Some philosophers- e.g. Frank Cameron Jackson, Howard Robinson- contend that the sense-data theory of perception is the only theory that can make proper sense of the term ‘illusion’. Take, for instance, two birds in a field on a sunny day. One bird, a cardinal, is perched atop a tree branch relatively close to our position; the other, a Blue jay, is enjoying a bird bath a little farther away. In our visual context, the cardinal will appear to be larger than the Blue Jay; however, the sense-data theorist argues, the cardinal is not larger than the Blue Jay, and thus what we are directly aware of in perception are not physical objects, but rather, sense-data.
It may be asked, however, that if what I perceive (am directly aware of) are and can only be sense-data, by what measure may I determine perceptual experiences to be in ‘error’ or ‘faulty’? In the above example, I submit that the sense-data theorist is not permitted to judge the cardinal to be “not larger than the Blue Jay”. At best he may assert that, in this particular perceptual context, the cardinal is larger than the Blue Jay. He will also have to permit other, more intuitively problematic, assertions, such as, “The Earth is larger than the Sun.”
In essence, the sense-data theorist must permit all perceptual experiences equal validity in any given, immediate context, and is thus unable to cogently use terms such as “illusion”, “delusion”, and “hallucination”.
It’s relative.
I think that there is something unfair about the epistemological pressure you are placing on the sense data argument.
Unless you are actually arguing that the world is so unstable as to include things like birds that change size depending in where we stand in relation to them, I wonder what other theories of perception could say that would do any better than the sense data theory in establishing grounds for believing in the objective existence of the world…
David, demanding logical consistency is hardly being epistemologically “unfair”. The point I endeavored to make in my post is simply that a sense-data theorist (SDT) cannot make consistent sense of some intuitively important terms. Indeed, the argument from illusion that a SDT gives in support for his position is premised upon a syllogism:
(1) ‘A’ appears larger than ‘B’
(2) But ‘A’ is not larger than ‘B’
(3) Therefore, in (1) we do not directly observe ‘A’ or ‘B’
However, by virtue of his own system, an SDT cannot assert (2). Thus, his argument is invalid.
On another note, a perspicacious SDT may attempt to use statistical terminology, for instance, “Blue Jays are ‘normally’ bigger than cardinals,” in describing sensory experiences. But it seems to me that such a turn is a rather trivial one.
As for your concern for sense-data theory possessing some pragmatic value, I would have to disagree. Sense-data, to me, seems an unnecessary theory that creates more questions than it resolves. If pressed, I would say a form of naive realism is far more pragmatic.
Aaron,
First - great post.
Second, I’m wondering if by, “the sense-data theorist must permit all perceptual experiences equal validity in any given, immediate context” you’re suggesting that an SDT could have no recourse to other perspectives in a given scenario, i.e., in your example, if I am standing directly beneath the tree and have “sense-data” that indicate equal size between the cardinal and blue jay I can not play arbiter in the evaluation of data?
If the SDT is committed to only considering the “sense-data” that she immediately perceives then I think you describe the situation accurately.
I wonder, in a separate thought, if an SDT is barred from including “rules of perception” in her evaluation of “sense data”? In your example for instance, the SDT may interpret the “sense data” with the rule “objects residing closer to me appear larger than objects further away” (or some such). Of course, I suppose that these rules would not make much sense in an SDT, since, as you illustrated, they are based on a syllogism that does not fit the SDT system.
Sorry if this is a bit indirect, hopefully we can continue the discussion nonetheless,
~Q
Quincy,
Thank you for your reply and, yes, we certainly may continue the discussion.
Perhaps I was a bit unclear in my initial post. When I said, ‘the sense-data theorist must permit all perceptual experiences equal validity in any given, immediate context,’ I meant that an SDT possesses no non-trivial way of asserting premise (2) in the syllogism above.
