Friday, August 15, 2014

Different Looks
In 1976, when I delivered the John Locke Lectures at Oxford, I often spent time with Peter Strawson, and one day at lunch he made a remark I have never been able to forget. He said, "Surely half the pleasure of life is sardonic comment on the passing show".  This blog is devoted to comments, not all of them sardonic, on the passing philosophical show.
Hilary Putnam

In my previous post I said that "what looks are, and the distinction I draw between objective looks and subjective looks, will be taken up next in this series of posts.” The present post cashes this promissory note.

Looks are dispositional properties of objects not qualia 
In accordance with the grammar of such noun phrases as “the pink look of your dress in this sunset light”, I take a look to be a property of an object. There is another use of “look” in which a real object is not required, viz in sentences of the form: “It looked to P as if…… The function of such locutions is usually to report a misperception or an illusion or a hallucination (“it looked to Mr. Tipsy as if there were a pink elephant in front of him”); the statement does not imply that any real object had a pink elephantish look. [As a comment from "Anonymous" reminds us, there is also the use to "hedge" an assertion - "It looks like the runner was safe".]   But if I truly say that an object looks so-and-so, then, regardless of whether it is actually so-and-so, in those circumstances, it has a disposition to seem so-and-so to the person in question. In precisely what sense it “seems so-and-so” distinguishes the different sorts of looks I will describe, but in all of these cases, if I speak truly, then I describe a disposition that a real object has at least momentarily.

It follows that whatever looks may be, they are not qualia. Qualia, in my view are properties of an organism (in today’s post, a human being). [In a forthcoming paper, Hilla Jacobson and I call this position “attributeism”, and we distinguish it from “adverbialism” which is a particular form of attributeism,  but that is not our subject today. Reichenbach, in Experience and Prediction was an attributeism but not an adverbialist. Attributeism is a position about the ontology of qualia (Reichenbach’s “impressions”), namely that they are attributes of organisms, while Adverbialism is a position, invented I believe by Ducasse[1], about the real logical form of statements like “I experience a red circular sense datum” – a logical form which is thought to be misleadingly expressed by the seven word sentence in quotation marks.] Others think of qualia as properties of “experience”, or as mental events. But in any case, no quale is a property of a vase or wall, although vases and walls are precisely the sorts of things that can have “looks”.

In my “promissory note” I spoke of a distinction between  “objective looks and subjective looks”, but a more complex taxonomy is certainly called for. In this post I will briefly indicate what I mean by “subjective looks”,  but leave further discussion to future posts.

Subjective looks
In the previous post the example of a “subjective look” was the exact shade something appears to me to be when I view it with my right eye closed and the exact shade it appears to me to be when I view it with the left eye closed. Because of variations in the pigmentation in the macular areas of the two eyes, those shades seem slightly different, but neither eye “sees it wrong”. The visual qualia are different, and neither eye “misrepresents” the object I look at (a gray wall or the sand on a beach), The difference in the qualia I experience are, to use Ned Block’s term, “ineffable”, not describable in public language as it stands now (see the previous post).  As Block points out[2], “…we can refer to [a quale] by saying ‘What it is like for that person to see red’.  What we cannot find is a color name ‘F’, such that what it is like for one of these people to see red can be expressed in the form ‘looking F’, and in that sense we can say that the experiential property is an ineffable quale.” Using this quale-terminology, subjective looks are dispositions to affect the visual qualia of the viewer; this is what Russell thought colors  are, but that was a mistake, as pointed out in my previous post. The visual system is designed to represent real colors, not subjective looks, and certainly not to represent “the sense data of normal viewers”.

More has to be said about subjective looks, in particular about what has been called their “transparency” or “diaphanousness”, but that is part of what I am leaving for later posts.
Intersubjective looks
I close with some remarks about looks that people with normal vision agree objects have under certain circumstances (which may be very deceptive, e.g., a white dress when  illuminated by a red light). But that is a genus that comprises more than one species (and I make no pretense to being able to list all of those species). For our purposes, we shall say something about just three species of intersubjective looks, namely:
(a) fully objective looks (ones that can be painted or photographed).
(b) illusions like the Müller-Lyer that do not depend on the background against which the object is viewed.
(c) background-dependent illusions.

Fully objective looks
So my readers don’t have to scroll back to my post of July 23rd (“Colors and Their Objective Basis”), let me remind you that in that post I endorsed something that Alex Byrne and David R. Hilbert wrote[3], namely:

But is there a physical property that all and only (actual and possible) green objects share? Given our assumption about the general correctness of our color perceptions, the answer is (plausibly) yes. The property that all green objects share is a type of SSR [=Surface Spectral Reflectance - HP]. Very roughly, this property -- call it ‘SSRGREEN’ -- is the type of SSR that allows an object in normal illumination to reflect significantly more light in the middle-wavelength part of the spectrum than in the long-wavelength part, and approximately the same amount of light in the short-wavelength part as in the rest. Obviously, particular reflectances meeting these specifications -- for instance those of frogs, lettuce, and dollar bills -- may be otherwise very different.”

However, in that same place, Byrne and Hilbert give a somewhat exaggerated account of color constancy, when they write,
If two objects have the same SSR, in all visible illuminations they will reflect the same amount of light. If one object is substituted for another with the same SSR (assuming they are the same size) in the scene before the eyes, no visible color difference will result. The SSR of an object is (typically) an illumination-independent property: the SSR of an object does not change if the object is taken from a room to a sunny street, or if the lights are turned out. And this, arguably, is also a feature of color.”

