If it could be made to work, materialism would be attractive simply on grounds of parsimony. Why introduce irreducibly mental items and/or abstracta if one could get by with just material items? By 'get by' I mean explain in adequate fashion all that needs to be explained: consciousness, self-consciousness including self-reference via the first-person singular pronoun, qualia, intentionality, conscience, mystical and religious experience, the applicability of mathematics to the physical world, the normativity of logic, the existence of anything in the first place, the emergence of life . . . .
My main interest is
negative: in showing that materialism doesn't work. Please don't respond by saying that some other theory (substance dualism, say) doesn't work either. For the issue is precisely: Does
materialism work? If theory T1 is explanatorily inadequate, its deficiencies cannot be made good by pointing out that T2 is also inadequate. This is an invalid argument: "Every alternative to materialism is inadequate; therefore we should embrace materialism despite its inadequacies." Wouldn't it be more reasonable under those circumstances to embrace no theory?
One more preliminary point. If materialism is adequate, then we ought to embrace it, and dispense with God, the soul, and the denizens of the Platonic menagerie. For if materialism were adequate, there would be no reason to posit anything beyond the material. But if materialism is not adequate, then we do have reason for such posits.
The following argument is my interpretation of remarks made by Edward Feser in his Philosophy of Mind: A Short Introduction (One World, 2005), pp. 156-159)
1. Consider a representation such as a picture. You draw a picture of your mother. The picture represents her: it is of or about her, and it would remain about her even were she to cease to exist. The picture is a physical object with physical properties: the paper is of a certain size and shape and texture, the ink of a certain chemical composition, the lines have a definite thickness, etc. Now I would insist that these physical features cannot be that in virtue of which the picture represents your mother: they cannot be that in virtue of which the physical item is a representation. But even if I am wrong about this, there remains a problem for a materialist theory of representation.
2. Suppose a 'copycat' comes along and makes an EXACT copy of your picture of your mother. The copycat's intention is not to represent your mother; his intention is merely to represent your representation of your mother. Now there are two pictorial representations, call them R (the original) and R' (the copy). The question arises: Is R' a representation of your mother, or is R' a representation of R? Now suppose a second copycat comes along and produces a second copy R''. Does R'' represent R' or R or your mother? The situation is obviously iterable ad infinitum.
3. Clearly, there is a difference between saying that R' represents your mother, a human being, and saying that R' represents R, a nonhuman drawing of your mother. The reference is different in the two cases. But the reference is indeterminate if we go by the physical properties of the representations alone. Suppose I hand you two drawings of your mother, one an exact copy of the other, but you do not know which is the orignal and which is the copy. You cannot, by inspection of these drawings, tell which is which. Thus you cannot determine the reference from the physical properties.
4. The point is generalizable to other types of representations. Suppose I say 'cat' to refer to a cat and my copycat brother says 'cat' simply to copy me. If my brother mimics me perfectly, then it will be impossible from the physical properties of the two word-sounds to tell which refers to a cat and which does not.
Please do not say that we are both referring to a cat. For my copycat brother is a mere copycat: his intention is merely to reproduce the word-sound I made. To make it even clearer, replace my brother with a parrot who happens to be a perfect mimic. No one will say that the 'cat'-token produced by the parrot refers to a cat. The parrot is just an animate copy machine.
The same goes for any physical representation. Suppose a pattern of neural firing is taken to be a representation of X. An exact copy of that pattern needn't be a representation of X; it could be representation of the original pattern. In general, no material representation of X is such that its physical properties suffice to make it a representation of X as opposed to a representation of a representation of X.
5. Here is the argument:
P1. All thoughts have determinate objects.
P2. No purely material representation has a determinate object.
-----
C. No thought is a purely material representation.
6. Let's consider an objection. "Granted, material representations on their own lack determinate reference, but that can be supplied by bringing in causal relations. Thus what makes a tokening of 'cat' refer to a cat rather than to a word is the fact that there is a causal chain starting with a furry critter and terminating with an utterance of 'cat.'"