Now, as you pointed out, an SDT may attempt to proffer certain ‘rules of perception’, but any such rule or rules would necessarily be a probabilistic inference. For the following reasons, I would very much like not to found perception upon probabilistic inferences.
(1) When one says, ‘Cordelia is taller than both Regan and Goneril’, one is not asserting an inductive inference, rather, one is asserting a certain, objective state of affairs.
If one intended such an assertion, one would properly use different language, e.g., ‘Cordelia tends- depending on one’s perceptual context- to be taller than both Regan and Goneril,’ or, ‘If observed, Cordelia will likely be taller than Regan and Goneril.’
I do not feel it would be prudent to collapse the meanings of the different assertions upon one another, for the intent and purpose of either is distinctly different from the other.
(2) Any ‘rule of perception’ would be arbitrary and thus trivial. Take, for instance, the rule: ‘objects residing closer to me appear larger than objects further away.’ We may certainly imagine visual contexts in which every object near to ourselves is smaller than every object farther away. This set of visual experiences would provide good reason for us to reject that rule.
Now, if we are to interpret the aforementioned rule probabilistically, we encounter the difficulty mentioned in (1) above and (3) below.
(3) Finally, as you noted, any such ‘rule of perception’ would itself be premised upon the above syllogism and, as we have discovered, that poses certain difficulties of its own.
The question, then, is how may one ‘play arbiter in the evaluation of [sense] data’ in a non-trivial, arbitrary manner?
I agree with your logic. An interesting example lives in all of us with the concept of color. What I call orange could be red to you, etc, depending on how we look at it. I have always wondered if there is a chance that we all see very different things, but we are taught to think of them as the same.
There is a very simple non-trivial way that one may be justified in asserting two. One my say that when all perceptual parameters are identical things that appear larger are in fact larger. There might be an epistemological problem in determining when those parameters are identical, but that is no more problematic than any other skeptical hypothesis.
You seem to have some belief that one cannot draw inferences from sense-data perceptions to objective facts. If that’s true, it is at the very least not trivially so. What (most) sense-data theorists say is that one cannot DIRECTLY perceive anything but sense-data. They have that ‘directly’ in there because they also hold that one can INDIRECTLY perceive objects via sense-data. So, it is a part of their theory that we do have indirect knowledge of objects and we can have such by fixing enough parameters as the same.
So, we can say that some things are larger than others because when all the parameters are identical, one appears larger than the other. We will not have a path to certainty unless we go the route of the phenomenalist, but that is not necessary.
Kevin,
First, I must thank you for your thoughtful response. Now, setting aside the difficulty of determining when perceptual parameters are indeed identical, you say: “One may say that when all perceptual parameters are identical things that appear larger are in fact larger.” However, in order to retain any coherence in that statement, Kevin, one must create some sort of synonymy between the words “appears” and “are” (and its cognate “is”).
Though, I am not persuaded that that can be done while also fulfilling the logical structures of ordinary language. For to say that the “Earth appears larger than the Sun,” and that the “Earth is larger than the Sun,” is to assert two entirely different propositions.
Furthermore, I am not persuaded that an SDT can coherently use the terms “objects” and “objective”, and thus the terms “directly” and “indirectly” in discussing perception.
If, as I firmly believe we do, we gain knowledge through perception, yet we only perceive sense-data, then what possible signification can “objects,” “objective”, or “physical objects” have?
If, then, an SDT cannot coherently use such terms, what possible sense can “one can INDIRECTLY perceive objects via sense-data” have?
I have yet to hear or read an adequate solution from an SDT to this problem. I am persuaded that when an SDT uses such terms he does so to make an untenable theory seem tenable.
Now, let’s reconsider the possibility of there being identical perceptual parameters. I, frankly, think that such perceptual parameters are chimerical. In any perceptual context, there are innumerable variables to consider; more variables than I think we have words to identify.
Now, undoubtedly, there are to a lesser or greater extent similar perceptual contexts. And, equally undoubtedly, an SDT will employ certain rules of perception to make sense of the many perceptual contexts. However that may be, any such rule or rules will be probabilistic inferences and thus epistemologically trivial.
