In 1980, the Philosopher, John Searle, published a paper criticising the famous “Turing Test” for artificial intelligence. Since the 1950s, the “Turing Test”, named after its creator, Alan Turing, essentially claimed that if an AI produces answers to questions which are indistinguishable from those answers given by a human being to the same questions, then it is obvious that a machine has “intelligence”. Searle’s “Chinese Room” presented a compelling objection to this idea. He asks us to imagine, instead of a computer, that he is the inhabitant of a room in which certain Chinese characters are slipped under the door. He responds to these characters by following a series of instructions (which essentially function in much the same way as a computer programme would), allocating specific Chinese symbols to be returned in response to particular Chinese character inputs. Searle manages to respond to each question satisfactorily, thus passing the Turing Test, and the Chinese Room appears to be understanding and answering all questions put to it in Chinese. However, Searle himself does not actually understand Chinese.
What Searle’s Chinese Room demonstrates is that the appearance of understanding does not always equate to actual understanding. A computer which provides answers indistinguishable from those of a human being does not actually provide satisfactory evidence of understanding the answers it gives, or the questions which cause them. To use Searle’s phrasing, they have “only a syntax but no semantics”. Success in Turing’s “imitation game” (as it is sometimes also known) tells us only that we have successfully programmed a machine to give us the sort of answers someone would give if they understood the questions, but is, by itself, no guarantee that any real understanding exists.
I think of the Chinese Room often as a teacher, especially around exam season, and wonder how much our obsession with testing has led to a Chinese Room approach to learning?
Through our focus on exam technique, and desperate attempts to quantify what is essentially unquantifiable in education by reducing our understanding of “learning” to a single grade, are we actually teaching our students anymore, or simply providing them with programming manuals to successfully manipulate correct output responses to certain preconceived input questions? Are we fostering true understanding of our various academic disciplines, or merely being fooled into believing we have done so by garnering responses indistinguishable from those of someone who has genuine understanding of the subject matter but for whom, in reality, beyond the empty words, there is no understanding at all?
For who among us has not had the unfortunate experience of speaking to a student who recently scored highly in an exam about something on which, on paper, they should now therefore be well versed in, only to see blankness and confusion in their eyes when you try to take that apparent understanding and get them to apply it to something new? Or been confronted, the lesson following some whizzy “AfL” activity has confirmed progress and learning for all, by a class now unable to speak meaningfully about the very same topic?
I have often laughed at the uselessness of lesson observations when lessons graded “outstanding” produce the illusion of understanding in the moment, for the observer, only to be completely undermined by clear student confusion and bewilderment the next time you see them and ask them something about the subject in which you were deemed “outstanding” in, apparently, making them proficient. And of course, I have cried too at the tragic lack of cosmic justice when one of my laziest, most uninterested pupils spent two years doing absolutely nothing before managing to somehow pull an A or A* out of the bag in their exam with some last minute, short-term memory reliant, revision.
Saying the right answers at the right time certainly makes our managers happy, and, whether a sign of true understanding or not, ultimately is all that is needed to get our students into the sixth form courses, university places, and jobs that they desire. But at some point those of us who love our subjects and believe in education as something more than a mere instrumental stepping stone to employment might want to ask if we have dangerously reduced learning to another syntax without the semantics? And if we are just being fooled in a Chinese Room imitation game of real learning by empty answers without real understanding, who will be the ultimate victims of this deception?