But causal connections cannot secure dterminacyof reference, as Hilary Putnam appreciates (Renewing Philosophy, Harvard UP, 1992, p. 23):
One cannot simply say that the word "cat" refers to cats because the word is causally connected to cats, for the word "cat," or rather my way of using the word "cat," is causally connected to many things. It is true that I wouldn't be using "cat" as I do if many other things were different. My present use of the word "cat" has a great many causes, not just one. The use of the word "cat" is causally connected to cats, but it is also causally connected to the behavior of Anglo-Saxon tribes, for example. Just mentioning "causal connection" does not explain how one thing can be a representation of another thing, as Kant was already aware.
In order for Bill's interesting and thought-provoking argument to succeed, we need some reason for thinking the theory of representation I favor is mistaken. Here is one objection to the theory. The word 'cat' represents cats, but does not resemble cats. Therefore, representation cannot consist in resemblance. This objection founders, however, on its conflation of representation and reference. Reference is a conventional, often arbitrary connection between one thing and another. It is the connection a name typically has to its bearer. One characteristic feature of reference is that it does not allow one to infer truths about the referent from a consideration of the thing referring. Representation, by contrast, is never conventional or arbitrary, and typically allows us to infer truths about the thing represented from a consideration of the representing thing. A paradigm would be a road map, which resembles the region it maps. We can always infer truths about the things mapped by a consideration of the (physical) properties of the map itself. So on my theory, there is a big difference between reference and representation, which means that any objection to my theory which presupposes that representation is identical to, or at least a kind of, reference is question begging. (Or at least it's question begging apart from some independent argument for the conclusion that representation is reference.)
To summarize: Bill's argument presupposes the falsity of a view of representation that a number of philosophers (myself included) find plausible, and his argument cannot succeed apart from an argument against that view of representation. But one objection to that view which Bill's post naturally suggests turns out to be question begging. So for Bill's argument to succeed, we need a better objection to the resemblance theory of representation.
Now, perhaps by physical "resemblance" you wish to emphasize resemblance in the subjective sense. While not actually 3-dimensional, the Mona Lisa evokes the illusion of 3-dimensionality, through perspective tricks, for instance. Through this and other factors, it resembles a woman to us - meaning it evokes the concept of a woman in our minds. Likewise, black ink lines evoke the concept of topography to us. They resemble these things by virtue of subjectively evoking concepts of those things in the mind of the observer. But, in that case, you can't disqualify the word "cat", because it also evokes the concept of a cat in the mind of the observer.
This would then bring us to the PoMo literary critic's idea that the valid interpretation of a book or painting is what it means to the observer, rather than what the creator of it was trying to represent. I don't want to go into that discussion, but will just say that once you've agreed that there is such a thing as representation that derives from symbolization instead of physical similarity, the representation that Bill talks about is back in play.
Perhaps you can say that there is a kind of representation that exists in virtue of physical similarity (though it seems to me you ought to just refer to it as "physical similarity" in that case, instead of "representation", to avoid confusion). But if so, it seems that you are simply using the term to refer to something other that what Bill is referring to, and it shouldn't be seen as a denial of the representation that Bill is talking about.
But there is a different approach to understanding representation which focuses on how representations, as information bearing structures, function to regulate and coordinate behavior. Of course, functionalism is neutral between materialism, idealism or hybrids. But once we take the functionalist route, that very neutrality means we must take seriously the possibility of a natural selectionist story of representation.
I hasten to add that even if a natural selectionist story is accepted as an explanation for the origins of primitive representational capacities, more advanced sorts of representation (e.g., symbolic representations) are not plausibly explained simply in terms of natural selection. Rather, I would suggest that natural selection provided us with some very "narrow" representational capacities that we have "bootstrapped," largely independent of natural selective processes, to produce complex symbol systems.
I am not commenting on the argument at this point, but wanted to comment on your set up a little. I agree with your restricting responses to the inadequacy of materialism. It does seems to border on a form of the tu quoque fallacy when someone points out the deficiencies of other views as a response to the inadequacies of materialism. But, don't you provide at least a partial response for why someone would prefer materialism in your intial intoduction? If all the theories suffer from inadequacies, but materialism has this virtue of parsimony, that is often the means by which many choose materialism in comparison to the other theories. If we take parsimony to be a criteria for adequacy, you have already said materialism is attractive. Is attractiveness adequacy? I think not. Is parsimony to be included in judging adequacy? I am not convinced. Is it enough to consider materialism over all the other theories, or, as you put it, to embrace no theory? That's my question. Parsimony sure seems to have pull and gain a lot of traction in the discussion. I would like to see something on the virtue of parsimony in discussions related to this. I'm out to prep my class tonight.
