Access Internalism, Attention, and Perceptual Expertise

Winner of the Outstanding Graduate Paper at the 62nd Annual Meeting of the Florida Philosophical Association

Adam Hauptfeld, University of Miami

The Expert and the Novice

Consider two people looking at an oak tree. One of them is an expert arborist who can tell by looking at the tree that it is an oak; the other is a novice who has not yet learned how to identify oak trees by their looks. There is an epistemological difference between them—the expert is justified in believing that the tree in front of her is an oak tree while the novice is not. But this epistemological difference is not basic; it must be grounded in some further difference between the expert and the novice. What other differences are there between them?

There is a historical difference: the past experiences of the expert are different from the novice’s and include tree identification training the novice hasn’t had.

There is a dispositional difference: the expert and the novice are disposed to believe differently in response to the oak tree stimuli.

There is a difference in stored representations: the expert’s mind has a stored representation of what oak trees look like while the novice’s does not.

There is a phenomenological difference: the character of their conscious experiences of the oak tree is different.

Of these, only the last option is available for providing a fundamental explanation of the epistemological difference for an access internalist like me. Since only a difference in consciously accessible evidence can ground a difference in justification, there must be a phenomenological difference between the expert and the novice if we assume that differences in consciously accessed evidence corresponds with differences in the phenomenology of consciousness. ¹ I think many philosophers will initially find this unappealing. Indeed, the expert and the novice case can appear to be a counterexample to access internalism precisely because it seems to be a case where two people differ in the justification available to them without differing in what they have conscious access to. Part of what pursuing this strategy must consist in is therefore motivating that there is this difference.

The Seemings View

One strategy for defending an access internalist solution to this problem is to posit the existence of a sui generis conscious mental state called a “seeming” that mediates between sensory experiences and perceptual judgements. ² Seemings share with judgements the qualities of being propositionally structured and conceptual, and of having a cognitive phenomenology. They differ from judgments by being reportive and a-rational, and in these respects are more akin to perceptual states. Their purported existence has been used to explain the epistemic difference between experts and novices, the idea being that seemings are phenomenally conscious and evidentially relevant mental states that differ in content between experts and novices. We can call this the “seemings view.”

According to the seemings view perceptual experiences have two parts: a sensory experience and a seeming. The expert and the novice have identical sensory experiences, but the expert also a seeming that the tree is an oak while the novice has no such seeming (although he has a seeming that the tree is a tree). This seeming is propositionally structured since it is predicating a property—viz. being an oak—of an object—viz. the tree. The seeming is conceptual since it involves the employment of the expert’s concepts, and it has a cognitive, as opposed to sensory, phenomenology, since what it is like to experience a seeming is more akin to having a thought than seeing an object.

The expert’s seeming is reportive in that it purports to convey to the expert something about the world, namely that the tree is an oak. Indrek Reiland thinks of seemings as being like testimony in this regard; experiencing a seeming is like being told that something is the case by your sub-personal system responsible for categorizing perceived objects.³ Since seemings are products of our unconscious systems, they are also a-rational in three respects. First, we do not arrive at our seemings via a process of deliberation. The arborist did not conclude that the tree is an oak tree via a process of reasoning, it just spontaneously seems like an oak tree to her. Second, it doesn’t make sense to criticize a seeming for being irrational since they are not doxastic commitments on the part of the subject. Like a hallucination or illusion, a seeming can be “criticized” for being inaccurate, but not for being irrational. Third, seemings are insulated from rational revision; they tend to persist even in the face of evidence that they are inaccurate. If the expert hears that all the oak trees in the area died in a drought decades ago, this acts as a defeater on her judgement that the tree is an oak tree, but it is likely that the tree will nevertheless continue to seem like an oak tree to her.

If seemings exist and have the properties ascribed to them above, why think they are epistemically relevant?

