The Contents of Perception

Susanna Siegel

Forthcoming in the Sage Encyclopedia of Perception, Ed. Bruce Goldstein


Suppose a subject recognizes George Washington on a dollar bill. It is an open question whether it is part of their visual experience that the person depicted is George Washington, or whether all that is visually experienced are colors and shapes, with the judgment that it is George Washington occurring farther downstream. According to Thin views about the contents of visual experience, these contents are limited to color, shape and illumination. According to Rich views, the contents of visual experience can involve more complex high-level features in addition to color, shape and illumination, such as personal identity (George Washington), kinds (banana, truck), semantic properties (meanings of words), and causation.  Rich views thus posit more informational richness in the contents of visual experience than Thin views. If Rich views are correct, then even if some visual experiences are thin (such as the experience of looking out of an airplane window into a clear blue sky, with nothing else in your peripheral visual field), visual experiences can represent that someone is George Washington, or that an object is a truck, or that the ball’s collision with the apple caused the apple to move.  If Thin views are correct, in contrast, then visual experiences can represent only that there is a layout of colored shapes, variously illuminated, where some part may be moving.


The debate between Thin and Rich views impacts the search for neural correlates of visual consciousness.  Antecedently held assumptions about whether visual experience is Rich or Thin may influence what researchers are prepared to count as a neural correlate of visual consciousness. If contents are Thin, then the neural correlates are more likely to be limited to early visual areas, such as V1 or V5. If contents are Rich, then the neural correlates are more likely to involve later areas, such as infero-temporal cortex and the fusiform face area.   


The debate between Rich and Thin views is of interest to psychologists because it bears on what the neural correlates of visual consciousness are, and on what cognitive processes are inputs to the underlying structures that give rise to visual consciousness. It also bears on the structure of disorders such as agnosia, in which subjects can see ordinary objects but cannot recognize them. If the Rich view is true, then agnosia may involve impoverished visual consciousness. If the Thin view is true, then agnosia may be a disorder downstream of visual consciousness, rather than a disorder involving visual consciousness itself. Finally, the debate bears on the cognitive structure of delusions such as Capgras syndrome, in which patients take a loved one to be an impostor. If a Rich view is true, so that visual experiences may represent such properties as ‘being an impostor’, then Capgras syndrome may be a normal response to an abnormal experience, consisting in taking things to be the way they appear in the (abnormal) experience. If a Thin view is true, in contrast, whether or not the visual experience of Capgras sufferers is itself abnormal, the response to the experience almost certainly is abnormal.  In that case, the disorder would consist in the abnormal response to evidence, perhaps in addition to an abnormal evidential base.


To philosophers the debate between Rich and Thin views is of interest because of its impact on epistemological issues concerning the evidence provided by perception. If a Rich view is true, then there are likely to be top-down influences on visual consciousness by mental states preceding perception, such as emotion, mood, desire, and beliefs. If such influences are pervasive, then the idea that our perceptual inputs are unaffected by antecedent mental states is probably false. While some top-down influences may help us perceive what we would otherwise be blind to (as when a radiologist sees a tumor on an x-ray), others may be epistemically pernicious (as it would if a depressive’s expectation that everything will look grey influences her experience, so that indeed everything ends up looking grey).


The case of causation illustrates both the debate between Rich and Thin views, and the difficulty in settling it. Whereas a Rich view allows that visual experience can represent that the ball’s movement causes the apple’s movement, a Thin view will say that only succession of movements is represented in visual experience. In a series of experiments published in the 1960’s, Albert Michotte showed adults a range of scenes that elicited causal descriptions. Many of these scenes consisted in ‘launching’: an object A moved toward a stationary object B, makes contact with B, and immediately afterward B starts moving in the same direction as A was moving at the time of contact. Through an extensive series of experiments involving a wide range of stimuli (including shadows, moving lights, marks on paper, and hefty objects), Michotte tried to isolate the parameters of motion that led adults to classify what they saw as causation. He discovered that no matter even when A or B or both are lights or shadows, adults report that object A launches object B exactly when the separate motions of A and B are consistent with a single motion that is transferred from A to B. This led him to posit an ‘impression of causation’, since subjects knew that shadows and lights cannot launch anything.

Though these results are suggestive, they do not conclusively favor Rich views over Thin ones. To get from Michotte’s results to a Rich view, what’s needed is a principle linking verbal reports to the contents of visual experience. The fact that the verbal report mentions causation does not show that the visual experience represents causation, without some further principle indicating that the reports are reports of experiences, rather than reports of a mental state downstream of experience. For all Michotte’s results show, visual experiences may represent the spatio-temporal parameters of motion that elicit causal reports, rather than representing causation itself. If so, then experiences in the ‘launching’ condition when the stimuli are lights and shadows will be accurate, even though there is no causation. In contrast, if the experiences represent causation (and not just the parameters that elicit causal reports), such experiences would be inaccurate, whether or not subjects would go on to infer that A launched B.

The issue between Rich and Thin views arises for auditory as well as visual experience. We can often recognize who is speaking by their voice. Rich views allow that auditory experience can represent that Frank Sinatra is singing, whereas Thin views allow that it can only represent the sounds and their qualities. The judgment that the person singing is (or sounds like) Frank Sinatra must come downstream of auditory consciousness. Rich views allow that auditory experience can represent that a truck is going by, whereas Thin views allow that it can only represent a gradually fading groaning noise. Rich views allow that auditory experience represent that a string of sounds was a meaningful sentence, whereas according to Thin views it can only represent that a string of sounds.


Like visual experiences, auditory experiences represent the world outside the perceiver, and thus can be accurate or inaccurate. Going with this, auditory and visual experiences have spatial content: we hear sounds as coming from different directions, and we see things as having specific spatial locations. Whether a standard visual experience such as the one had when viewing a cube is accurate depends on whether the cube has the properties (including spatial properties) that the visual experience represents it as having. Whether standard auditory experiences such as the ones had when listening to a lecture are accurate depends on whether the lecturer made the noises that the subject auditorily experiences her as making. The exact accuracy conditions of the experiences thus depend on whether Rich or Thin views are true. For instance, suppose you are talking to an impostor on the telephone, but the disguised voice fools you and you wrongly take the impostor to be your spouse. There is clearly an error at the level of judgment, but is there an error at the level of auditory experience as well? If a Thin view is true, then your auditory experience may be correct in this case, because it represents the tone of voice that the impostor actually has. In contrast, if a Rich view is true and your visual experience represents that the person you’re hearing is your spouse, then your error will reach all the way down to the auditory experience itself. 



Susanna Siegel


See also: Intentionality and Perception; Consciousness: neural correlates; Vision: Cognitive Influences; Philosophical Approaches to Perception


Further Readings:

Bayne, T (2009) “Perception and the Reach of Phenomenal Content”. Philosophical Quarterly.

Byrne, A. (2009) “Experience and Content”. Philosophical Quarterly.

Matthen, M. (2009) “The Feeling of Presence”, In N. Gangopadhyay, M. Madary, and F. Spicer (Eds). Perception, Action and Consciousness. Oxford: Oxford University Press.

Siegel, S. (2005). Which Properties are Represented in Perception? In T. Gendler and J. Hawthorne (Eds), Perceptual Experience (pp. 481-503). Oxford: Oxford University Press. 

Siegel, S. (2009) The Visual Experience of Causation. Philosophical Quarterly.