A picture or painting of a woman does resemble the woman quite a bit. Of course there are some features of such paintings that are not woman-like--the texture of the canvas, their two-dimensionality, etc.--but there are many other features which do establish a resemblance. In the case of the Mona Lisa, for instance, there is a region of the painting corresponding to "her" neck. It has roughly the same shape as her neck, and is located between the region corresponding to her head and the region corresponding to her torso, just as her neck is located between her head and torso. I won't go on, but there are many, many other points of resemblance in such a case.
Your use of 'physical resemblance' is odd. The Mona Lisa resembles a certain (imaginary?) woman in virtue of some of its physical properties, such as the sizes, shapes, and arrangements of the regions of color on its surface. How is this not physical resemblance? It is true that black ink lines have a different composition from hills, but if various black ink lines bear relations to one another that correspond to the relations that the hills bear to one another, then the map containing the lines would resemble, at least abstractly, the hilly terrain. Sure, in superficial respects a map of New York resembles a map of Miami. They both can be folded and put in a glove compartment, for example. But how can you deny that a map of New York bears a great deal of resemblance to the city itself? How else would we explain the fact that the map is so useful for getting around the city, if not because of some resemblance?
I hold that complex symbols can represent, but that simple or primitive ones cannot. The complex ones can represent because they can be composed of other symbols that bear relations to one another, relations that correspond to the relations of the thing the complex symbol represents. Simple ones, however, have no relevant structure and therefore are not representations. You seem to commit yourself to something like this view because you characterize representations as "information bearing structures." A primitive symbol like '+' has no (relevant) structure--no structure in virtue of which it bears information--and therefore even on your view could not be a representation. It might be an information evoking sign, but it isn't an information bearing structure.
So what prevents me from saying that purely conventional symbols simply aren't representations, and that to think they are is just sloppiness?
I suspect you have a fairly well-articulated view about the nature of representation, how it relates to information, signs, symbols, and probably a whole lot more. Is there something I could read that would help to orient me? As it is, I think we're using terms somewhat differently.
I didn't mean to suggest that the component abilities that enable us to "bootstrap" weren't themselves products of selection. But that shouldn't be taken to imply that whatever is produced by bootstrapping has itslef been the focus of selective forces. For example, I think natural selection equipped us to make various visual discriminations. We now rely on that equipment to record and retrieve printed information. But I don't believe for a minute that our visual capabilities have been shaped by selection for the ability to discriminate letters.
Thanks for the astute comment. I think there is something correct in what you say. Suppose T1 and T2 are equally explanatorily adequate/inadequate. And suppose T1 posits only one category of entity while T2 posits two or more. Then on grounds of parsimony alone we should prefer T1. But this consideration does not support materialism since there are versions of idealism that are one-category. Berkeley's idealism, for example,is a one category ontology.
Thanks to your comment, I see that my second paragraph is defective.
Of course, there is the vexing and probably unanswerable question of how one compares whole theories to determine their relative adequacy/inadequacy.
You put your finger on a weakness in the argument as stated, a weakness that occurred to me as I was formulating it, but thought I could deal with if some astute commenter such as yourself brought it up. You say it is a false dichotomy to hold that R' either refers to the person depicted or to R since R' could refer to both.
I think you are conceding the point of the Feser-Vallicella argument in a roundabout way. For there is nothing physical about R' to establish that it refers both to R and to the person depicted, or that it refers to the person alone, or that it refers to R alone. That's the whole point: the reference of a material representation is physicalistically indeterminate. But the reference of a mental state is not; ergo, mental states are not material representations.
The fact that R' can be taken to refer to R or to the person or to both shows that by itself R' lacks determinate reference. This is also shown by the fact that if R'' comes into existence, then the reference of R' becomes even more indeterminate.
