This time from Can Digital Computers Think?, a lecture in 1951. This passage is particularly important for my thesis too (my emphasis).
Certainly the machine can only do what we do order it to perform, anything else would be a mechanical fault. But there is no need to suppose that, when we give it its orders we know what we are doing, what the consequences of these orders are going to be. One does not need to be able to understand how these orders lead to the machine’s subsequent behaviour, any more than one needs to understand the mechanism of germination when one puts a seed in the ground. The plant comes up whether one understands or not. If we give the machine a programme which results in its doing something interesting which we had not anticipated I should be inclined to say that the machine had originated something, rather than to claim that its behaviour was implicit in the programme, and therefore that the originality lies entirely with us.
The use of mechanisation that Turing alludes to here is completely different from the usual bashing that Stengers et al affords. It’s a deep commitment to deterministic machinery, but with a weird philosophical (almost phenomenological) understanding of indeterminacy which originates not only with the observer, but also the thing thats being observed. It’s a fascination with the machinic novelty, which comes from the machine itself, and from the observer who observes it. The machine just is what it is, executing blind recursive rules; the execution of the rules are performative to others and (this is my own reading) to itself. The machine cannot know, the outcome of it’s own rules, so to speak.
I wonder if the last passage may have been misunderstood. Like most of the philosophical insights the later Turing came up with, he never elaborated or confirmed certain inclinations. The passage: “and therefore that the originality lies entirely with us.“, could be taken two ways; either it could be misconstrued that the machine is devoid of intelligence, and human beings delude themselves into thinking that the machine is surprising, or (and this would be my wager) that the passage fits in with the general tone of the argument. When Turing uses that statement, it’s used to argue against the idea that a human observer can understand that its behaviour ‘contains’ intelligence (or human thinking) in the programme.
This is why early on, Turing disagrees with Ada Lovelace’s definition of a machine, that ‘it can only do what humans can program it to do’. Whatever it is humans programme into the machine, this does not entail that the machine’s real execution conforms to human knowledge. To do so, results in a reductionist argument of human knowledge (if only we had a better understanding of the rules, and more ingenuity, we can engineer the phenomena we want to see). Turing knew perfectly well that any observation of a machine is a second hand performance of it – and this is the whole point of a Turing test. Understanding or having perfect knowledge of what the machine is capable of is impossible. It can only resemble, or manifest as a translation, and for Turing as long as the translation of the machine was enough to appear intelligent, then in-itself, this was enough to be intelligent. Intelligence then, is only related to performance; it could be ubiquitous in the world, despite scientism’s claims to the contrary (funny how Turing, one of the usual poster boy for analytic scientism, implicitly refutes a scientism view of reductionism).
Often enough Turing is referenced as being a materialist, especially when his work is summarised as a simple study into surpassing or meeting human intelligence with computation. These views are usually espoused by technologists and philosophers who regard human intelligence as a failure anyway, and suggest that we ‘grow up’ and embrace techno-science progression.
But when you actually read Turing, one never feels that this was Turing’s intention. Dare I say it, it’s more like reading Whitehead – another English thinker who perhaps no-one would dare link to Turing in a month of Sundays. For a start, it’s fun to read. Turing’s rhetoric is witty, astonishingly clear (I cannot think of many thinkers who came up with as many dynamite thought experiments and comical analogies as Turing) and also he was screamingly funny. You can get a sense that of that famous Turing maniacal giggle, behind the thought processes. I firmly believe that there are crucial misunderstandings on how Turing actually envisioned intelligence taking place in machines. At no point was he concerned that machinery intelligence would supersede human ability – in fact Turing envisaged it as akin to (of all things) a student/teacher relationship. The quote below is taken from the same lecture;
We must not always expect to know what the computer is going to do. We should be pleased when the machine surprises us, in rather the same way as one is pleased when a pupil does something which he had not been explicitly taught to do.
It’s perhaps the greatest irony, that a great thinker who had the greatest obsession with formal language and formal systems, wrote in such an informal manner.