The Problem With Mary's Room

Metaphysics & Epistemology
1

First, the background:

The thought experiment describes Mary, a scientist who exists in a black-and-white world where she has extensive access to physical descriptions of color, but no actual perceptual experience of color. Mary has learned everything there is to learn about color, but she has never actually experienced it for herself. The central question of the thought experiment is whether Mary will gain new knowledge when she goes outside of the colorless world and experiences seeing in color.

Now the problem:

We are told that she is in a “black and white” room. But what exactly does this mean ?

Do we imagine taking a tour of the room as a group and agreeing, verbally, that “yes it’s all black and white in here” ?

But I don’t see the room through your eyes nor you through mine. I just believe that we all called the room “only black and white.” We made similar sounds. That was enough for us.

In the quote, this black-and-white-ness is suggested/assumed as “absolute” or “objective.” Fine maybe for other topics, but this paper is about the perception of color.

When Mary leaves the room and “experiences color” ? What does that mean ? Perhaps after the tour that we’ve taken, we leave the room together and agree, verbally, that “there are colors other than black and white out here.”

Would Mary upon exiting say “oh so that’s what red is like?” Because she recognizes a rose ? Or has a device that associates a frequency with what she perceives ? And she would perceive this frequency as a numeral on her device.

What are “physical descriptions” of color ? What can they be ? Mary might learn from a book that blood is “called” red. That most people call objects that emit certain wavelengths “red.” And so on.

The problem with Mary’s Room, which is great anyway, is that we “look down” on Mary with an amusing arrogance about “our” knowledge of color. But “our” knowledge of color looks much like Mary’s when she’s in the room.

And yet “things are colorful,” yes ?

2

And perhaps continue… “Funny, because it looks the same as my grey table? Although… now that I look, the table seems a bit duller than before. So do a lot of things in the room. Hmm. I suppose I might be re-calibrating?”

2 Likes
3

Her exit gets weirder the more I think about it.

If she’s a philosopher, she might say : so that rose would probably be called “red.” But can I “say” more ?

I think it’s reasonable to say that Mary doesn’t “learn” anything. If “knowledge” is generalizable or structural. The logical positivists talked like this. Acknowledging “ineffable quality” in some sense but seeing that nothing could be done with it. Sort of “the myth of the given.” I can reason from “the rose is red” but not from the pre-articulated “quality” that “drives” such an assertion.

4

It is, of course, impossible to have someone in Mary’s situation in the first place. Her skin has been something other than the exact black or white of the room she’s always been in. If she ever got a cut anywhere in her entire life, sheer saw red blood.

Maybe a better idea is that she was born with eyes that were entirely unable to see anything but black and white. Then, after having learned all that she is said to have learned, she got an eye transplant. So she would have seen roses, stop signs, and tomatoes, known they are all said to be red, and understood that she was seeing red when seeing them all.

5

So I guess the issue is that she is seeing what-tends-to-get-called “red.” And maybe upon reflection she comes to believe that this is only “structural” or “behaviorist” red. The “surplus” is “ineffable” in that it “exceeds” the pragmatic concept. Yet presumably that “ineffable quality” (“personal red”) is important for her ability to recognize the object as one that people call “red.” The rose is “there for her” in that “personal red.” The “personal red” is an indicator of the “public color” of a thing. The public color is in some sense “not a color.” In another sense, it’s the “essence” of the color.

6

The question is of what relevance is the thought experiment. If we root meaning in use, the existence or non-existence of qualia becomes irrelevant, not non-existent. Their irrelevance protects against the incoherence of a system that attempts to root meaning in publicly inaccessible mental images.

So, sure, Mary before seeimg red knows more than before, showing that propositional statements cannot fully educate without additional experiential information.

The bigger issue (for me at least) is what we say of that inner state other than to say it’s there but is can’t impact how we form language.