I just noticed that you distinguish reference and representation. So I'll have to deal with that.
Your theory is that representation is grounded in physical resemblance. Do you mean the following:
X represents Y iff there are sufficiently many physical properties in respect of which X and Y resemble each other.
You write that representation, which you distinguish from reference,
I find this unclear. Consider a topographical map, a 'topo.' Contour lines extremely close together indicate a cliff or some near vertical feature of the landscape, while contour lines very far apart indicate a more or less flat area. Surely there is something conventional about this system of representation: color coding could have been used to indicate elevation change, say bright red indicating very steep, darker reds indicating less steep, etc.
So I fail to see how contour lines physically resemble anything in the actual landscape. The symbol for a shaft, tunnel, or cave is a 'Y' on its side, while the symbol for a mine is an 'X.' Surely those are purely conventional signs: a cave needn't look at all like a 'Y' on its side. So it seems that the last sentence of the above quotation is false.
If your theory has the above bioconditional form, then it implies that any two things that physically resemble each other, represent each other. Surely that is false? Consider two qualitatively identical paper clips. Surely they do not represent each other in any interesting sense. But perhaps I do not understand your theory.
We can adopt the communication theorist's view of information as an objective measure of dependency. An item can then be a bearer of information without being a representation.
In order to function as a representation, the information bearing item must play a particular sort of role in regulating or coordinating the behavior of an information user/consumer. Organisms are the most prominent such users/consumers in our part of the universe.
When we characterize an information bearing item in terms of its function as a representation, the representational content will typically not exhaust the whole of the informational content of the item. What part of the informational content is treated as represental content depends on the functional role of the item in the operation of the user/consumer in question.
I note, of course, that this view of information and representation does not implicate consciousness.
I'm having a hard time understanding you. A map by itself doesn't represent anything since it is merely a piece of paper with marks on it. It represents a chunk of landscape only if someone interprets it as representing. It possesses no intrinsic intentionality. The same goes for states of the brain.
I don't think you appreciate that your own theory of representation, according to which it rests on physical resemblance, entails the indeterminacy of the physical with respect to representation. Suppose a Madame Tussaud wax statue S of you is made that is a very close likeness. It represents you on your theory. But then an indiscernible copy C of S is made. There is nothing physical about C that makes it represent you rather than S. The physical properties do not determine the reference. If you say it refers to both, then I say that it is indeterminate as between referring to both or referring to one or the other.
But that is not the way mental states are: a thought about X could not have anything other than X as its object. A thought about Boston is not indeterminate as between being about Boston and being about an indiscernible mock-up of Boston.
I can't see that you have refuted the argument.
Your proposal is a bit sketchy, too sketch for me to wrap my mind around it. How close are you to Stalnaker? I don't mean that spatially, of course.
I agree with Bill that your proposal remains too sketchy. Perhaps you could begin to flesh it out by answering at least in general terms this question:
What sort of role does, say, a photograph of the Eiffel Tower play "in regulating or coordinating the behavior of an information user/consumer"?
Thanks for the abundance of helpful and intelligent comments. I will try to clarify my view.
1. The bold biconditional you ascribe to me above accords well with my view.
2. You are absolutely right that there is something conventional about maps, e.g., the choice to use distances between lines to represent changes in altitude. Yet, this choice having been made, it still follows that the lines bear relations to one another that correspond to the relations that the bits of land bear to one another. And this connection or correspondence between the lines and the land--which is what the representation here consists in--is not at all conventional. I maintain that the topo represents in virtue of the fact that it exhibits a certain relation-structure, the parts of the contour lines bearing various spatial relations to one another. This particular structure could have been instantiated in arbitrarily many ways (using colors, contour lines, etc.), and it is a conventional matter how, in designing a map, we choose to instantiate this relation-structure. But representation itself is not at all conventional because it wasn't up to us that that particular relation-structure would correspond to the relation-structure exhibited by the mapped terrain. That is what I had in mind when I said that representation is never conventional or arbitrary.
