In a 2015 article entitled ‘Humans Pretending to be Computers Pretending to be Human,’ Melbourne writer Oscar Schwartz examines the repercussions of the age-old, ‘almost mythological dawn’ of the age of artificial intelligence. Schwartz recounts the tale of Wolfgang von Kempelen’s Schachtürke, a chess-playing ‘mechanical man’ known to English-speakers as the Turk. Appearing before the court of the Empress Maria Theresa of Austria in 1770, von Kempelen made a remarkable claim: this supposed automaton, he declared, was capable of beating humans at chess. And indeed it (he?) was – the Turk humiliated some of the fiercest minds of its time, including (most famously) Benjamin Franklin, Catherine the Great and Napoleon Bonaparte.
The story of von Kempelen’s Turk, with its star-studded cast and its air of intrigue, is the stuff of legend. Sadly, though, the whole operation turned out to be a scam; inside the supposed automaton was a human man, a flesh-and-blood chess maestro, who, interestingly enough, is roundly excluded from retellings of this remarkable tale. When the ‘myth’ of the Turk is recounted, Schwartz notes,
there is a certain character that always remains invisible and voiceless. No one ever talks about the person sitting inside the cabinet controlling the Turk, scurrying among the levers and cogs to avoid detection. The Turk was essentially just an elaborate puppet, and all of its many great achievements were thanks to a real human hiding inside. History has made the Turk into a star, whereas the person inside is a ghost in the machine. We are willing to ignore the human for the romance of a thinking machine.1
Our human fascination with the concept of artificial intelligence has endured well into the modern day. Von Kempelen’s infamous invention itself spawned a slew of literary and cultural phenomena, inspiring the likes of Edgar Allan Poe, Walter Benjamin and Ambrose Bierce. Nothing, however, has ever been written about the nameless genius responsible for the Turk’s celebrity. The history of translation has suffered from a similar invisibility. Like the art of chess, the art of translation was once considered an intrinsically human endeavour. With the steady advance of technology, however, significant dilemmas have begun to surface. Just as the Turk brought questions of human intelligence bubbling to the fore, so does the nexus of machinery and translation cast doubt upon our definitions of human creativity and the value attributed to it.
A curious science fiction novel entitled Galápagos, penned by Kurt Vonnegut in 1985, offers an interesting (if improbable) contribution to the machine/translation conundrum. One of the novel’s many ill-fated protagonists is Japanese computer genius Zenji Hiroguchi, who has recently invented a pocket computer capable of translating many spoken languages instantaneously. Zenji names this machine Gokubi. Five years later he comes up with a pilot model for Gokubi’s successor: a new generation of simultaneous voice translators named Mandarax. Mandarax has been programmed not only to instantaneously identify and translate between over one thousand human languages, but also to diagnose diseases, recite poetry, and even instruct its users on the venerable art of ikebana. Zenji’s wife Hisako, herself a talented kadoka, is naturally upset when she discovers this last, surprising feature. For Hisako, the reduction of such a nuanced and ineffably beautiful art form to the cold, sterile language of the machine betrays an inexcusable contempt for human artistry. ‘You, Doctor Hiroguchi,’ she snarls, ‘think that everybody but yourself is just taking up space on this planet… That wonderful Mandarax you’re scratching your ear with now: what is that but an excuse for a mean-spirited egomaniac never to pay or even thank any human being with a knowledge of languages or mathematics or history or medicine or literature or ikebana or anything?’2
The point Hisako raises is an important one. Does the very supposition that computers – now or in the not-too-distant future – might be capable of convincingly imitating, mastering, or indeed improving upon human practices inevitably diminish the value of human enterprise? Possibly. There is certainly ample historical evidence to suggest that professions tend to become obsolete once they are deemed replaceable or, at the very least, roughly replicable by technology. Perhaps, though, we should be asking a different question altogether: is any effort to reproduce acts of human passion and creativity – acts that verge on the spiritual, like the composition (and translation) of poetry, or the resurrection of flowers – necessarily futile? Does the idea of machine-made poetry miss the point of poetry entirely? Is there in fact an inherent value in human endeavour, in l’art pour l’art, precisely because it is human?
If von Kempelen had presented his marvellous Schachtürk at a 2017 technology trade fair, no one would be impressed. As we all know, in the years between 1770 and the modern day humans managed to design computers advanced enough to trounce their human opponents in chess; in 1996, IBM’s ‘Deep Blue’ defeated Garry Kasparov to become the first computer system to win a chess game against a reigning world champion under regular time controls, and since then the technology has improved exponentially. Machine translation, while somewhat more complex from an algorithmic perspective (the logic of language, it seems, is less straightforward than the logic of chess) is following a similar path. Along with remarkable improvements in the quality and accuracy of Google translations, increasingly sophisticated speech recognition technologies have paved the way for ground-breaking translation software (and hardware) such as Google’s recently released Pixel Buds: wireless headphones featuring real time language translation from Google Translate. Douglas Adams’s Babel fish is looking less and less like science fiction every day. But these technologies, mind-boggling as they are, may still exist on a plane that is tangential – perhaps even unrelated – to the artistic plane. Where translation is concerned, the crucial divergence occurs at the intersection between language and art: precisely at the place where poetry is located.
In Vonnegut’s Galápagos, Zenji Hiroguchi’s wondrous translation device is proven hilariously wanting when it comes to the selection of stirring poetic sentiments. ‘When Mandarax was asked to come up with quotations from world literature which could be used in a celebration of some event,’ Vonnegut writes, ‘… the instrument almost always came up with clunkers.’ Its suggestion when one character gives birth, for example, is the following excerpt from a rather grim Kipling poem:
If I were hanged on the highest hill, Mother o’mine, O mother o’mine! I know whose love would follow me still, Mother o’mine, O mother o’mine!
The machine’s ‘expert’ medical advice, administered to Zenji’s wife (now widowed and stranded on the remote Galapagos Island of Santa Rosalia) is similarly lacking in tact: ‘When Hisako Hiroguchi entered into a deep depression … Mandarax recommended new hobbies, new friends, a change of scene and perhaps profession, and lithium.’ The inadequacies of Vonnegut’s fictional machine, outdated and waggish as they are, hint to a question that has nagged thinkers ever since the dawn of the mechanical age: can technology ever breach the realm of the phenomenological? And, if not, what does it mean to translate human thought and experience? What would such a thing entail?