Lilac Tuesday and Artificial Intelligence

Every May, without fail, the Dowbrigade drifts into a fragrant nostalgic reverie when the lilacs come into bloom. We grew up in upstate New York, near Rochester’s Highland Park, which together with the Royal Botanical Gardens in Burlington, claim to be the largest collections in the world, and in our neighborhood every house had at least a few bushes (Mom counted 19 at our house on East Boulevard). Peak lilac season comes later in Rochester; the last week of May or the first week of June. For fifteen days the normally dowdy aging urbe would be decked out in garlands and leis and smelling like Aphrodite in heat. I remember that each year at the height of lilac season the P’s would invite my elementary school teacher over for dinner. By that time the interminable upstate soggy spring had given way to better weather and we usually ate out on the porch, with lilac bushes pressing against the screens and a few sprigs of purple, or white, or vermilion in vases on the table.

New England’s best lilac collection is at the Arnold Arboretum, the Harvard University botanical collection in Jamaica Plain, which boasts over 450 bushes, including a magnificent variety of colors and forms, both single flowering and the rarer double flowering “French lilacs”. We had hoped to take Norma Yvonne last weekend, but the poor woman is currently working a full-time bank job and teaching 10 courses in the evenings and consequently basically rolls up into a ball when she finishes at 12:45 Saturday and tries to recuperate as much strength as possible before she has to start again on Monday morning.

Besides, how could I miss Bar Camp?

So, taking advantage of a break in our teaching schedule, we went out yesterday. After an hour or so of tramping around the Arboretum listening to Jimi Hendrix on our chi-pod, we wandered into the lilacs. We approached the collection, strewn on a sunny hillside, sort of like an art museum. We would approach each bush like the masterpiece it was, peering from various distances. By squinting, or letting our orbs slide slightly out of focus, we could make each flowery explosion of color into a Cezanne, or a Monet, or a Pissaro. Finally, we would stick our head into the fattest and most fecund bunch of blossoms and inhale deeply.

Our head was spinning by the time we plopped down next to a particularly pungent purple exemplar and opened the book we had brought with us, Marvin Minsky’s “The Emotion Machine”, which is modestly subtitled “Commonsense Thinking, Artificial Intelligence and the Human Mind.”

We must say we are disappointed, so far (we are still on Chapter 2). His basic idea is that the human mind is very complicated and contains many specific abilities which he calls “resources” and active agents he calls “critics” which determine which resources will be brought to bear on each situation or problem.

Our first impression is that this ground was covered more comprehensively 40 years ago by John C. Lilly, who argued in Programming and Metaprogramming in the Human Biocomputer (1967) that because we don’t have time to think out each step of our reactions in real-time situations, we rely on pre-established programs of how to behave, and that which of these programs snap into place when we are confronted with specific cues or triggers is controlled by metaprograms.

“Each mammilian brain functions as a computer, with properties, programs and metaprograms partly to be defined and partly to be determined by observation. The human computer contains at least 13 billion active elements….Among other know properties are self-programming and self-metaprogramming.” (Lilly ’67)

Lilly goes on to argue that humans can gain access to and actually reprogram these metaprograms through a variety of techniques including meditation and other rigorous mental disciplines, transcendent religious experiences and directed drug use, an approach which seems to offer more possibilities for actually improving our system software than Minsky’s.

In fact, we have spent the past 35 years, since we first read Lilly’s seminal paper as an undergrad, trying to implement its implications and modify our metaprograms. We can report mixed results: we no longer become a stuttering idiot in the presence of beautiful women, but we still break out in a cold sweat at the thought of an approaching dental appointment.

Still, perhaps we are jumping to conclusions. The remaining 7 chapters in the Minsky book may contain some new ideas or useful elaborations.

We didn’t bring any lilacs back to Norma or to brighten the apartment. Breaking branches off a living bush seems like a sacrilege today, although we know they will fade and die on the vine soon enough. It’s enough to know that they’ll be there still next May, filling the late spring air with the scent of seasons long past and far away.