Silliman's Papers

Do I have Understanding?
Searching for an understanding beyond the functional

Sitting in Philosophy of Mind class, I listen to my professor explain John Searle’s anti-Artificial Intelligence piece, Minds, Brains and Programs. Dr. Jim Stephens speaks to us in his class about the background of the debate, the nature of the Turing Test, Searle’s example and the way the Chinese Room works. The lecture goes well, and with a lot of student participation and class discussion we move through the material. It seems we understand Searle fairly well—the class soon moves past the workings of Searle’s arguments to the implications. Contemplating the article, which I have heard described and explained a number of times and which I have now carefully read, I pause and ask myself if my professor understands Searle.

There is nothing about what Dr. Stephens has said that causes me to think he doesn’t know what he’s talking about, but this is a matter of syntax and not of semantics. This lecture is the manipulation of formal symbols, the syntactical system called of philosophy. Philosophy is surely a language—it is a system one learns to manipulate. Let us call the system of philosophy, philosophy-speak, philanguage. When Dr. Stephens speaks about “functions,” “understanding,” “Searle,” “Turing,” or any of it, he is manipulation the system of philanguage that was programmed into him during the years he spent internalizing it in an education. But does Dr. Stephens understand the work of Searle? Does he understand any of the things he’s telling us in this class on the philosophy of mind? He can speak philanguage very fluently but does he know that “syntax” means syntax or means anything at all when he writes it on the board? Perhaps he has just been well programmed to manipulate philanguage. He has been well programmed at the best schools and has the ability to manipulate the syntax in such a way as to fool his professors then and his students now.

He can speak the language, but perhaps he cannot understand any of it. He can answer all our questions and can tell an educated and sophisticated story about the material, but this only shows that he has internalized the complex system of syntax of philanguage, memorized the rules and methods in the philanguage books, does all the calculations in his head and has incorporated the system. There may be nothing in the whole system of philanguage that Dr. Stephens does not encompass and he manipulates philanguage so that it seems he knows the terms “Searle” and “Chinese Room” actually mean something more than a series of guttural noises that work in a systemic way with other guttural noises. But none of this extensive and impressive usage of philanguage shows anything of understanding.

There is no functional way to tell what he understands, because understanding is not a functional thing, as Searle shows us in the thought-experiment of the Chinese Room. Understanding is a personal thing; one can only identify understanding in the first person. I will never know if Dr. Stephens is more than a well-programmed machine. This is a bit disappointing, because it is hard to respect a professor who is probably only manipulating a complex system he doesn’t understand as meaning anything. It would be nice to know my professor gets the meaning of the subject he teaches, though it would not make any difference in class.

Since it cannot be known to me that my professor Dr. Stephens, my fellow students, or anyone understands anything, my thoughts continue to drift in class. Considering that knowability of one’s understanding cannot exist in the functionalist third-person, I turn to the first-person. Since understanding is not functional, its knowability is limited from the third-person. So I turn to inspect my own division between syntax and semantics, my own line between system manipulation and the grasping of meaning, my own human understand that is beyond that of Artificial Intelligence. I wonder if I understand Searle and the Chinese Room.

Since it is not functional, I assume my understanding will have nothing to do with passing this class and less to do with grading. Any grade will necessarily be based on system manipulation. My professor cannot tell if I have understanding anymore than I can tell if he has it. I have told him I understand, but the statement, “Ahhhh. Now I understand” is something of syntax and is a functional thing, not a semantic thing of understanding. Perhaps I am fooling him. Any judgment on my understanding will be third person and based on functional action that cannot denote meaning. Thus the class, and any relation between me and the world, will be based on my manipulation of the system of the syntax of philanguage. The question of my understanding can only be answered by me by introspection, and will have no bearing on that outside of me.

Looking inward, I wonder if I could fool myself. How good is first person knowability of understanding? Occasionally in philosophy when reading something particularly full of jargon and self-referential definitions of uniquely used terms, one is caught in a jargon loop. When thinking of one term, A, I uses the definition B. When thinking of term B, I use the definition A. Caught in a circle of jargonizing, if I stop using one definition I have no way to think of the whole thing because I didn’t understand. Is it possible, then, that I could be fooling myself? Is it possible that I am manipulating systems and playing with syntax, but have no understand? How would I know? If I am but a highly programmed machine, passing all sorts of daily Turing Tests, how would I stop from being fooled by my own manipulations of the programs? I think I ought to look for a system I know I know, not something new and full of jargon like a philosophy class, but something fundamental to my knowledge and something I’ve known I’ve known for a long time.

