Intentionality or the "causal" and "inferential-role" theories of meaning (quote)

You still might wonder how the marks being scribbled and erased by demons inside the computer are supposed to represent or stand for things in the world. Who decides that this mark in the system corresponds to that bit of the world? In the case of a computer, the answer is obvious: we get to decide what the symbols mean, because we built the machine. But who means the meaning of the symbols allegedly inside its'? Philosophers call this the problem of "intentionality" (confusingly, because it has nothing to do with intentions). There are two common answers. One is that a symbol is connected to its referent in the world by our sense organs. Your mother's face reflects light, which stimulates your eye, which triggers a cascade of templates or similar circuits, which inscribe the symbol mother in your mind. The other answer is that the unique pattern of symbol manipulations triggered by the first symbol mirrors the unique pattern of relationships between the referent of the first symbol and the referents of the triggered symbols. Once we agree, for whatever reason, to say that mother means mother, uncle means uncle, and so on, the new interlocking kinship statements generated by the demons turn out to be uncannily true, time and again. The device prints Bella mother-of Me, and sure enough, Bella is my mother. Mother means "mother" because it plays a role in inferences about mothers. 

These are called the "causal" and the "inferential-role" theories, and philosophers hostile to each have had fun thinking up preposterous thought experiments to refute them. Oedipus didn't want to marry his mother, but he did so anyway. Why? Because his mother triggered the symbol Jocasta in him rather than the symbol Mom, and his desire was couched as "If it's Mom, don't marry her." The causal effects of Jocasta, the woman who really was Oedipus' mother, were irrelevant; all that mattered was the inferential role that the symbols Jocasta and Mom played inside Oedipus' head. A lightning bolt hits a dead tree in the middle of a swamp, and by an amazing coincidence the slime coalesces into a molecule-for molecule replica of me at this moment,  memories included. Swampman has never been in contact with my mother, but most people would say that his mother thoughts are about my mother, just as mine are. Again we conclude that causation by something in the world is not necessary for a symbol to be about something; its inferential role is enough. 

But, but, but! Suppose the sequence of information-processing steps in a chess-playing computer turns out, by a remarkable coincidence, to be identical to the battlefield events in the Six-Day War (King's knight = Moshe Dayan, Rook to c7 = Israeli army captures the Golan Heights, and so on). Would the program be "about" the Six-Day War every bit as much as it is "about" the chess game? Suppose that someday we discovered that cats are not animals after all, but lifelike robots controlled from Mars. Any inference rule that computed "If it's a cat, then it must be an animal" would be inoperative. The inferential role of our mental symbol cat would have changed almost beyond recognition. But surely the meaning of cat would be unchanged: you'd still be thinking "cat" when Felix the Robot slunk by. Score two points for the causal theory. 

A third view is summarized by the television ad parody on Saturday Night Live: You're both right—it's a floor wax and a dessert topping. Together the causal and inferential roles of a symbol determine what it represents. (On this view, Swampman's thoughts would be about my mother because he has a future-oriented causal connection with her: he can recognize her when he meets her.) Causal and inferential roles tend to be in sync because natural selection designed both our perceptual systems and our inference modules to work accurately, most of the time, in this world. Not all philosophers agree that causation plus inference plus natural selection are enough to nail down a concept of "meaning" that would work perfectly in all worlds. ("Suppose Swampman has an identical twin on another planet . . .") But if so, one might respond, so much the worse for that concept of meaning. Meaning might make sense only relative to a device that was designed (by engineers or by natural selection) to function in a particular kind of world. In other worlds—Mars, Swampland, the Twilight Zone—all bets are off. Whether or not the causal plus inferential theory is completely philosopher-proof, it takes the mystery out of how a symbol in a mind or a machine can mean something.

  -Steven Pinker, Enlightenment Now, ch 2 pg 81