An x-ray picture of a man's head, which is filled with various tools that resemble the interior of a normal head.

Clockwork Conversation: Not Everything Could Be Half of Something

In 1562, Don Carlos, the seventeen-year-old heir apparent to the Spanish throne, falls down a flight of stairs. Tragically, he sustains a terrible head wound. His father, King Philip II, orders physicians to attempt every available cure. Nothing works – not the birthwort powder applied to the skull, not the mixture of turpentine and egg yolk to the lips. What to do?

Philip turns to religion. Monks bring in the physical remains of a deceased member of their order: the holy Brother Diego de Alcalá. This is a century after Diego died, a decade before he was sainted, and forty years before San Diego, California was named after him. The monks lay the remains beside the suppurating, head-swollen Don Carlos. Desperate and distraught, Philip promises God a miracle in exchange for a miraculous recovery. The deal is accepted; Don Carlos is up and about the very next day, telling of fever-vision conversations with a monk. It is clear to all involved what has occurred. The king must now keep his side of the bargain – he owes God a miracle.

Philip commissions Juanelo Turriana, noted maker of clocks and automatons, to create his gift for God: a wood and iron mechanical monk. The monk is an automaton and, once wound up, moves about nearly on its own. Remarkably, this thing survives and today lives at the Smithsonian’s National Museum of American History. Here he is doing his thing:

This fifteen-inch masterpiece goes through the motions of penitence: the robomonk beats his chest, bows his head, rolls his eyes, kisses his cross, and shuffles to and fro. Philip II has fulfilled his promise with the ultimate votive offering – the gift of perpetual prayer.

The monkbot’s clockwork prayer derives its religious and sociological power from the magical dimension of mechanical repetition. This is the power source of the Catholic rosary, Bhakti yoga japa beads, Tibetan prayer wheels, and shamanic drumming. The automaton’s prayers are formally correct and therefore technically effective. Philip’s gift to God was, indeed, a steady stream of supplication – so long as the little guy is occasionally rewound.

As a very early immigrant from the uncanny valley, the thing is a spectacle – then and now. An engineering miracle and an engineered miracle, this particular automaton skirts the border between science, magic, and theatre. And as with all mimics, puppets, and other moving simulacra, we wonder at the technique of the illusion while also getting lost in the performance. It’s not real; it’s real.

But does it have soul? Does it touch the divine and sublime on Philip’s behalf? Do the mechanical monk’s prayers count as prayer? And if so, for whom? The king? Juanelo Turriana? The beholder? The monk itself? Is a good mimic good enough?

I don’t know.

Automata have come a long way. No longer an obscure curiosity of the clockmaker’s art, human mimics move through our lives most every day. The artificial intelligences of the internet, chatbots, self-aware robots, and smart cars are far more central to our daily lives than was any gear-driven doll in the sixteenth century. A Bloomberg report earlier this week about Google’s dedication to AI underscores the business giant this technology has become.

It takes more than being intelligent, however, to truly mimic humanity. Due to its awkward immediacy, the automaton monk may in fact be creepier than the Facebook News Feed Algorithm. Set up computer intelligence to do something more tangibly human, however, and the uncanny flag unfurls. For example, watch this video of two Cleverbot chatbots conversing with one another using real language:

Their conversation takes me right back to the mechanical monk (who was, admittedly, better dressed than these business-attired avatars). There are times in this video when I am truly gobsmacked. “Not everything could be half of something” and “don’t you wish you had a body” – these novel expressions are haunting. The conversation among the bots is, frankly, as deep and confusing as one I may have on any given day.

Almost.

So, have the Cleverbot clockmakers crafted a thinking and speaking machine? Is language really just this? The syntax and diction are correct. The mechanics of conversation are present, or at least seem to be.

While watching this, I am always aware of the engineered nature of the thing. I know it’s a puppet. But it’s conversing! Right? It’s not real; it’s real.

But does it have soul? Do its words and phrases count as expressive speech? And if so, for whom? The engineers? The beholder? Cleverbot itself? Is a good mimic good enough?

I don’t know.

But I suspect not. With both the Diego bot and the Cleverbot, there is something that feels out of place, or not in place at all. In both cases, the ghost in the machine is not a ghost at all, but a human touch. While they’re clearly amazing and magical in their way, they lack the mundane magic at the core of actual vitality. But is approximation enough? Not everything could be half of something. Right?

Corey Pressman is the Director of Experience Strategy at Neologic, a member of CSI’s Imaginary College, and the director of Poetry for Robots, a digital humanities experiment that is trying to teach poetry and human metaphors to robots and computers.