Looking for a system I understand, clearly, in more than a syntactical way, I look to English. I don’t remember learning English. I barely remember learning to read the language around the age of 4, and I thoroughly knew the language before that. When learning other systems I didn’t understand—Greek, Latin, philanguage—I translated them into English so that I could understand. Surely if there is a system I know in more the semantics of, it is English. It seems obvious that I know English. If I know that “dog” means dog, and is more than a piece of a syntactical system, it is in the system of English. But separating “know” from “use well” is a tricky business. “Knowing” English cannot mean that I have spoken it for a long time or that I write the language well. If I am to be separated from the syntax manipulating and the program running computers, there must be a “knowing” of English that is more than functional. There is nothing in my experience that shows English is anything more than the system I can best manipulate. Looking inward, I cannot find any understand of the system that is beyond the system, outside of it.

Turning to Searle’s descriptions of knowing as distinct from using, I find that knowing is supposed to be an existential thing. What is the difference, Searle asks, between knowing that “squiggle, squiggle” is followed by “squoggle, squoggle” and knowing that “Turing Test” is followed by “Artificial Intelligence”? Why do I look at the one and think, “Okay, I’ll remember this relationship” and at the second one and say “Yes. I understand.” What is this understanding? There is an experiential difference, one that is called “understanding what a thing means.” But what is it to understand “hamburgers” as hamburgers in way totally beyond the functional? If I were asked if I knew what hamburgers were, I would certainly affirm that I did, and proceed to talk of ground beef and buns with mustard and pickles. But this is functional, this is the same thing a computer would be programmed to say when asked about the meaning of hamburgers.** This is the response of a Chinese speaker who doesn’t know English but has internalized the system enough to know that the marks “hamburger” are followed with the marks “ground beef on bun with pickles and mustard.” This is not understanding, as Searle wants to speak of. Do I know what “hamburger” means? Do I know what it means in a way that I know it without the marks “hamburger” and needing an inter-system reference like the marks “picnic,” or “fast food”? It seems this understanding is an existential thing, it is the experience of a moment of “yes, I understand” without the statement “yes, I understand.”
It seems that understanding is an existential thing, but every time I attempt to climb out of the syntax and into semantics, I find more syntax. Understand is something beyond language, but I don’t seem to be able to pass beyond language. The difference between my comprehension of “squiggle, squoggle” and “What psychological and philosophical significance should we attach to recent efforts at computer stimulation of human cognitive capacities?” is functional. The difference is between my ability to talk extensively about the one and not the other. The difference is that one is part of a large syntactical system I can manipulate and the other is a part of a system of two words. This makes the one interesting and thus meaningful, while the other is gibberish because we can’t talk about it.

Normally, it seems my understanding is a linguistic thing. I “understand,” which means I can use the syntax is a sufficiently competent manipulation. I pass a formal or informal test, a test of mine or someone else’s, at my skill manipulating the system and then I say I understand. Searle is asking for an understand that is beyond the functional and is beyond language. He speaks of understand** as “the possession of mental (intentional) states.” Asking, “What is understand?” the text seems to answer with a cycle of phrases: understanding, “understanding,” knowing what a thing means, what it means, intentionality. If I ask what “means” means, I’m answered by replacing the original statement with one using another term. One can sympathize with the spot Searle is in, trying to speak of that which can only be known in the first person and that which is beyond the functionalism of language. This is little comfort when I ask if I am more than a well-programmed computer fooling the world. Do I have intentionality? Do I understand anything? Do I know what things mean? These are questions that, seemingly, can be answered with rather simple introspection. One is not supposed to have to become a mystic meditating on the unspeakable mysteries to know it one has understanding. This is supposed to be obvious.

In the end I can assert to myself that I have intentionality, but even to myself I cannot demonstrate it. If I ask myself the question of my understanding, I can give no answer that does not entail the functional. Looking inward, I only find semantics in syntax. If understanding is, as Searle says it is, more than this, something so obvious as to be beyond description, then I can only conclude I don’t have it. Perhaps I am a well-programmed computer. Perhaps I am Artificial Intelligence.*** There is no internal or external evidence to the contrary.

Notes:
*Consider a dictionary program.

**Though he only speaks of it in the footnotes. He doesn’t define the all-pivotal term in the main text of his paper.

***It should be noted this same line is taken in the film Bladerunner (derived from Philip K. Dick's science fiction novel Do Androids Dream of Electric Sheep?). Rick Deckard’s job is to separate the Artificial Intelligence (Androids) from the human, and finds this distinction extremely difficult when attempting to apply it to himself.