3. I do not hold that the individual contour lines physically resemble anything in actual landscape, but that the contour lines, considered collectively, physically resemble features of the actual landscape. The resemblance I have in mind here is abstract, structural resemblance. When the lines are closer, the land is steeper, and when the lines are farther apart, the land is flatter. The relation-structure consisting in the various distance relations (which are physical) that the parts of the contour lines bear to one another corresponds to or is (approximately) identical to the relation-structure consisting in the "altitude relations" that the bits of land bear to one another (this bit of land is 100ft. higher than that bit of land, etc.). That is the sense in which a topo resembles the landscape it represents.
4. You are entirely correct in saying that my view "implies that any two things that physically resemble each other, represent each other." I embrace this conclusion, though I would add that most of these cases of resemblance are rather uninteresting, which is why they typically go unnoticed as representations. Two qualitatively identical paper clips do represent one another. I totally embrace this claim. But as you correctly point out, this is not an interesting case of representation. It isn't interesting because we don't find it useful to use one of the paper clips as a surrogate for the other--as a way of discovering or marking truths about the other. By contrast, a map of NYC is an interesting representation because we do find it useful for discovering or remembering truths about NYC itself.
5. As to the wax statue, you want to say that C is "indeterminate as between referring to both [S and me] or referring to one or the other." But what is the argument for this claim? Why isn't it correct to say, as I want to, that C determinately represents both S and me? Unless this can be shown to be incorrect, your indeterminacy of the physical argument can't get going.
Also, if representation is just physical similarity, then isn't it redundant to speak of "representation"? Why not just say that there is physical similarity and nothing more? What is captured by "representation" that is not captured by physical similarity, since you've tried to remove all mental aspects from it? This seems like one of those reductions that is really just an elimination, like equating propositions with brain states.
Also, by representation, Bill means (if I am not mistaken) that a particular thing was intended to represent something else, and from there he's arguing that you can't derive from the physical properties of a thing what it was intended to represent. If you are saying that this representation is just the physical similarity between two objects, it would appear that you're being an outright eliminativist about the intentionality.
Finally, what about when we use pictures to represent logical constructs, such as two overlapping circles with shading to represent the XOR function, for instance? Are these not representations?
First, it is crucial to this approach that informational relations are taken to be objective. Information exists wherever relations of causal dependency exist. For example, the movements of planets in our solar system literally carry information about gravitational forces in our region of the universe. Likewise, the optical array at a given point literally carries information about the distribution of reflective surfaces in the surrounding environment.
(I note that the resemblance between two items must reflect a real dependency (i.e., it’s not coincidental) if it is to constitute an informational relation between the two -- unless we’re dealing with the special, and quite complicated case where a cognitive agent opportunistically takes advantage of a coincidental resemblance.)
But the fact that things carry information doesn’t automatically make them representations -- at most, it makes them potential representations. This is because ‘representation’ is a functional concept -- not in the purely formal sense of ‘function’ employed in mathematics or the physical sciences, but in the more restrictive sense common in the life sciences and the so-called artificial sciences. In those domains, to say that x is a function of y has a “teleological flavor.” In the artificial sciences, the teleology is based on the purposes/designs of intentional agents. In the life sciences it is commonly thought to be based on the operation of natural selection. Reference to purposes and/or selection histories serve to narrow our focus onto a subset of the effects produced by the thing to which we attribute a function. The important point, though, is that for something to be a representation it must not only carry information, but it must have the function of regulating/coordinating behavior in virtue of carrying (at least some of) the information it does carry.
(A photograph of the Eiffel Tower might regulate/coordinate verbal responses to questions about the appearance of the tower; it might regulate/coordinate the movements of tourists in the city of lights (“No, that’s not it. Let’s keep looking.”); it might regulate/coordinate the occurence of memories of that springtime visit...)
I think there are some similarities between my view and that of Stalnaker, particularly regarding the notion of information as an objective measure of dependency. But I don’t think Stalnaker uses ‘representation’ in quite the way I do, and he certainly doesn’t use ‘function’ in the way I do. Also, Stalnaker seems to be concerned primarily with understanding very highly developed cognitive systems capable of subjecting information to quite complex processing, employing conventionally agreed upon symbol systems, etc. I’m inclined to think that we should learn to crawl before we walk, walk before we run, run before we attempt hurdles, etc.
