[H]e is proclaiming his new project, the Wolfram Language, to be the biggest computer language of all time. It has been in the works for more than 20 years, and, while in development, formed the underlying basis of Wolfram’s popular Mathematica software. In the words of Wolfram, now 54, his new language “knows about the world” and makes the world computable.
From the point of view of the philosophical debates on artificial intelligence, the crucial bit is the claim that his new language, unlike all other computer languages, “knows about the world”. Could it be that this language does indeed constitute a convincing reply to Searle’s Chinese Room argument?
To be clear, I take Searle’s argument to be problematic in a number of ways (some of which very aptly discussed in M. Boden’s classic paper), but the challenge posed by the Chinese Room seems to me to still stand; it still is one of the main questions in the philosophy of artificial intelligence. So if Wolfram’s new language does indeed differ from the other computer languages thus far developed specifically in this respect, it may offer us reasons to revisit the whole debate (which for now seems to have reached a stalemate).
But over at Slate, David Auerbach is not convinced:
In short, the users of the language “know about” the world, not the language itself. Baking particular sorts of data into the language does not help it “learn” or “understand.” It is not able to generalize or deal with exceptions to its rules. Wolfram has shown how the language can grab the flags of different countries and treat them as a dataset that can be manipulated for geographical or aesthetic analysis—but that’s only because the code for Wolfram has specific handling for associating specific visual images of flags with specific country names. But if the nation of Davidstan decides to change its flag to an animated, rotating sphere with ultraviolet paint on it, the Wolfram Language won’t be able to handle that without modification.
At first sight, this may seem to be yet another version of a Chinese Room-like argument. Yet, the real question seems to be not whether the Wolfram Language itself knows about the world, but whether the agents using it – computers – will ‘know about the world’. Will it be a phenomenon sufficiently similar to how competent speakers of English ‘know about the world’ when using English – or any other vernacular language they master, unlike the man in the Chinese Room who does not speak Chinese? (Putting aside the fact that Searle’s characterization of computation as manipulation of meaningless symbols is somewhat simplistic.) To claim that a language, not the users of a language, ‘knows about the world’ is arguably a category mistake, and Auerbach is right to call Wolfram on his bloated claims. (“The intellectual dishonesty in the presentation of the Wolfram Language, whether intentional or unintentional, disturbs me, as I’m sure it does many other computer science professionals.”)
But the general question of whether non-human agents using the Wolfram Language, or any other computer language, ‘know about the world’ in some suitable sense, that one still stands.