1 Like
7

At least part of this issue can only be addressed empirically. What Mary or other humans would be likely to say in the situation described is a matter of fact, not philosophy. Unfortunately, as far as I can tell, there are no studies that address this exact issue. There are some which address somewhat similar situations which might be of interest, but which won’t satisfy our curiosity about this specific set of circumstances.

8

So do I. How would we describe what Mary learns? It would have to be, more or less, “Mary has learned what it’s like to see red,” or perhaps, “Mary has learned what red looks like.” Both these suggestions involve the tricky “what it’s like” construction, which may or not be about knowing something.

But . . . Mary has absolutely had a new experience, which may be all the thought experiment is trying to demonstrate. To deny that (assuming you’re willing to buy the whole “black and white world” set-up) would be to deny that experience outstrips knowledge, and I can’t see how to defend that.

9

Yes, that was my point in “the harder problem of quiddity.” And this works fine for some sort of sheer phenomenal taste, sight, etc. (although whether this is true to how the world is actually experienced is another question).

The problem, from where I stand, is that the very shape of the description already assumes a lot. Mary “already knows” all there is to know about color because she knows what exactly?

Quiddity, the whatness of a thing, that it is a dog, a daffodil, a dove, deuterium, or even a color or ‘the experience of color,’ turns out to be seemingly quite impossible to deduce from physics qua “fundamental” mechanism/microphysics. I will put in defense of this claim that fact that most of the advocates of reductionism seem to gravitate towards either extreme forms of nominalism (human desire is the ground of anything being any thing) or a very austere realism (reduction to mathematics, which then seems to exclude all things as such). The work in analytic philosophy on composition also points in this direction, as mereological nihilism and universalism would seem to render the denoting of any particular whole or kind arbitrary, whereas popular forms of restricted composition (e.g., Fine, Jarwolski, etc.) find themselves wholly unable to resist the arbitrariness trap as well.

So to get to “color” we would have to seem to somehow know what we wanted to explain in the first place, in order to pull it out of the particle/field soup. Similarly, since, once cause is reduced to accidentally ordered temporal mechanism, causes become seemingly impossible to isolate (another point often embraced by physicalists), and so it becomes impossible to explain the causes of any one thing or type, without one sort of extrinsic imposition of finality to scope the analysis.

I would then say that the theologically motivated separation of man from nature in early modern thought has to do a ton of heavy lifting here, because man has to stand outside all this to impose any meaning at all. That is, finality still has to come from somewhere. But does this become perhaps viciously circular if man is also in the order of nature?

D.C. Schindler discusses this in The Catholicty of Reason:

As for final causality, it represents an explanation of the meaning of things, and not simply an arbitrary imposition, only insofar as teleology is taken to be most fundamentally intrinsic.If there is no intrinsic relationship between a being and the purpose it serves, if, in other words, the purpose is simply extrinsic to a being, then it becomes wholly accidental that it happens to be this particular being that serves the purpose, and not some other. Things become interchangeable with respect to their purpose, and represent nothing more than instruments in its service. The purpose, in this case, does not illuminate the meaning of the being, which is to say it has no strictly theoretical role, but as we saw above dissolves into a kind of positivistic pragmatism that is never truly self-explicating but only ever endlessly self- justifying, and indeed, always in terms other than itself…

As we mentioned above with reference to Spaemann, even a wholly “positivistic” view of causality derives whatever intelligibility it possesses from an implicit affirmation of teleology. One cannot distinguish a cause from the essentially infinite number of conditions preceding the effect with- out some minimal reference to final causality: this reality differs from the others in that it acts “for the sake of” this effect; its activity has the purpose of producing such and such an effect.If there is nothing but wholly extrinsic relations, it would make no sense to distinguish a “post hoc, propter hoc” fallacy from a valid analysis of a causal relation, because there would be only “posts” and no “propter.” Thus, not only would we lack a basis for attribut- ing any necessity to the connection between cause and effect, but we would in fact have no way of identifying any causes, which means we would also lose the ability to identify something as an effect, insofar as doing so depends on identifying a cause. Along with necessity, there would be no such thing as probability.