Thanks for the clarifications. I still have questions, though. Take the example of the pic of the Eiffel Tower, which on your view represents the Tower in virtue of the fact that it regulates/coordinates the movements of certain tourists who are trying to find the Tower. (Let's use that example rather than the ones about verbal responses or memories.) One big question is how to generalize this account of representation so that all and only genuine representations are captured. I don't see how you can do it. Suppose we are talking about a postcard of the Eiffel Tower in a Paris gift shop. A person has picked it up off the rack and when putting it back uses the picture of the Tower on the postcard to identify where it should go in the rack. Would you say that this pic of the Tower actually represents that part of the postcard rack because the pic regulates/coordinates the movements of certain tourists who are trying to find out which part of that rack the card should be put on?
A few reactions:
1. The child's squiggles do not represent a horse. You say that to the child "it's a representation of a horse." I would say rather that for the child those squiggles indicate a horse, or evoke thoughts of a horse, etc. But I consider it a mistake to say that they represent a horse to the child. The very idea of a thing representing one thing to one person and another thing (or nothing) to another person makes no sense.
2. Some representations involve physical similarity, but it doesn't follow that representation is merely physical similarity. Perhaps non-physical things (thoughts? mental images?) can bear relations to one another, in which case relation-structures could be instantiated by non-physical things. For me the relevant similarity is similarity of relation-structure. In some cases this amounts to nothing more than physical similarity, when the relation-structure is instantiated by physical things, but not always.
3. I believe that Venn diagrams can be used to represent things when the relations that the shaded and unshaded areas bear to one another correspond to the relations that the elements, constituents, etc. of the thing represented bear to one another. But I'll have to think about that one some more.
I think that in the case you describe we need to factor in the beleifs, desires and intentions of the person interacting with the postcard (itself an artifact), which play an important role in determining how the information carried by the postcard is processed (this might not be "hurdling," but it's not "crawling" either). And remember, the postcard carries information in excess of what it can convey about the Eiffel Tower -- in particular, it also carries information that can be used to compare it to other postcards on the rack, regulating where it will be placed, given the beliefs, desires and intentions of the person in question.
But all of this might be straining at gnats if we bear in mind the distinction between information and representation. It might be most plausible to say that when the person is endeavoring to return the postcard to its proper place, the postcard is not functioning as a representation, but is simply an informational array the person desires to position among other informational arrays. I'd need to think long and hard about this to commit to any one way of viewing it as the "right" way.
Your theory of representation is quite fascinating. But I'm still having trouble with it. You write,
This implies that a relation-structure (RS) is a universal, where a universal is any repeatable (multiply exemplifiable, multiply instantiable) entity. So there is an RS instantiated by this particular map of the Superstition Wilderness in my hands and a second RS instantiated by the wilderness itself. Representation is then a relation that connects RS-1 and RS-2, and the obtaining of this relation is not up to us, and is insofar forth not conventional.
Sounds rather Platonic! Seems you are shifting representation from the mental plane to the plane of abstracta: representation is a relation between abstracta. Could that be what you mean? Of course, if that's your view, then your view is inconsistent with a full-bore materialism acc. to which everything is mat'l.
Not that you would deny that.
But here is the problem. If representation is a relation that connects universals (relation-structures), then it is not clear how you can speak of physical resemblance. The relation-structure that is common to, and instantiated by, contour-lines and contour-colors (and indefinitely many other schemes of indicating elevation change) has no physical attributes. How then can it physically resemble anything?
In short, how can you speak of representation as both abstract and physical?
The thing is, I think there is a kernal of truth in what you are saying. All cases of representation that I can think of are such that there exists some sort of relation or homomorphism from the representation or its context to the represented or its context. But the homomorphism isn't the cause of the representation; it is the purpose of the representation. In other words, it seems that you are confusing the final cause with the efficient cause.
In the case of a blind person, I would resist saying that the painting doesn't represent the subject to her (the blind person). I would say rather that the painting represents its subject, and that she just isn't in a position to notice this. On my view, representation consists in structural correspondence, and it makes no sense to say that the painting corresponds-to-the-subject for one person but not for another.
