by Carolyn Dicey Jennings
The following ideas and arguments were central to my dissertation work, and are now published as an article in Philosophical Studies. I include them below in a much shortened format for those readers short on time, but high on interest (but hopefully not literally).
The ultimate claim of this work is that top-down attention is necessary for conscious perception. (I argue elsewhere that attention is not necessary for conscious experience, in general.) That is, we might ask the question: what is the contribution of attention to perceptual experience? Within cognitive science, attention is known to contribute to the organization of sensory features into perceptual objects, or object-based organization. I argue something else: that attention enables the perceptual system to achieve the most fundamental form of perceptual organization: subject-based organization. That is, I argue that subject-based organization is brought about and maintained through top-down attention. Thus, top-down attention is necessary for conscious perception in so far as it is necessary for bringing about and maintaining the subject-based organization of perceptual experience.
So what is "Subject-Based Organization"?
Here is an example: Imagine that you are in a room with dark gray walls, ceiling, and floor and a single source of dim overhead light. In that room is a sitting cushion and a small table. On the table is a steaming pot of tea. Imagine that you are sitting on the cushion facing the teapot with your eyes open, fixated on the teapot, breathing steadily and slowly. To yield the phenomenal contrast, compare two versions of this experience in your mind. In the first version of this experience, imagine that you are interested in observing the steam from the pot of tea. In the second version of this experience, imagine that you are interested in observing your breath, as in a session of yogic meditation. Nothing else about the situation should be different except for this difference in interest. In that case, does your perceptual experience differ between these two scenarios? I suspect that for many readers the steam is in the foreground when interested in the steam, and the breath is in the foreground when interested in the breath. What do I mean by foreground? One might experience foreground in terms of epistemic proximity: the stimulus of interest might appear more available for knowledge, whereas the background stimulus might appear less available for knowledge. Or one might experience foreground in terms of importance, meaningfulness, or richness: the background stimulus might seem less subjectively important, meaningful, and rich than the foreground stimulus. However one experiences the difference between foreground and background, the difference reveals a structural feature of experience that corresponds with the subject’s interests, which I am calling “subject-based organization.”
Here is some empirical and conceptual background: work on the perceptual organization has mostly centered on the organization of perceptual objects, as in the the Feature-Integration Theory (Treisman & Gelade 1980). Although FIT has been widely criticized in philosophy and cognitive science, some of its claims withstand those criticisms. Specifically, it is generally accepted that 1) certain visual features are first processed in relative isolation and only later grouped into objects, 2) salience information is processed alongside this feature-based information, enabling visual search, and 3) top-down attention stimulates and/or suppresses this salience information according to the current goals of the subject. These claims leave open the question of how early sensory processing relates to perceptual experience. That is, if we assume that perceptual experience is rooted in early sensory processing, we might ask whether that processing is sufficient for perceptual experience (e.g. Lamme 2004; Block 2007) or whether some further processing, such as the top-down feedback of attention, is required to bring about conscious perception (e.g. Dehaene & Naccache 2001).
This is my addition to the debate: top-down attention is necessary for subject-based organization, and subject-based organization is essential for conscious perception. Importantly, although subject-based organization is argued to be essential to conscious perception, it is not here argued to be essential to consciousness, in general (see the link above).
My conceptual reasoning is basically this: the concept of perception essentially includes the concept of subject-based organization. That is, the concept of perception is normally understood in contrast with the concept of sensation, and the key difference between these is that perception, unlike sensation, is informational to the subject of perception. For perception to be informational to the perceiving subject, it must have some minimal degree of organization that accords with the perspective of the subject. In other words, it must have subject-based organization.
But I also argue that subject-based organization is crucial to conscious perception by ruling out other forms of organization that might provide the essential structure:
1) Feature-based organization is the organization of sensory input into sensory features. This type of organization, seemingly basic, actually presupposes other forms of organization: we perceive a bright red line or a dull gray cross, but not simply redness or dullness. Further, conscious perception does not always involve feature-based organization.
2) Object-based organization is the organization of sensory features into perceptual objects. Although an important feature of conscious perception, this form of organization is neither universal nor fundamental to conscious perception.
3) Space-based organization is the organization of features, patterns and objects into a common space-time matrix. This form of organization is nearly universal to conscious perception, but it is not the most fundamental form of perceptual organization, since it cannot account for whether a percept is or is not perceived when viewing, for example, a bistable stimulus.
4) Field-based organization is the organization of parts of a perceptual field by other parts of the perceptual field or by the perceptual field as a whole. However, even if field-based organization can explain organic switching between bistable percepts, it fails to explain the subject’s apparent control over this switching.
5) Subject-based organization can account for perceptual differences that other forms of organization cannot, such as the difference between a split experience and a switch experience, and seems to be present in all of the above cases. Thus, I conclude, subject-based organization is the best candidate for the most fundamental and universal form of perceptual organization.
Here is the overall argument for the ultimate claim (that top-down attention is necessary for conscious perception):
1) Every perceptual experience must have some degree of subject-based organization to count as an instance of perception.
2) Early sensory processing does not have subject-based organization.
3) Some process is required to bring about subject-based organization for sensory input in order to bring about conscious perception.
4) Only top-down attention can organize sensory input according to the current tasks or interests of the subject.
5) Top-down attention is required to bring about subject-based organization for sensory input in order to bring about conscious perception.
6) Therefore, attention is necessary for conscious perception.
This is, of course, a highly controversial ultimate claim. (I presented this work at two separate symposia that involved shouting!) I present objections and responses for nearly every claim that is made in the paper (which have been helpfully supplied by audience members at different presentations), which readers can also find at my website. But I love to hear new objections, and this is a work in progress. Object away! Alternatively, if you have work that is complementary to this, I would love to hear about it!