Can a machine think? Can a thinking machine tell us something about ourselves – or does it need to ask its own questions?
Abstract: The Turing test poses problems for explanation in supposing that different causes, synthetic and biological, can converge upon the same effect, namely intelligence. In particular, the recognisable change from cause to effect depends upon a capacity for recognition that cannot be subtracted from the appearance of intelligence or its explanation.
Historical background: Alan Turing was a pioneer in the field of Artificial Intelligence. In 1950 he devised a famous thought experiment as an objective means of assessing the equivalence between an intelligent machine and a human competitor: If an independent examiner, who can’t see whether he is dealing with a man or a machine, cannot discriminate between their performance, then it is reasonable to assume that the machine is intelligent, indeed thinking for itself, and that thinking and intelligence can be explained as a programme. Various advances have been made since then plus diverse claims about the prowess of thinking machines. There is an annual competition called the Loebner Prize which is broadly based on the Turing test. The 2014 competition was held at Bletchley Park and won by a machine called Rose, which was awarded a bronze medal. If a machine eventually passes the Turing test, a special prize will be awarded and the competition will end.
Causality is a conundrum. Everything we know about atoms and molecules doesn’t tell us what comes next until we see what comes next. So it is by the nature of the effect that we are able to establish the nature of the causal properties of the atom – observable in the changes attributed to it. These attributions reassure us that the change is within bounds, but those bounds are discoverable by observation of the recurrent fact of the effect, not the continuing presence of the cause in its original state. So all we have really established is that things change, with implications for both the cause and the effect.
In fact all we see is change as a fact of change. The causality is a fact we have construed. Our causal proof is based on the fact that the cause came first, in its unchanged state, plus the ‘fact’ that there is nothing else to observe – save for the effect that results. But this observational framework is directly challenged when it comes to our understanding of our own thinking – because here the effect comes first, as an active observational prerequisite, and all the facts we can observe as causes can’t match the nature of the effect we seek to explain.
Likewise, we attribute consciousness to its causes knowing that any comparison is based on the change to consciousness – an effect that supplements the cause, meaning the cause is less than the effect – a difference that calls into questions the explanatory power of our causal proofs. However, the Turing test proposes a situation in which the difference disappears when comparing our version of thinking with an ‘intelligent machine’ – suggesting that if we can’t tell the difference we can dismiss it, and implying that we can also ignore the physical differences in its causes. Also, we can begin to understand the mind as a physical process, knowing that the Turing machine is an entirely material thing.
But what does the machine know? Is it aware that it knows – of knowing as a state of awareness? Is awareness no more than a physical process? Presumably an alert machine would recognise this much of itself and could simply tell us or show us, providing the answer from the material information in its operating systems and programmes – the corollary being that the reality equates to its physical activities, that there can be nothing more to our thinking, knowing and understanding. In this vein, the ‘change’ to awareness is seen as a feature and function of its physical causes; but it was the change to awareness that led us to identify this feature with its causes, and without the effect emerging as something else we have no cause to attach any function to the cause. Furthermore, the possibility of different causes, biological and synthetic, leading to the same effect also endorses the significance of the effect as a real change. Whichever way we look at the facts we can’t escape the fact that the special nature of our state of awareness is really its special nature as something else, which can’t be the nature of the cause as it was.
An explanatory gap lurks within our theories of where thinking comes from, as if its properties can and need to be known as the output of something else – to the point that we can also identify our awareness with that something else, in its cause. Current reasoning avers that we can’t really know what knowledge amounts to in ourselves, subjectively, without observing it objectively as an objective fact in the world, with the unknowing unconscious physical cause instated as the complete explanation. This is the same reasoning that subsumes the nature of the effect to that of the cause. Accordingly, reason itself is seen as belonging to the properties of an external world and as such gives our reasoning its authority. However, the universe is an unthinking fact occupied by a thinking fact – in one sense it can be seen to give the thinker their thoughts, but in another, very real sense, it cannot – for it has no thoughts to convey.
Even though science hasn’t explained the physical nature of consciousness ‘it’ remains sure that there is a physical explanation in its causes. But causality is a conundrum. It comes down to the fact that we know the conscious mind is possible, but we are expecting the impossible of it in trying to explain it away in the properties of something else that is less than conscious – in the unconscious causes of the physical world. These material proofs attempt to reconstruct the mind as an avatar, by which we might know it better. However, even the ‘thinking machine’ cannot show us what its thinking is really like without doing something really unusual, without joining us in the speculation over its own faculties and their origin – for it is evident that the reality is simultaneously one thing and another, being one thing as the cause and another as the effect. It appears that we introduce our own reasons for wanting the appearance of appearances to be a reality we must explain in terms of its insentient causes.