This can only be an intentional oversimplification. If things were that simple the white dress would still look white even when there is a gorgeous red sunset! In fact, writers on color seriously qualify what they say about color “constancy”, (although the ones I have seen fail to distinguish objective failures of “constancy”, failures that can be “captured” by a good painter or photographer, and subjective and even idiosyncratic failures of subjective colors to be “constant”.  An example of such qualification is the following by David Briggs (Dimensions of Color,

For an object in a natural scene, the capacity of our visual system for colour constancy means that its perceived hue is determined to a large extent by the dominant wavelength of its reflectance under white light, as long as the illumination of the scene is not too strongly coloured. Nevertheless, an object can vary considerably in perceived hue depending on surrounding, interspersed and previously viewed colours, and even on the attitude of the viewer. In viewing an image, our capacity to extract object colour information from a visual scene can cause us to perceive hues very different from the actual image hues.”  [emphasis added –HP][4]
As for the “ontology” of such objective looks: when the look of a color (hue) in a particular situation can be displayed by a photograph or a painting, the “objective look” is simply the color shown by the photograph or painting; the “look” of one hue can sometimes be a different hue.
illusions like the Müller-Lyer[5] that do not depend on the background against which the object is viewed
When we move from hues to lengths a difficulty emerges! A photograph of a Müller-Lyer illusion is just another Müller-Lyer illusion; what makes it an illusion is that two lines appear to be of unequal length when they are actually the same length; but the lines in the photograph are also of the same length (measure them!) and they too appear to be of unequal length.  The photograph qua physical object does not show “the unequal lengths the lines appear to have”, in the way that a photograph of a white dress in red light can show the color the dress appears to have. In that sense, the “look” of the lines is not fully objective, yet, since everyone appears to describe the illusion in the same way, it is an intersubjective look.
Background dependent illusions
Andrew T. Young writes (“Introduction to Color”):
Incidentally, what the human visual system considers “light” or “dark” depends very strongly on the perceived context. If I had displayed the orange square here[6] against a white background instead of a medium-gray one, you'd have perceived it as brown rather than orange. There are some vivid demonstrations of this context effect at Edward Adelson's checker-shadow illusion,, and the paper by R. Beau Lotto and Dale Purves, “An empirical explanation of color contrast” (Pub. National Acad. Sci. 97, 12834–12839 (2000), which you can download as a PDF file if your institution subscribes to PNAS.”
Although the “objectivity” of the look (orange or brown) is virtually nil, the intersubjectivity is apparently as great as that of the Müller-Lyer illusion.

The account according to which all this is explained by similarities in the sense-data (qualia) of normal subjects suffers from a lack of present empirical evidence. What is clear is that (1) our visual systems are built to construct representations of the color – the real color – of surfaces from data such as arrays of light falling on our retinas, data that woefully underdetermine the “right” representation, and they get it right in an enormous range of cases. They do that by using clever algorithms (Edwin Land was the pioneer here with his Retinex – retina+cortex – algorithm). But the algorithms work only under certain conditions of illumination and background. Thus various representations and misrepresentations sometimes arise. This explanation` does not require us to know how similar or dissimilar the qualia of different people actually are.
to be continued

[1] See
[2] Ned Block, “Wittgenstein and Qualia,” in M. Baghramian, Reading Putnam. However, Block identifies subjective looks with the corresponding qualia, and I disagree for the reason just given.
[3] Byrne and Hilbert Readings on Color, Volume 1: The Philosophy of Color, MIT Press, 1997; online at My agreement with Byrne on the physical nature of the colors of surfaces does not imply agreement with his “intentionalist” view of the nature of subjective  colors, and of phenomenal experiences generally, namely that “the propositional content of experiences in a certain modality (for example vision) determines their phenomenal character. In other words, there can be no difference in phenomenal character without a difference in content.” [Alex Byrne, “Intentionalism Defended”, The Philosophical Review, Vol. 110, No. 2 (April 2001) , p. 204. Emphasis in original.] This disagreement will be the subject of a future post.
[4] For a review of the literature see David H. Foster, “Color Constancy”, Vision Research 51 (2011) 674–700.


  1. It looks like the use of the lexeme 'look' in that sense indicates a hedge on the assertion, which would be whatever is within the scope of that verbal element. In using such a hedge (e.g., "That cup looks red." Also, 'seems' seems to have a similar function.), the speaker signals a less than normal degree of confidence in the accuracy of the report. This sense of reduced confidence could be because of nonnormal environmental conditions between the eye and the object (e.g., nonnormal illumination), or due to nonnormal physiological conditions within the subject (e.g., perhaps jaundice). The surface reflectancy profile of the object remains constant under all of these distorting conditions, but the distortions could also, in principle at least, be given an objective empirical analysis and explanation. One could discover and state a causal law determining the effect of varying illumination conditions on the normal visual system, and in the subject one could explain distortions of the normal experience of the reflectancy profile as caused by changes in the cellular physiology of the relevant visual pathways. So the normal and the distortions could all be given an objective description. As for the experiences, the observer must rely on the reports of the subject, but perhaps this could be helped by allowing the subject to match the experienced colour with a palette of possibilities, which can be compared by the observer to the object under normal conditions.

  2. nice and informative blog post.