Friday, December 18, 2015

The Brooksy awards 2015. I

Walter Pitts, via La Settimana Anacronistica.
The only slightly amusing thing at first glance in part 1 of David Brooks's annual survey of magazine articles he sort of enjoyed reading, or "Sidney Awards", isn't really amusing at all: a summary of the tale of the cognitive scientists Warren McCulloch and Walter Pitts, as reported by Amanda Gefter in "The Man Who Tried to Redeem the World with Logic":
These two geniuses fit together perfectly. They performed amazing intellectual feats, the first of which was coming up with a working model for how the brain works and laying the groundwork for artificial intelligence.
They also developed an amazing friendship. At one point when they were apart, Pitts wrote McCulloch, “About once a week now I become violently homesick to talk all evening and all night to you.”
Only one person was unhappy with this arrangement: the wife of a third colleague who was jealous of her husband’s academic relationships. She told her husband, falsely, that their daughter had been seduced by his colleagues. That ruptured the whole network of ties.
Violently jealous wife disturbs intensely beautiful intellectual collaboration, which, knowing what we know and what we think we know about Brooks's recent intellectual collaborations and his personal life sounds a lot like projection. An amazing lot, if you know what I mean.

The story of McCulloch and Pitts is a little more complicated. The "third colleague", for one thing, was the most permanently famous member of the team, the inventor of cybernetics, Norbert Wiener, and the effect of his rage wasn't to break the relationship between McCulloch and Pitts, who continued to work together in Jerome Wiesner's brain science project at MIT with Jerome Lettvin, Humberto Maturana, and others. Rather, Wiener cut himself off from all of them, and it was a boil that had been going on for years (1947-52), involving a lost Wiener manuscript (Pitts or Lettvin had lost it), and a good deal of paranoia on Wiener's part; Mrs. Wiener's false accusations were just the last straw in a series of slights and alarms. And while Wiener's biographers, Flo Conway and Jim Siegelman, got the story for their 2005 Dark Hero of the Information Age directly from Lettvin (then in his 80s) himself, there's a good deal of unclarity in the dating.

Trying to tell the story without Lettvin makes it harder. Lettvin and Pitts were both kids, Pitts a 16-year-old homeless runaway, when McCulloch took them into his noisy (four kids already) household in Chicago, and their friendship is as important an element as their relationships with McCulloch and Wiener, but Pitts's contribution to the intellectual endeavor is important in a very signal way: there would literally be no computers without his idea of how to put a memory into a finite-state machine (by looping it), and in that sense he's the missing link between Alan Turing on one end and Wiener and John von Neumann on the other in the development of all information science.

What I get from Gefter's article is an alternative hypothesis about what happened to Pitts after 1952 and the break with Wiener, and it has to do not with vicious rumors and crazy cyberneticists but science. The excitement over the new ideas that McCulloch and Wiener and their students were working over wasn't really about building computers, at least for some of them, but about understanding the workings of the human brain—not how to process information, but how we process it, and it seemed likely that Pitts had discovered the actual mechanism, which is why they referred to his strings of symbols from the start as "neural nets". They weren't as interested in building tools as they were in unpacking the mysteries of the human mind. Pitt's dissertation project was to work out the mathematics of three-dimensional neural nets, an order of magnitude more complicated than the two-dimensional nets they were working with at the time, and all his friends believed it was a task only he could do.

He's supposed to have been nearly finished with this work at the time of Wiener's eruption, and afterwards burned it and the rest of his papers as he turned into the lost and lonely drunk who died alone in a hospital 17 years later, but it's not obvious when that happened. What we do know, though, is about the frogs: Lettvin and Maturano were working with frogs, using electrodes attached to individual fibers in their optic nerves to get a sense of the structure of the information packages sent from the eye to the brain, and they found that even the frog brain was not a finite-state machine but a sophisticated analyst, filtering and categorizing, published in a 1959 paper in which Pitts and McCulloch were listed as authors though neither had done much of the work.

Gefter writes:
The results shook Pitts’ worldview to its core. Instead of the brain computing information digital neuron by digital neuron using the exacting implement of mathematical logic, messy, analog processes in the eye were doing at least part of the interpretive work. “It was apparent to him after we had done the frog’s eye that even if logic played a part, it didn’t play the important or central part that one would have expected,” Lettvin said. “It disappointed him. He would never admit it, but it seemed to add to his despair at the loss of Wiener’s friendship.”
I think he realized, in a way Lettvin and McCulloch failed to understand, that his immense work on the three-dimensional neural net was no good for the purpose he had intended it for—however useful it might be for some future artificial intelligence, it could never duplicate the real thing—and that is when he began to despair, some years after the break with Wiener, and destroyed his work, in around 1959. Lettvin writes, indeed, in his MIT potted biography of Pitts,
All that vanished before the end of the 1950s. He died alone in a boarding house in Cambridge after doing his best for close to a decade to avoid being found by his friends. Nothing of his work was left. 
If I'm right, it wasn't about Wiener at all, let alone Wiener's nasty wife (who does seem to have been an awful person, a German anti-Semite married to an absent-minded Jew hopelessly dependent on her, and it's plausible her real jealousy might have been less over her husband's friendships than those of her daughter, who had loved the year she spent in Chicago in the McCulloch household with Lettvin and Pitts, and didn't like her mother much at all)—though it was an intellectual impasse he might have felt he could work himself through if he had only had Wiener's support, which is pretty deeply sad.

But Brooksy will settle for the bad wife story, because science is hard, and bad wives are, you know. I think I'll just put this up now, and get back to the rest of the column, if at all, later.

No comments:

Post a Comment