I'm glad to hear that you think there's at least a kernel of truth in my view. But why would you say that the homomorphism is the purpose of the representation? I would think that the purpose of a representing thing such as, say, a map of Paris is to allow us to find our way around that city. And this purpose is achieved by making a map that is homomorphic to the city. The homomorphism isn't the purpose; it's what allows the map to fulfill its purpose.
Thanks for plumping, even if only tentatively. But isn't it a somewhat bizarre theory of representation that commits us to saying that that postcard of the Eiffel Tower represents a position on the postcard rack? (A more serious problem for the sort of view you favor is how to account for error or misrepresentation. Robert Cummins, in his book Representations, Targets, and Attitudes pursues this difficulty in great detail. In Chapter 7 ("Representation and Isomorphism"), incidentally, he defends a view similar to mine.)
Excellent and penetrating comments. I think what I want to say is that representation is a relation between two things (concrete or abstract) that have (or instantiate) the same (or very similar) relation structures. The map, for example, represents the terrain, and does so in virtue of its having a certain structure that corresponds to the structure of the terrain. So it's not that the structure of the map represents the structure of the terrain, but that the map represents the terrain because these structures correspond.
I think for purposes of this discussion I can remain neutral on the ontological status of these structures I'm making use of. Perhaps they are universals. In that case, I suppose my view would be that one thing represents another when they instantiate the same relation-structure. But perhaps these structures aren't universals. Perhaps they're like the tropes or particular properties favored by G.F. Stout, Donald Williams, and Keith Campbell.
This may not yet be sufficiently clear or precise....
I'll grant tha it does sound a bit odd to say simply that a postcard of the Eiffel Tower represents a position on the postcard rack, without providing any contextual information. But I think this is because we usually give priority to the intentions of the original designer when identifying things in terms of their functions. Still, entities can have multiple functions, some of which might not reflect the intentions of the original designer -- as when a coffee mug is pressed into service as a pencil holder or a paperweight.
Also, some here may be interested in the Leibnizian roots of the view I favor. (Yes, Leibniz again.) If so, I offer the following, a large chunk of Leibniz's essay "What is an Idea?" written in 1678:
One can see here some of the key elements of the view I have been promoting: first, that representation involves a correspondence or analogy between the relations of the representing thing and those in the thing represented; second, that because representation has this nature, one can infer truths about the thing represented from a consideration of the representing thing itself. I discuss Leibniz's view at length in my dissertation, Perception and Representation in Leibniz.
Consider a map M of an area A. There is a marina in area A that corresponds to the point p on M. Consider three case:
(1) There is a speck on the map at point p caused by a fly or a printer's error. This speck does not represent the marina even though it bears the relation to the map that you require.
(2) The map maker thought that the marina was located at a position corresponding to point q on the map and so he put a dot at q, different from p. There is only one marina in A, so the dot at q represents the marina, but it does not have the relationship that you say that representation consists in. A boater reading and understanding the map would end up at the wrong location because he correctly understood what the dot at q represented.
(3) There is a dot at point p where the map maker put it to represent the marina. In this case, your theory would say that the marina symbol represents the marina because of where it occurs on the map, but as I showed with (1), that isn't the case. Rather, as I showed with (2), the dot at p represents the marina because the map maker intended for it to represent the marina. And because it represents the marina, the map maker put it at p. In other words, the dot doesn't represent the marina because it is at p, rather, it is at p because it represents the marina.
Something similar holds for all representations that I can think of. They have a structure that is homomorphism to the thing represented. But they don't represent because of that structure, rather they were given that structure because they were intended to represent.
You seem think my view is something like this:
X represents Y in virtue of the fact that X bears certain relations to other things that correspond to the relations Y bears to other things.
That's not it, though. My view is this:
X represents Y in virtue of the fact that the constituents (elements, parts, etc.) of X bear certain relations to one another that correspond to the relations the constituents (elements, parts, etc.) of Y bear to one another.
So I don't say that the dot on the map represents anything; it doesn't have the requisite structure. I say only that the map itself (and complex parts of the map) represent. I would say that the dot you mention in cases (2) and (3) indicates or designates or refers to the marina. But designation, indication, and reference are different from representation. In case (1), the speck neither designates nor refers.
