Wednesday, September 17, 2014

If a Lion Could Talk

Ludwig Wittgenstein made many statements that became famous in philosophical circles. At the conclusion of the Tractatus Logico-Philosophicus he wrote "Whereof one cannot speak, thereof one must be silent." This is the odd ending to the quasi-logical tract that he wrote while he was a soldier and prisoner during World War I, which later became his Ph.D. thesis at Cambridge. Towards the end of his career he wrote, "If a lion could talk, we could not understand him." This was published posthumously in Philosophical Investigations.

The "If a lion could talk" statement occurred in the context of Wittgenstein's thoughts about language games. Although it is an ambiguous sentence, it is generally thought to mean that the frame of reference of a lion is so different from that of a human that lion language would be unintelligible to us. The theory is that meaning is related to the use of words, and that lions, if they could speak, would have entirely different uses for language than we do.  I have tended to agree with this idea but was thinking about it again recently and am no longer sure that I do. Language involves more than words and grammar: it contains symbols. Most other animals have little or no capacity to understand symbols, but some do. I found it instructive that no animals besides humans, including all other primates, understand the meaning of pointing, with the exception of dogs. Wolves don't either. Among humans, pointing is a symbolic shorthand used to indicate that an object worthy of attention is located in a specific direction. In the case of people it might be the location of a gas station. In the case of dogs, it might be the location of prey. For all intents and purposes, a pointing dog might be saying "There are quail in these bushes."

What is different about the case of dogs is that, through habituation and breeding, they have interacted with humans enough to understand some basic symbolic communication, which I consider a form of language. Thus, in theory, if a dog could speak in sentences, we might understand some of what it says exactly. I'm sure many dog owners would agree. Some cat owners would probably be able to understand at least part of what their cats were saying if they could speak: "I'm hungry" or "I want to play," for example. Mutatis mutandis, the same might be true for lions, making Wittgenstein technically incorrect.

This brings me to Noam Chomsky, whom I recently saw speaking in a video about artificial intelligence. Apparently he thinks it isn't going anywhere any time soon. He believes that humans have unique genetics that support language, and that we are nowhere near figuring out how it works. Since language is central to our thinking abilities, Chomsky is also skeptical about our ability to create meaningful artificial intelligence that might actually rival our own intelligence. According to him, there may never be a singularity. I am not up to date on research in the fields of language simulation by machines and artificial intelligence, but I suspect that Chomsky will be proven wrong. For example, if computers can be taught to reconstruct languages based on their exposure to grammar and word usage, it seems to me that they might be able to use language just as well as humans. One of the advantages of computer learning is that if a process works at all, it can be sped up and developed rapidly simply because of the processing power and memory capacity of a large computer. While an infant might pick up grammar and usage gradually over a period of years through real-time experiences, a computer might work its way through an enormous database in a short amount of time, permitting more rapid language acquisition than humans. Once computers are able to perform a broad range of learning activities comparable to humans, it seems probable to me that they will quickly surpass humans in cognitive functions.

Whether that will happen or not remains to be seen. At the moment I am interested in the reasons why Chomsky, more so than Wittgenstein, thinks that human language is unique. It seems to be because we are the only animal that we know of that has true language. Certainly language has given us an enormous advantage, but Chomsky seems to project a sort of hubris about it. If you look into Chomsky, he is quite an odd fellow. By contemporary standards he is an off-the-charts radical who advocates anarchism and the end of the nation state. The reality is that he is a dyed-in-the-wool Enlightenment thinker who still thinks that man is the center of the universe. I think this is a better explanation for his minimization of the potential value of research in artificial intelligence than any particular knowledge that he may have of the subject. What I have noticed over the years is that even when an intellectual legitimately reaches prominence through hard work and important discoveries, once he is given a pulpit from which to speak, his ideas for the public often bear no relationship to his research and in fact may be ideas that he arbitrarily absorbed as a teenager. In Chomsky's case, he was strongly influenced by a radical uncle when he was growing up. Most of his political ideas originated then and have nothing to do with his research. Thus, as I suspect is true in the case of many public intellectuals, there is no direct link between their areas of expertise and the public opinions for which they are known.

No comments:

Post a Comment