The systems objection, which I take to be identical (or at least nearly so) with the virtual mind reply, suggests that the Chinese Room argument is entirely beside the point. When Searle says “It seems to me quite obvious in the example that I do not understand a word of the Chinese stories,” the computationalist merely need reply, “So what?” The English speaker has a representation of Chinese-speaking syntax, a phenomenon which is entirely independent of the representation which is supposed to be created by the latter. This confusion of representation/consciousness/understanding of syntax and the syntax of representation/consciousness/understanding is the primary motivation which makes Searle’s conclusion so ‘obvious’ or intuitive.
The irrelevance of the understanding of the English speaker can be illustrated by simply replacing him with a Chinese speaker or, for the sake of convenience, make the English speaker run an English-speaking program rather than a Chinese-speaking program. Let us follow Searle’s example in calling the man in the room ‘John Searle.’ It is important to note that when one asks the English Room what its name is, it will not answer ‘John Searle.’ The name which will be given will be that of the program which John is running (let us call it ‘Al’), not the program which (supposedly) is John. John, in as much as he understands anything, will see the interaction between Al and the outside interlocutor as a conversation between two intelligences, neither of which is him. In order to better appreciate this point, we will have to clarify another ambiguity which lies in Searle’s account.
Before we do this, however, lets first point out that the difference between the program which John is running, Al, and the program which is John is an illustration of the difference between the representation of syntax and the syntax of representation. Once that this difference is appreciated, the fact that John understands nothing in the Chinese Room setup becomes entirely irrelevant, for even if he did understand the symbols which he was manipulating, we have no more (or less) reason to believe that Al understands anything. John’s understanding is completely independent of, and therefore completely irrelevant to the understanding of Al.
Of course both of these situations, the Chinese Room and the English Room which I have set up, are almost certainly wrong. The fact that the input and output of a program are Chinese characters gives us no reason to assume that the program itself runs on some syntax composed of Chinese characters. While Searle never claims any different, this points to a significant gloss in Searle’s account: what is the nature of the program which he is running in the Chinese Room? Sure, it gives Chinese output in response to Chinese input, but what happens between the two? This is hardly a trivial point since this is where the relevant understanding (Al’s as opposed to John’s) is supposed to happen. If all we are concerned with is John’s inability to understand the input and output of the Chinese Room, all we have said is that John does not understand Chinese. We have said nothing at all about the program itself or its formal structure.
To recap, John’s instructions are in English. The input and output are Chinese characters. But what is the actual syntax which the Chinese speaking program is running on? Searle’s Chinese Room has both the input/output and the syntax as being Chinese characters, but this only serves to confuse the issue. Let us, instead, change the scenario a bit in a way which the Searle’s response to the Robot reply seems to demand: let us suppose that the syntax of the program is entirely different from the form of the input/output. His counter-reply seems to merit such a distinction, for he suggests that the sensory input/output of the robot is just more Chinese characters. If this is so, then John, from inside the Chinese robot, has no way to distinguish between sensory input which originates from Chinese characters, trees, robot-breathing or, more importantly, spoken or written English. In other words, we could have an English Room in which the English-speaking John runs a program written in a syntax composed of Chinese Characters.
Let us now ask Searle’s all important question: Does John understand? The answer is both yes and no, but neither answer is at all relevant. It is for this reason that I suggested that when John runs the English Room scenario he would likely not understand very much. Inasmuch as John could understand the input/output (since they are written in English), he would merely understand it as a conversation between two intelligences, neither of which is him. One intelligence would reside outside the room. The other intelligence would be that in, or responsible for the program which he is running. Whether John understands the input/output or the syntax of the program is entirely independent of whether the program responsible for the input/output is intelligent or not.