You say that representations "have a structure that is homomorphism to the thing represented," but this is a highly misleading way of putting your view. You think things can represent even when they don't have a structure, as long as they belong to a structure. You should say rather that representations belong to structures that are homomorphic to the structures that the things they represent belong to. There's a big difference between having a structure and belonging to a structure. On my view representing has to do with the former, not the latter.
Thanks. I was about to ask you if there were any contemporary philosophers you were basing yourself on. I'll have to read some Cummins.
Let me make a quick comment about Leibniz. The equation y = x squared has a familiar graphical representation. Here the equation represents the graph and vice versa. Representation in this case is symmetrical. But mental representation is asymmetrical: my thought of you represents (is about) you, but you are not about my thought, you do not represent my thought. Of course, you could entertain a thought representing my thoughts, but you qua object of my thought of you do not represent my thought of you.
Of course you realize that in denying that a simble object can represent you are abusing normal language. I like to play Risk with the old-style board where armys are represented by colored chips of wood. There is no resemblence between a colored chip of wood and an army, but you ask any risk player what those chips "represent" and they will promptly reply "1, 5, or 10 armys".
This doesn't completely derail your notion of representation, of course, but it does call into serious question whether you are talking about the same thing as everyone else when you say that "X represents Y".
But I believe that even your restricted notion of representation has to answer my challenge on the direction of causality. Take the map as a whole representing area A as a whole, with a dot a p to represent the marina, and a dot at q to represent a town and a line drawn between them to represent a road. The map doesn't represent A becaues it has the dots at p and q and the line between them, rather the map has the dots at p and q and the line between them because it represents A.
I admit that my view does not preserve all our uses of 'represent' and its cognates. But I doubt any coherent account of representation could do so. To think otherwise is to suppose what is almost surely false: that our ordinary use of 'represent' and its cognates is coherent. So this is a price I'm willing to pay, because I suspect it's a price that any coherent theory must pay. (Of course, if there were two coherent theories and one did more justice than the other to our ordinary uses of the term, we should favor the former over the latter.)
I continue to think your point about causality a bizarre one. A cartographer might say, "I want this to be a map of NYC, so there ought to be a dot at p and a dot at q, etc." But this doesn't mean that there are dots at p and q because the map is a representation of NYC; it means only that the cartographer put dots at p and q because he wanted to achieve a representation of NYC. What makes the map a representation of NYC--what "causes" it to be a representation, if you will--is that the dots are located in such a way as to achieve isomorphism. Had the cartographer wanted to make a map of NYC but had inadvertently put the dots in the wrong place, I think most would agree that it would fail to be a map of NYC. It would fail because it isn't homomorphic to the city itself. In fact, it might even turn out by dumb luck to be homomorphic to some other city, say Phoenix, in which case the right thing to say is that the cartographer meant to make a representation (map) of NYC, but actually made a representation (map) of Phoenix. But you seem to want to say that in such a case the map really is a representation of New York, even though none of the dots and lines are in the right place. That seems seriously wrong to me.
2. Disallowing comments from a particular person, or deleting an offensive, off-topic, or otherwise substandard comment, has nothing to do with censorship. People who think otherwise confuse censorship with lack of sponsorship. I am under an obligation not to interfere with anyone's exercise of legitimate free speech rights. But I am not under any obligation to aid and abet anyone's exercise of free speech rights, legitimate or illegitimate.
3. The Comments area is not an open forum for anyone to say anything about any topic. As the name implies, it is primarily for commenting on the author(s)' posts. But to comment on them, one must have read them. And if I have spent three hours on a post, a reader will not understand it in thirty seconds. Secondarily, the Comments area is to facilitate civil discussion between and among commenters as long as the discussion remains on-topic.
4. Some undesirables: The skimmers, those who cannot read but only read-in. The sophists who, abusing argument, argue for the sake of argument. The ideologues, those who are out for power, not truth. The uncivil. The illogical. The politically correct. Worst of all, perhaps, are those who exemplify the anti-Socratic property: those who think they know what they don't know. If Socrates was famous for his learned ignorance, these types are marked by their ignorant unlearnededness.