Perceptual dogmatism about perceptual justification is a popular internalist theory that the defender of the seemings view could appeal to. According to perceptual dogmatism, having prima facie justification for a belief that p is as simple as it perceptually seeming to you that p. ⁴ You do not need to reflect on your experience to gain the justification, but if asked to do so, you are justified in citing how things seem to you—your seeming counts as evidence. Justification is therefore obtained rather easily. If this seems profligate, the dogmatist rejoinder is to point out that the justification is merely prima facie and is easily lost in the face of defeaters. If this seems ad hoc or under-argued, dogmatism gains its plausibility from the intuitive and parsimonious way it handles the threat of skepticism, as well as the array of cases where we have justification without engaging in deliberation. The main challenge taken on by the dogmatist is to define in a principled way the range of beliefs that are justified in this way, have it be broad enough to accommodate the perceptual judgements we intuitively take to be justified on the basis of perceptual experience alone, but not so broad as to entail that beliefs that are not intuitively justified on the basis of perception alone are so justified under the theory.

On the seemings view, the sensory component of perceptual experience is responsible for presenting an objective world with a certain arrangement of features and the seeming is responsible for providing the perceptual content that, according to dogmatism, provides immediate prima facie justification for beliefs. Since seemings are propositionally structured states that employ the concepts possessed by the subject, this view can explain how our perceptual experiences can have a range of contents broad enough to justify the correspondingly broad set of perceptual judgements we take to be immediately justified simply by pointing out that our perceptual seemings can employ the same concepts that get employed in our perceptual judgements. The worry remains that seemings can have such a broad range of contents that the dogmatist must accept unintuitive consequences about which judgements can be immediately justified on the basis of perception alone (e.g. that a cup of water is my cup of water, that a purse is expensive, that a cheese pairs well with Pinot Noir), but there are strategies available to the seemings theorist to restrict either the contents of seemings or the range of seemings that provide immediate justification. ⁵

To sum up, the seemings view consists of two parts. The first part is the positing of the existence of seemings—sui generis conscious mental states that are propositionally structured and conceptual but nondoxastic, reportive, a-rational, and spontaneously generated by our sub-personal categorizing systems. The second part is the marriage with perceptual dogmatism. Seemings contribute to epistemic justification because, according to dogmatism, perceptual experiences justify beliefs in virtue of having corresponding conscious content and seemings are the part of perceptual experiences which provide the experience’s conceptual content. If this is right we have an answer to the problem of the expert and the novice. The expert is justified in believing that the tree is an oak because the expert enjoys a seeming whose content is that the tree is an oak while the novice enjoys no such seeming.

Against the Seemings View

Although I am sympathetic to dogmatism about perceptual justification, I think the seemings view, as stated here, is wrong. ⁶ Either seemings must provide justification all by themselves or they provide justification only in combination with the sensory element of perceptual experience. The first view is unappealing for a number of reasons which will be discussed below. The second view raises the question of what kind of combination between a seeming and a sensory element is required for justification. I argue that a mere combination won’t do, and a non-seemings alternative is more appealing.

The possibility that seemings provide justification all by themselves is separate from the question of whether we can ever experience a seeming all by itself, with no corresponding sensory phenomenology, but it’s instructive to consider this latter question as well since reflecting on what a mere seeming would be like sheds light on why they cannot be epistemic justifiers all by themselves. Chris Tucker argues that the phenomenon of blindsight provides a reason for thinking that seemings can be experienced in the absence of sensory phenomenology. ⁷ Blindsighters have a condition that leaves them without conscious sensory experience of the world. They still have some working connections between their eyes and their brain, however, and although their conscious sight is gone, blindsighters do better than chance when guessing the kinds of objects placed in front of them. Tucker suggests that blindsight is explained by the presence of a seeming in the absence of sensation.

I consider this interpretation of blindsight implausible because it conflicts with the blindsighters’ own self-reports of what their experiences are like. Blindsighters report experiencing nothing during the trials, and that their responses are merely guesses.⁸ Seemings are supposed to be conscious though, so why wouldn’t they report them if they had them? But suppose that Tucker is correct. Are blindsighters justified in their beliefs? Intuitively they aren’t. Blindsighters are in no better a position than Bonjour’s clairvoyant when he spontaneously believes that the president is in New York, and this is a paradigm case of unjustified belief for the access internalist. ⁹ The blindsighter, as understood by Tucker, is different from Bonjour’s clairvoyant in one respect, and that is that the blindsighter’s seemings are conscious and pre-doxastic, whereas the clairvoyant has lost his doxastic agency—his clairvoyance puts beliefs right into him. But the motivating force behind our judgements of clairvoyance cases is that there is that there is nothing in the subject’s awareness that’s indicative of the truth of the proposition at hand, i.e. that is independent evidence for the proposition. Does a mere seeming indicate the truth of the proposition that is its content? No. The seeming represents its content as being the case, but it is not by itself an indication that things actually are that way anymore than an a posteriori belief represents its contents as being the case, but without being independent evidence for the truth of its content. ¹⁰

