Sunday 13 May 2018
















































Murmur by Will Eaves  {Reviewed by THOMAS}
No algorithm can entertain a proposition as being both true and not true at the same time. This necessary computational allergy to contradiction enabled Alan Turing during his time at Bletchley Park in World War 2 to significantly decrease the time it took to break the German codes (“from a contradiction you can deduce everything,” he wrote), thus saving many lives. It would also provide a good test for ‘thought’ as opposed to ‘computation’ in artificial intelligence, and have implications for the eventual ‘personhood’ or otherwise of machines. “Machines do nothing by halves,” writes Will Eaves in Murmur, a beautifully written, sad and thoughtful novel based on Turing. Machines cannot but incline towards explication, whereas it is our inability to access the mind of another that verifies its existence as a mind.* “It isn’t knowing what another person thinks or feels that makes us who we are. It’s the respect for not knowing,” writes Eaves. In 1952 an English court found Alan Turing guilty of ‘Gross Indecency’ for admitted homosexual activity (then a crime in Britain), and he submitted to a year-long regime of chemical castration via weekly injections of Stilboestrol rather than imprisonment. Turing’s chemical reprogramming, speculates Eaves, struck at the core of his identity, his mind at first barricading itself within the changing body and then seemingly inhabiting it once more, but resourceless and compliant. Is personhood always thus imposed from without, or does personhood lie in the resistance to such an imposition? What conformity to expectations must be achieved or eschewed to accomplish personhood? The murmur inMurmur is an insistent voice that rises from Alec Prior’s (i.e. Alan Turing’s) sub-computational mind as it reacts to, and reconfigures itself on the basis of, its chemical reorientation. A narrator in the third person, waiting in ambush in mirrors and other reflective surfaces, Prior’s reflection, assails and supplants Prior’s first-person narrative, breaches the functional boundaries of his identity, describes Prior as “a man in distress, a prisoner of some description?”, unpicking his autonomy, and acting as a catalyst for the emergence of material (memories, voices, impulses) from the deep strata of Prior’s mind, much of it foundational (such as Prior’s formative relationship with a fellow student at high school), atemporal or, increasingly, counterfactual (a series of imagined letters between Prior and his friend and colleague June, with whom he was briefly engaged (parallelling Turing’s relationship with Joan Clarke, June was unconcerned by Prior’s homosexuality but he decided not to go through with the marriage) veers towards a confused and non-existent future in which a child of theirs remarks to Prior, “You’re changing. You’re lots of different people, lots of things, and all at once.”). What is the relationship between memory and fantasy, and what is the pivot or fulcrum between the two? When the first-person narration restabilises it is a new first person, one constructed from without (“There was another me, speaking for me.”). Consciousness is detached from what it contains, but made of it. “I am the body in the bed. I’m what sees him. I am the room.” But it is consciousness’s detachment from its object, its resistance to connection (a machine cannot help but connect), its yearning for what it is not and what is not (“yearning is a sort of proof of liberty”), its inaccessibility, its ability to see itself from the couch of its exclusion (“a shared mind has no self-knowledge,” writes Eaves-as-Prior-as-Turing), its cognisance of the limitations of narrative, its capacity to suspend disbelief in fictions, its ability to use a contradiction as a stimulant to thought rather than a nullification, its fragility and tentativeness that distinguishes thinking from computation. Artificial intelligence will not achieve personhood through mimesis, learning or algorithmic excellence, but only, if ever, through qualities that eschew such virtues: “We won’t know what machines are thinking once they start to think.”



* “As soon as one can see cause and effect working themselves out in the brain, one regards it as not being thinking, but a sort of unimaginative donkeywork. From this point of view one might be tempted to define thinking as consisting of ‘those mental processes that we don’t understand’. If this is right, then to make a thinking machine is to make one that does interesting things without our understanding quite how it is done.” - A.M. Turing (‘Can Automatic Calculating Machines Be Said to Think?’ (1952))

No comments:

Post a Comment