10

No, that’s not what is meant. The point of Mary’s room is to highlight the distinction between light as a phenomenon, and the experience, and how complete knowledge of the former does not entail the latter.

So, in terms of the set up of the room, there’s no need to go into subjective descriptions. We can say she’s in a room where there is a mix of photons of wavelength 500, 600 and 700nm such that all the cones of a human eye would fire equally. Or, she’s just wearing glasses that can only emit equal amounts of these three kinds of photons at each pixel.

disclaimer: I don’t think RGB maps exactly to 500, 600 and 700 nm; it’s a rounding to make things simpler

11

I’m not so sure Mary could actually learn “what red is” or “what red is like” since both those descriptions imply a sort of universal objectivity to the “red” experience she is having. But Mary, like all of us, can’t have proof that the “red” of others is the same as her “red”. With this in mind, the only consistent/objective facts about red are the properties that can be measured, like wavelength etc. With that said, Mary probably(?) experiences something different when. she sees a tomato or a cherry than when she looks at a table which is “black” by wavelength. I’m not sure if this difference counts as knowledge or is simply an adjustment in her perceptive “filter” so to speak (like when you adjust the saturation of a digital image).

12

Fun fact: the philosopher Frank Jackson came up with Mary’s Room as an argument against Physicalism. He has since become a physicalist and thinks Mary’s Room is a poor argument against Physicalism.

1 Like
13

Good point. To suggest she knows black and white better because she has experienced it but to know red less only because her knowledge of it is propositional, bakes in the conclusion that the experential qualitative knowledge exists.

The best that can be said of qualia is that they may exist, but to say more attempts to use language, a public use tool, to describe a private state. It is a category error.

14

Not sure about that. One can produce perception of color by dividing a flat sphere into two halves, with one half painted black and the other painted white. Then a series of black lines are drawn on the white half , running parallel to the curvature of the disc. When this disc is spun, the black curved lines appear colored, and which color they appear as depend on whether they are perceived as emerging from the black side of the disc or the white side. Warm colors appear as a movement from dark to light, and cool colors from light to dark.

This suggests color is a perceptual construction simulating a perceived dynamism within objects, their apparent receding away or coming toward, activity vs passivity. This may be why warm colors are likened to warm temperature and ‘cool’ colors to cold temperature. Hot and cold are heightenings vs lessenings of intensity of activity. The same is true with emotions (feeling blue or red with rage).

From this vantage, experiencing color for the first time is like detecting a movement within objects which was undetectable previously. That sounds like a more useful form of learning than some ineffable qualia.

1 Like
15

Yes, good point. But what then is “public?” Aside from having the primary/secondary qualities distinction driven into us from grade school, I’m not sure how exactly it continues to justify itself. Identifying “public” correlates of redness is radically dependent on the experience of redness by the investigators.

Plus, in virtue of what is redness private, but dogness or treeness not? Certainly anyone with a functioning set of eyes can publicly point out either. But perhaps dogness and treeness are also private. Or why is mathematics paradigmatically public? Is it? And even if it is, the application of mathematics in the natural sciences seems irredeemably bound up in the private.

Plus, it’s hard to see how one could learn or identify one, two, three, etc., or half, a quarter, without the seemingly “subjective” notions of wholes, since you need “duck” to even conceive of “three ducks” or “half a duck.”

Dennett has it that the redness is irrelevant, and all we have is the data from recordings of people saying “this looks red.” But then the meaning of those words is also seemingly private in the way redness is.

I’m not convinced the primary/secondary qualities distinction isn’t ultimately arbitrary.

16

Let me stress that I completely reject the postulation of “qualia.” For me this is instead about the “quality” or “presence” of shared lifeworld objects. I once knew a guy that “hated jazz.” I like jazz. Is jazz “not really there” in terms of a “pleasant” or “unpleasant” quality respectively ?