Perhaps perception is not the right comparison for seemings, and a better comparison of their epistemic force is with testimony. Reiland agrees, saying “In fact, I find it illuminating to think of seemings on the parallel of testimony. In the case of testimony, someone tells you that p…. Similarly, in the case of seemings your sub-personal system tells you that p.” ¹¹ I think this is right, given what mere seemings would be like. I therefore find it plausible that the justification that would be provided by a mere seeming would be of the same kind and strength as the justification provided by someone’s testimony. But this just means that a view on which seemings alone explain expert perceptual justification must be false since the justification provided by expert perception is very different from testimonial justification. For one thing, it is stronger. If I see a laptop in front of me, I will continue to be justified in believing that there is a laptop in front of me even if a trusted friend tells me that there isn’t a laptop in front of me. Only massive testimonial evidence is capable of defeating simple cases of perceptual justification. Another way that expert perceptual justification differs from testimony is by being immediate. The justification of testimony depends on background beliefs about the reliability of the testimonial source while the justification from perceptual experience does not rely on background beliefs about the reliability of perception (pace Crispin Wright?). Finally, testimony is arguably only capable of transmitting justification, whereas perception is capable of generating it. ¹²

A final reason for doubting that seemings alone can justify perceptual judgments stems from the demonstrative nature of many of those judgments. While some of our perceptually justified judgements are not demonstrative (e.g., “sheep exist”) some of our perceptually justified judgements are demonstrative (e.g., “that is a sheep”). This poses a problem for seemings-only theories of perceptual justification. Perceptual judgments with a demonstrative character rely on attention for fixing reference to the demonstrated object, and paying (visual) attention to something requires sensorily experiencing it. ¹³ Seemings, in the absence of sensory experience, are incapable of having demonstrative content. The thesis that seemings do all the epistemic work in perceptual experience is motivated by the assumption that seemings are responsible for grounding the conceptual content of the perceptual experience all by themselves, but since they don’t have a sensory phenomenology, seemings can’t have demonstrative content in the absence of sensation. But then we can’t say that seemings (all by themselves) justify our demonstrative judgements.

Seemings alone cannot justify beliefs, but the seemings view isn’t committed to this. A more plausible view is that combinations of seemings and sensations justify beliefs. But what kind of combination? Perceptually justified demonstrative judgements are a reason for thinking that it cannot be a mere a co-occurrence of the seeming and the sensations. Demonstrative judgements can only be justified by seemings that cannot exist independently of the appropriate sensory experience because the seeming is dependent on the sensory experience for its demonstrative content. But then what kind of relation must obtain between the seeming and the sensory experience? What establishes this relationship? An attentional link between the seeming and the sensory part is plausibly what the seemings view needs, but I think that the correct view of attention ultimately allows us to abandon our commitment to seemings as sui generis mental state altogether in favor of a view that attention links sensory experience simply with thoughts.

Attentional Structure

The “folk” or “spotlight” theory of attention conceives of attention as a “directing” of consciousness: to attend visually to an object is to direct our gaze at it, to attend to some idea is to think about it. This theory conceives of attention as something that consciousness does (or something that the subject does with his or her consciousness).

According to Sebastian Watzl, attention is not something that we do with consciousness; it is a structure of consciousness. ¹⁴ The elements of our consciousness stand in “priority relations” towards each other, with some elements of our consciousness being at the “attentional center” (these are the elements that have priority over all the other elements of our consciousness), others are at the “attentional periphery” (these are the elements that have less priority than some elements but more than others), and still more are at the “attentional fringe” (these elements of consciousness don’t have priority over any of the other elements). “Attending to x” can be analyzed as “the x elements of my consciousness have been given priority over the non-x elements.” As an example, Watzl discusses listening to a jazz album. At any given moment you are auditorily aware of many different sounds produced by the various instruments. You could attend to the entire ensemble as a whole—in this case all the sounds that make up your conscious experience of the music would have equal priority. But you could decide to focus on the trumpet—in this case the sound of the trumpet would be given priority over the sounds of all the other instruments, although awareness of the other instruments wouldn’t disappear altogether. In both cases the sounds of the music have priority over the feeling of your butt in your seat, which may be so peripheral in your consciousness as to have effectively disappeared.