Aaron,
You say:
“Though, I am not persuaded that that can be done while also fulfilling the logical structures of ordinary language. For to say that the “Earth appears larger than the Sun,” and that the “Earth is larger than the Sun,” is to assert two entirely different propositions.”
OF course, those are two different propositions, but that is not the issue. There does not need to be a synonymy between ‘appears’ and ‘are’ for the inference to be drawn. Consider these two statements:
(1) All mammals are warm-blooded.
(2) All dogs are warm-blooded.
You can make the inference from (1) to (2) without there being any synonymy between ‘dogs’ and ‘mammals.’ All you need is this principle:
(1.5) All dogs are mammals.
So, just because a sense-data theorist makes a move from:
(3) In identical perceptual parameters, John appears taller than Mary
To:
(4) John is taller than Mary.
He is not thereby committed to any synonymy between ‘appears’ and ‘is.’ What he (may) be committed to is a principle like (3.5) (although this version leaves out complications regarding the relation between objects and sense-data, which could be fixed but are not necessary for the current discussion):
(3.5) For all objects x and y and perceivers S and perceptual parameters P and Q (If (x appears taller than y to S, S perceives X in P and S perceives y in Q and P=Q) then x is taller than y)
Further, a sense-data theorist MUST NOT hold that we get the inference from (5) to (6):
(5) John appears taller than Mary.
(6) John is taller than Mary.
The whole point of appealing to sense–data was to provide us with an account that allowed for us to make sense of illusions.
You also say:
“Furthermore, I am not persuaded that an SDT can coherently use the terms “objects” and “objective”, and thus the terms “directly” and “indirectly” in discussing perception.”
I am not sure why you say this. There are problems with sense-data theories, to be sure, but I don’t see why they have this one. An object is something that has a particular causal relation to sense-data. What that relation is might be in question but there is at least an intuitive pull that there is such a relation.
Lastly, you say:
“Now, let’s reconsider the possibility of there being identical perceptual parameters. I, frankly, think that such perceptual parameters are chimerical. In any perceptual context, there are innumerable variables to consider; more variables than I think we have words to identify.”
I don’t see the relevance here. There being a large number of parameters does not preclude there being identical parameters. And, it should not matter whether we have words to identify those parameters. The sense-data theorist need not do so, either. All he needs to do is appeal to an idealization, as we often do in philosophy and they regularly do in the sciences.
Kevin,
I admit that my second to last paragraph was a bit unclear. If we consider the extent and magnitude of a single perceptual parameter, that is, the amount of external and internal factors play in determining perception, one would be surprised to find any two parameters to be identical. In fact, I would say the probability of such an occurrence is fairly (to say the least) low.
Recollect the many occasions in which you have had perceptual contact with your lover or best friend. In no two instances were the perceptual parameters identical. At best, Kevin, you may speak of perceptual parameters being more or less alike, or rather, similar.
Even in a more controlled environment, say, a laboratory or a particle accelerator, perceptual parameters are never identical. Indeed, in many cases, and I may speak from experience from my time in nuclear engineering field while in the Navy, we have to give probability factors to represent experimental similarities. For, even atoms (and not only isotopes) of the same chemical species vary from one to another and we can only properly speak of them being more or less similar.
Now, your entire argument relies fairly heavily upon the notion of identical perceptual parameters. If perceptual parameters, however, are not identical, then they, as previously stated, are more or less similar. As such, I am convinced, an SDT can only use probabilistic terminology in referring to perception. E.g., “The Earth tends to appear smaller than the Sun.” Yet, we are not apt to speak in such a manner. Obviously, I place greater epistemological faith in ordinary language than sense-data theory.
Thus, I do not see how you may make the inference from (5) to (6).
