic philosophers are not particularly good about citing the works of members of groups already underrepresented in their discipline.[2] So, in fairness to Searle, these deficiencies with the book are not exceptional or particularly notable — to the contrary, they are the norm in philosophy. Still, many of us hold out hope for a future vision of the landscape of philosophy in which we look out and do not simply see ourselves.





Josh Armstrong is Assistant Professor of Philosophy at the University of California, Los Angeles. His research concerns language variation and change, social cognition, and animal communication.