For me the relevance is to empirical science. What is perception ? A measurement seems to involve a categorization of “too much quality” so that a useful quantitative “extract” results. What is left of the object if the quality is removed ? Yet we tend to agree on the extract ? Is the quality unnecessary ? Can I measure the length of what is not visible or tactile ?

Do we need to do philosophy of science ? No. Part of the “glory” of science is that power is knowledge. That inversion is meant to suggest that we trust what gives us what we want and fear what can destroy it. But why do we want what we want ? What is wanting ? How do signs have a “use” if we aren’t trying to do something ?

If an android avoids rain, because it is programmed to flee humidity, we don’t “feel bad” for it if it gets wet. But we feel bad for those who suffer. Or do we feel bad for those who look like the humans we call “suffering” ? This is a sincere question, if also a “useless-to-most” question. But I ask it. So I have some “use” for the signs. But what if I’m a P-zombie ? (I can’t believe that, but others might play with the idea. Do P-zombies have a “use” for anything ?)

Right. But what is experiential information ? As opposed to “pure inference,” as in a formal system ? Is “empirical science” a delusion ? Is there just “grammar” that is independent of the “quality” of objects altogether ?

To me that “inner state” is the quality-for-me of the object itself. Surely empirical objects affect the signs we use and how we use them ? But are there empirical objects without “quality” ?

17

I basically agree. I mean that what is said is relatively objective. Weirdly, as Saussure points out, this involves a researcher “categorizing” the sound that streams from a mouth sa the saying of an iterable sentence. This iterable sentence is something like an equivalence class of soundsteams, an equivalence class that plays a “role” (like [ friendly greeting ] ) in the “pre-theoretical lifeworld” of that community. This knowhow with language is presupposed by science, and, if science investigates it, it depends on it as it does so.

18

Right. “Knowing” is often framed “logocentrically.” It’s conceptual and “above mere quality.” Yet one “knows one’s way around” and “knows how to ride a bike.”

I definitely personally believe that experience outstrips knowledge. I’m trying to point out a certain “forgetting” in the set up. We ourselves are — sort of — in the position of Mary. But we “pretend” to “directly experience” an “absolute black and white.”

Yet it seems that we too can only “trade words.” I can’t see the rose through your eyes. But I can learn that you call it “red.” I can learn that you also call ripe tomatoes “red.”

19

I hear you, but why are mathematical wave models “knowledge” ? As in warranted beliefs ? Why does empirical science tell us the “truth” about colorful things ? For context, I am very pro-science. So I’m interested in perception as the basis of science. What sets it apart from “just some pretty math I like.” I’m suspicious of “metaphysical forgettings” of this perceptual basis.

How do you know, if you claim to know, that equivalent firings of cones are correlated with equivalent manifestations of objects for a plurality of observers ? I grant, for the sake of discussion, that equivalent wavelengths and equivalent firings and equivalent verbal categorizations of color are all sufficiently correlated.

Can I infer from these correlations to another correlation ? That between these “public” determinations and “the color of the object for me” ? Perhaps I cannot. And if I cannot, does that mean that object doesn’t manifest to me ? Or just that only my report matters ? Only the firings matter ?

And yet all of these reports are reports for an “average” witness. The world itself evaporates if “quality” does. Or so I am inclined to suggest so far.

Many philosophical explications of science strike me as “Pythogoreanism” in that “only numbers are real.” But this is weirdly idealism ? Because a “pure meaning,” the kind that plays the role of “mind stuff,” is presented as “objective reality.” As if “the meaning of the equations” is “more real” than the perceptions that they summarize and predict. Or, more cautiously, the reports of perceptions and actions explained in terms of such perceptions.

20

Personally I think “the metaphysical subject” is completely “empty.” No entities are “private” in a strong sense. An empirical subject, however, is a thing in the world with “properties.” Jeff’s sadness is not “inside” Jeff. You and I can discuss it. We can even convince Jeff that he is sad.