This also touches upon my concern over the synonymy between “appears” and “are / is”. I am well aware of your inductive inference from (1) to (2). However, Kevin, you beg the question.
The issue at hand is how may an SDT assert the first premise, “All mammals ‘are’ warm-blooded”? What signification does “are” have in sense-data theory? I see none. Rather, an SDT may only assert, “Mammals thus far examined ‘appear’ to be warm-blooded.” If you cannot transform your “appears” in an “is” via some conditional inference, then you must assert synonymy. You, of course, do not want to assert synonymy.
This leads to another concern. The difference between sense-data and physical objects.
You assert, “An object is something that has a particular causal relation to sense-data. What that relation is might be in question but there is at least an intuitive pull that there is such a relation.”
The phrase, “What that relation is might be in question”, is an understatement. In fact, the epistemological problem presented to such a causal connection between physical objects and sense-data is grave.
Secondly, how are we to begin to determine or make sense of any idea of causal connection between sense-data and physical objects if we are incapable of observing physical objects? You may say that the problem is similar to determining the causal connection between, say, a black hole and the observable mass in its gravitational field; but, that would be imprecise.
For black holes are, in theory, directly observable, whereas physical objects, according to most SDTs, are not.
For if the immediate objects of perception are sense-data which are supposed to represent physical objects other than themselves, it impossible to know if they do in fact represent those physical objects.
I’m not going to go on and on about this, but it seems pretty obvious to me that you are merely appealing to a general skeptical hypothesis. That’s fine, but it’s no problem in particular for sense-data theorists. Someone who thinks we directly perceive objects is going to have problems with skeptical hypotheses and showing that there are any objects that we directly perceive.
Further, you say:
“Thus, I do not see how you may make the inference from (5) to (6).”
But read what I said. I said we SHOULD NOT get the inference from (5) to (6).
You also say:
“This also touches upon my concern over the synonymy between “appears” and “are / is”. I am well aware of your inductive inference from (1) to (2). However, Kevin, you beg the question.
The issue at hand is how may an SDT assert the first premise, “All mammals ‘are’ warm-blooded”? What signification does “are” have in sense-data theory? I see none. Rather, an SDT may only assert, “Mammals thus far examined ‘appear’ to be warm-blooded.” If you cannot transform your “appears” in an “is” via some conditional inference, then you must assert synonymy. You, of course, do not want to assert synonymy. ”
But this entirely misses the point of the example. My point here was that you DO NOT NEED synonymy between two terms in order to draw inference of the form exemplified by (1) and (2). All you need is a principle which you think is pretty plausible. If what you want is a first principle for which there can be no doubt, I think you’ll be very disappointed in the long run; you will not find one.
Kevin,
I think this discussion has run its course as well, but allow me to point out a few things.
1. When I say, “For if the immediate objects of perception are sense-data which are supposed to represent physical objects other than themselves, it impossible to know if they do in fact represent those physical objects,” I am not making an appeal to a general skeptical hypothesis. Rather, I am pointing out the primary epistemological difficulty of the sense-data theory.
2. I am aware that you noted we are not permitted to infer (6) directly from (5), but that instead we must introduce (5.5). Needless to say, Kevin, I do not see how an SDT can reasonably do that without identical perceptual parameters.
3. Again, I understand that one does not need to assert synonymy to infer (2) from (1) and (1.5) if one can establish the premises upon a sure epistemological foundation. However, as I noted in my previous post, the sense-data theorist is unable, as far as I can detect, to do that.
4. Lastly, I do not dismiss sense-data theory completely. I continuously peruse various arguments in favor of SDT. I find some arguments rather persuasive, while others I find gravely deficient.
5. I am not a skeptic and my purpose in the initial post and the replies that followed was not to argue in favor of skepticism. I believe we directly perceive objects, just not mind dependent objects. I shy away from saying that I hold any specific philosophical position, but as of this moment, I find more epistemological work being done in naive realism than I do in sense-data theory.
Again, allow me to thank you for your posts.