Watzl’s theory of attention is different from the folk theory in ways I don’t have space to discuss, but one key difference is worth pointing out before I put his theory to work. According to the folk theory attention plays a merely causal role in determining what the phenomenology of an experience is like. The experience of attending to a trumpet is phenomenally different from the experience of attending to drums because trumpets sound different than drums. Sensory qualities constitute the phenomenology of experience; attention simply determines which ones are make a contribution at any given moment. But according to Watzl, attentional structure itself makes a distinct contribution to the phenomenology of an experience. There’s something it’s like to listen to a jazz album while giving priority to no instrument in particular that’s different from what it’s like to listen to the album while giving priority to the trumpet that’s explicable only in terms of how one structures one’s priority relations differently in the two cases. If Watzl is right, he has provided a tool for explaining the epistemic difference between expert and novice perceivers in terms of only what obtains internal to the subject. I think he’s right, and I think that attentional structure is the key to understanding how perceptual experience gains its content.

Go look at Figure 1 in the appendix. What is it an image of? Try to figure it out before moving on to the next paragraph.

Could you tell what it is an image of? I’ll tell you what it is, and when you look at it again it will click. It’s a picture of a man’s face. He’s looking straight out at you and he’s wearing a cowboy hat. The parts you probably thought were wings are the bills of his hat.

This example is useful because it recreates in rapid succession and in a single subject the difference between expert and novice perceivers and it does so in a way that illustrates the value of Watzl’s theory of attention for explaining this phenomenon.

First, before being able to see the image correctly you lacked justification for the proposition “that’s a man wearing a cowboy hat”—you were like a novice. But once you were able to see it correctly you gained justification for that proposition—you became an expert.

Second, the change in justification is plausibly explained by a change in the conceptual contents of your perceptual experience. At first, the conceptual contents of your experience were confused or impoverished. If you are like me you thought it was a silhouette of some sort of bug. But after learning to see it correctly the contents of your experience came to include “cowboy head.”

Third, there is a clear phenomenological difference between one’s experience prior to seeing the image correctly and one’s experience now that one can see it correctly. But this difference is not explainable by a difference in sensory quality. And it’s not explainable merely by a difference in the cognitive element of your experience. It’s a difference that’s inexplicable without the resources of Watzl’s structural theory of attention, but that is easily explained by it. The phenomenology changed because the way the attentional structure of your consciousness has changed. In the earlier experience, you probably gave priority to the black splotches in the image, relegating the white parts to the attentional periphery, or, in traditional gestalt terms, the “background.” Seeing the image as a cowboy head required a restructuring of your consciousness’s priority relations. Equal priority had to be given to the black and white sides of the image in order for them to form two complimentary sides of a face. And priority had to be given to the eyes, nose, and mouth parts of the image in order for the face to be seen correctly.

The above considerations make attentional structure a very promising candidate for explaining the relevant differences between expert and novice perceivers. If what makes someone an expert perceiver is their ability to structure the priority relations between the elements of their consciousness in the right way, then we have the desired result that our explanation also explains the phenomenological difference between the expert and the novice. If attentional structure can also ground the contents of perceptual experience, as the above consideration suggests, then adopting dogmatism will also allow us to explain the epistemic difference between the expert and the novice in terms of a difference in their attentional structures.

