“Men can do all things if they will.”
It might almost be a motto for Humanity+ itself. But it was written by Leon Battista Alberti, a fifteenth-century Italian polymath whose interests would have seemed equally at home in the twenty-first. An architect, artist, and philosopher, he was also a cryptographer who invented the first polyalphabetic cipher, and he made the first scientific study of perspective. The investigations Alberti began on how painters could represent three-dimensional space in a two-dimensional image have revived today, in the field of machine vision. And Alberti’s obsession with vision and the eye as the symbols of understanding have their echoes in modern research that views perception as central in artificial intelligence.
Alberti’s varied interests were linked by a personal belief, new in its time and still not fully appreciated in ours: the conviction that, with effort and foresight, human beings could avoid the common calamities that might otherwise be chalked up to “Fate” or “Fortune.” His advice to young men in The Book of the Family (a sort of proto-self-help book) urged them to “conquer Fortune with patience.” Where medieval philosophy urged resignation and endurance, humanists like Alberti urged their ambitious merchant readers to outwit and outplan Fortune. In a sense, that’s not too far from the transhumanist project – the aspiration to do something about death, aging, limited resources, natural disasters, etc., instead of merely enduring them. To see the present as humanity’s troubled childhood, which someday we might outgrow.
Alberti was an almost comically self-invented man. He was illegitimate, his family was expelled from his home city of Florence, and his mother died of the bubonic plague when he was young; his faith in human ingenuity may have been based on his own experience with overcoming harsh circumstances. He gave himself the name “Leon” because he admired the regal qualities of the lion, and in self-portraits drew himself with a lion’s mane of blond curls (which he likely didn’t have.) His self-portrait is marked with his symbol: a winged eye radiating thunderbolts, and his chosen motto, “Quid Tum?” This translates either as “What next?”, a challenge for the future, or “So what?”, daring the viewer to call him a bastard. Quid tum? I define myself. My origins, my body – that’s old news. I’m looking ahead. The same motto might hold for Aimee Mullins with her twelve pairs of prosthetic legs, redefining the scope of the human body. It’s a motto that says nature can be hacked in the interest of human self-invention.
Alberti’s insights into perspective, which sparked a century of innovation in Italian painting, today form the basis of much of the modern science and industry of computer vision and graphics. The central question, then and now, is the relationship between 3-d and 2-d. Given a three-dimensional object, how do you show it in a two-dimensional image that gives the illusion of depth? Given a two-dimensional image (or, in fact, many of them), how do you reconstruct the three-dimensional object they represent?
Alberti, in his On Painting, develops the mathematical model for painting linear perspective. We imagine a “veil” – a plane, vertical like a TV screen – between us and the scene we want to paint. Draw a straight line from any point in the scene to your eye; then mark where the line intersects the “veil.” This is where that part of the 3-d scene projects onto the “veil.” The “veil” will become your canvas. This model is enough to produce perspective drawings and paintings; with a little calculation you can verify that faraway objects will converge to a “vanishing point” just at eye level. After Alberti’s work was published, Italian painters quickly adopted the new style of strict mathematical perspective; Western art was never the same.
The corresponding challenge for contemporary researchers is to do this in reverse – when you have two-dimensional images (say, photographs or videos) from different viewpoints, how can you infer depth and three-dimensional structure? This is called the structure from motion problem. Just as in Alberti’s model, 2-d images are considered as “veils” or planes, onto which 3-d scenes are projected. About twenty years ago, researchers showed that if you have at least three different viewing planes, you can solve for the positions in 3-d space of all your observation points (Kontsevich et al 1987, Tomasi and Kanade 1992). This has major applications; imagine being able to take a picture with your smartphone and recognize your location by how the photo corresponds with a 3d database inferred from other people’s photos. Or imagine getting a 3d virtual tour reconstructed from tourists’ photos, like this model of the Colosseum. This November, researchers at UNC and ETH Zurich published 3-d models of Berlin and Rome stiched together from millions of user-generated photos. Microsoft’s Photosynth has put this into practice; user-generated photos are stitched together into panoramas of sites around the world. Similar algorithms can be used for orienting anything from archeological digs to crime scenes from photo data. Six hundred years later, Alberti’s influence keeps yielding new technologies.
Alberti wrote, “the eye is more powerful than anything; swifter, more worthy; what more can I say? It is such as to be the first, chief, king, like a god of human parts…we are reminded to be wide-awake, all-embracing as far as the power of our intelligence allows, in order to find out all things that lead to the glory of excellence.” An inspiring metaphor – but is it any more than that? As a matter of fact, yes; there are modern researchers who see in vision (and sensory perception more generally) the key to understanding and ultimately replicating intelligence. Jeff Hawkins, who invented the Palm Pilot and currently works on neuroscience full-time, argued in his On Intelligence that our awareness of external objects depends on our ability to detect patterns across space and time from the stream of sense data entering our brain. “Patterns,” he wrote, “are the fundamental currency of intelligence.” He argues that higher intelligence, the ability to make predictions about more abstract patterns, proceeds on the same algorithms as low-level perceptual inference, just at a higher level in the brain’s hierarchy. Before we can make a machine that can reason the way we do, we might have to make a machine that can understand what it sees. Vision might not just be a metaphor for the power of a perceptive mind, but the literal key for how that mind works.
There’s some benefit, even for futurists, to looking backwards to see where our current ideas originated. And transhumanist ideas grew out of very old humanist ones – belief in human ingenuity, self-creation, and the ability to use our minds to overcome harsh circumstances or unpredictable fortunes. When we want to improve the human condition by extending the limits of technology, we’re no more radical than Alberti or Pico della Mirandola were when they expressed the shocking idea that men of their time should seek to surpass the ancients. Leon Battista Alberti – with his quirky audacity, far-reaching innovation, and surprisingly prescient insights – was, perhaps, one of us.