Attentionalism about Conceptual Content

I claim that attention is the resource best suited to explain the epistemic (and phenomenological) difference between expert and novice perceivers, once it is understood as a structure of consciousness as Watzl has proposed. I think attention can do the explanatory work that other internalists hope to do by positing sui generis seemings. But seemings did at least one bit of work that attention alone isn’t obviously capable of doing—they conveyed information. When you look around your environment you learn about what kinds of things populate it. Seemings theorists can explain this because seemings report their conceptual content. On the seemings view, learning about the kinds of things in your environment is simply a matter of taking your experience at face value. Attentional structure is neither propositional nor reportive, and it doesn’t seem to be necessary for structuring your consciousness a certain way that you employ some concept (even though top-down psychological effects can facilitate structuring your consciousness in a particular way, as is illustrated by the “clicking” phenomenon that occurred when you were told that image is a cowboy head). If attention can’t do this, then maybe, contrary to the suggestion above, attentional structure doesn’t ground the contents of perceptual experience.

I think attention does have a role in grounding the content of perceptual experience, though my view is unorthodox for a dogmatic internalist. On my view only thoughts have original conceptual content, only thoughts have concepts as parts. But our concepts for visually recognizable kinds have an attentional organization as part of their non-cognitive content. I call this attentionalism about conceptual content. The attentional content of a concept is what enables a thinker to have demonstrative thoughts utilizing that concept, since it’s the attentional content that determines how the thinker needs to organize their attention in order to pick out an instance of that concept. Perceptual experiences can then be said to have conceptual content derivatively when the attentional organization of the experience matches the attentional organization imprinted in the concept of an occurent thought. The ultimate explanation for why the expert arborist is justified in believing that the tree in front of him is an oak is that his perceptual experience is attentionally isomorphic to the attentional content of the [oak] concept that is part of his occurent thought that is an oak tree.

Consider sortalism, the view that demonstrative reference to an object relies on the classification of that object under a sortal. Amie Thomasson argues that sortalism must be correct for our nominative terms since otherwise reference would be ambiguous.¹⁵ For example, if the sortal “dog” wasn’t associated with the name “Spot,” it would be radically ambiguous whether a puppy, a location in space, the act of lying on the ground, or something else was named when a family named their puppy “Spot.” Thomasson then argues that the sortal associated with the name at the moment of the object’s baptism becomes part of the term’s content, continuing to serve as a necessary condition on the successful reference of the term, as well as establishing analyticities

I propose that something very similar occurs with attention and the initial learning of visually recognizable kind concepts. Learning to visually recognize instances of a kind requires being able to first attend to objects of that kind. As we have seen, this is a much more involved phenomenon than simply having the object focused at the spatial center of your visual. It means establishing the right priority relations between the parts of your consciousness so that some are more central and others are more peripheral. This “pattern” to the attentional structure of one’s experience is “imprinted” onto the concept when the concept is ostensively taught (by pointing out a paradigm instance) and becomes a non-cognitive part of the concept’s content. Not only is there something it’s like to experience the world with a certain attentional structure, there’s something it’s like to think demonstrative thoughts when you are attending to the world in the way imprinted in the concepts that compose the thoughts.

Bibliography

Audi, Robert. “The Place of Testimony in the Fabric of Knowledge and Justification.” American Philosophical Quarterly 34.4 (1997): 405-422.

BonJour, Laurence. “Externalist Theories of Empirical Knowledge.” Midwest Studies in Philosophy 5.1 (1980): 53-74.

Brogaard, Berit. “Phenomenal Seemings and Sensible Dogmatism.” In Seemings and Justification: New Essays on Dogmatism and Phenomenal Conservatism, edited by Chris Tucker, 270-289. Oxford University Press, 2013.

Campbell, John. “Visual Attention and the Epistemic Role of Consciousness.” In Attention: Philosophical and Psychological Essays, edited by Christopher Mole, Declan Smith, and Wayne Wu, 323-41. Oxford University Press, 2011.

Chudnoff, Elijah. Intuition. Oxford University Press, 2013.
Chudnoff, Elijah, and David Didomenico. “The Epistemic Unity of Perception.” Pacific Philosophical

Quarterly 96.4 (2015): 535-549.
Cullison, Andrew. “What Are Seemings?” Ratio 23.3 (2010): 260-274.
Lackey, Jennifer. Learning from Words: Testimony as a Source of Knowledge. Oxford University Press, 2008. Lyons, Jack. “Seemings and Justification.” Analysis 75.1 (2015): 153-164.

Montague, Michelle. “Cognitive Phenomenology and Conscious Thought.” Phenomenology and the Cognitive Sciences 15.2 (2016): 167-181.

Reiland, Indrek. “Experience, Seemings, and Evidence.” Pacific Philosophical Quarterly 96.4 (2015): 510-534.

Smithies, Declan. “The Phenomenal Basis of Epistemic Justification.” In New Waves in Philosophy of Mind, edited by Mark Sprevak and Jesper Kallestrup, 98-124. Palgrave Macmillan, 2014.

Stoerig, Petra. “Phenomenal Vision and Apperception: Evidence from Blindsight.” Mind & Language 12.2 (1997): 224-237.

Thomasson, Amie. Ordinary Objects. Oxford University Press, 2007.
Tucker, Chris. “Why Open-minded People Should Endorse Dogmatism.” Philosophical Perspectives

24.1 (2010): 529-545.

Watzl, Sebastian. Structuring Mind: The Nature of Attention and How It Shapes Consciousness. Oxford University Press, 2017.

Werner, Preston J. “Seemings: Still Dispositions to Believe.” Synthese 191.8 (2014): 1761-1774.

See Declan Smithies, “The Phenomenal Basis of Epistemic Justification,” in New Waves in Philosophy of Mind, ed. Mark Sprevak and Jesper Kallestrup (Palgrave Macmillan, 2014), 98-124, for an argument that a phenomenal requirement on justification best motivates access internalism.
Indrek Reiland, “Experience, Seemings, and Evidence,” Pacific Philosophical Quarterly 96.4 (2015): 510- 534, gives an authoritative statement of what seemings are supposed to be and a defense of their status as epistemic justifiers. Here I diverge from his terminology and emphasis only slightly.
Reiland, “Experience, Seemings, and Evidence,” 510-534.
See Elijah Chudnoff, Intuition (Oxford University Press, 2013) and Chris Tucker, “Why Open-minded People Should Endorse Dogmatism,” Philosophical Perspectives 24.1 (2010): 529-545.
See Berit Brogaard, “Phenomenal Seemings and Sensible Dogmatism,” in Seemings and Justification: New Essays on Dogmatism and Phenomenal Conservatism, ed. Chris Tucker (Oxford University Press, 2013), 270-289, for a strategy for restricting the range of seemings that provide dogmatist justification in a principled way.
There is much debate over what exactly seemings are. See Tucker, “Why Open-minded People Should Endorse Dogmatism,” and Andrew Cullison, “What Are Seemings?” Ratio 23.3 (2010): 260- 274, for defenses of the view assumed here. See Elijah Chudnoff and David Didomenico, “The Epistemic Unity of Perception,” Pacific Philosophical Quarterly 96.4 (2015): 535-549, and Preston J. Werner, “Seemings: Still Dispositions to Believe,” Synthese 191.8 (2014): 1761-1774, for two different views.
Tucker, “Why Open-minded People Should Endorse Dogmatism.”
See Petra Stoerig, “Phenomenal Vision and Apperception: Evidence from Blindsight,” Mind & Language 12.2 (1997): 226-227, for excerpts of blindsighters’ self-reports of experiencing nothing.
Lawrence Bonjour, “Externalist Theories of Empirical Knowledge,” Midwest Studies in Philosophy 5.1 (1980): 53-74.
Jack Lyons, “Seemings and Justification,” Analysis 75, no. 1 (2015): 153-164, and Reiland, “Experience, Seemings, and Evidence.”
See Reiland, “Experience, Seemings, and Evidence,” 13-14.
Robert Audi, “The Place of Testimony in the Fabric of Knowledge and Justification,” American Philosophical Quarterly 34.4 (1997): 405-422, makes this claim about testimony merely transmitting justification, but see Jennifer Lackey, Learning from Words: Testimony as a Source of Knowledge (Oxford University Press, 2008) for dissent on this point.
See John Campbell, “Visual Attention and the Epistemic Role of Consciousness,” in Attention: Philosophical and Psychological Essays, ed. Christopher Mole, Declan Smith, and Wayne Wu (Oxford University Press, 2011), 323-41, for an argument on the importance of attention for demonstrative reference.
Sebastian Watzl, Structuring Mind: The Nature of Attention and How It Shapes Consciousness (Oxford University Press, 2017).
Amie Thomasson, Ordinary Objects (Oxford: Oxford University Press, 2007).