At the international conference on Communication with Extraterrestrial Intelligence [CETI] held at the Byurakan Astrophysical Observatory, Yerevan [Armenia] USSR, 5-11 September 1971, Marvin Minsky contributed not only his usual wealth of original ideas (including the statement that "instead of sending a picture of a cat, there is one area in which we can send the cat itself") but an actual flying rocket, as was documented in the proceedings. I asked him about this in 1994, and he replied:

"Yes. I brought a miniature water rocket to the meeting, assuming that it would amaze everyone, because it embodied such high technology in a plastic children's toy.  The second stage was automatically launched from the top of the first stage when the first stage pressure fell low enough to release the docking device.  The same company also actually made a 3-stage toy rocket -- and I don't recall which version I took to the meeting.

“The rocket worked very well, and the upper stage ended up on the roof of the Observatory.  It may still be there, for all I know.

“The meeting was in all respects thrilling.  I returned with Crick to Moscow for a few days.  We got tired of being supervised, so one day we went for a stroll by leaving from a rear exit of the Academy of Sciences Hotel.  When we returned, our "interpreter" was in tears. She said she feared she'd be punished severely if she "lost" us again.

“I argued that eventually we should be able to make nanorobots for interstellar travel, but most of the participants were skeptical about the possibility of intelligent machines of any size." 

As a person and a scientist, Marvin was always a true original, out-of-the-box thinker. While he is of course widely recognized as one of the founders of the field of Artificial Intelligence (AI), his views were often at odds with the majority of the AI community.

For many decades, and even today, AI was plagued by what some of us refer to as "physics envy." Researchers sought a handful of universal principles or mechanisms that could explain and produce human-like intelligence. Marvin constantly reminded us that the real solution was likely to be a lot more complex. He described a myriad of different mechanisms that may be involved in producing intelligence in his books The Society of Mind and The Emotion Machine and emphasized the importance of giving computers large amounts of "common sense knowledge," a problem few AI researchers, even today, have attempted to tackle.

I suspect that gradually the field will come to align with his views, recognizing that his thoughts and writings have a timeless, deep quality.

I’ve just sent some reminiscences of Marvin to his children, and have been rummaging through my memories, and I’ll share some of what I have recovered with Edge. My first meeting with Marvin, in the AI Lab at Tech Square back about 1973 or 74, was disconcerting. Everybody was talking about “the Frames paper” (eventually “A Framework for Representing Knowledge” MIT AI-Lab Memo 306, June 1974), and Marvin was having a great time in the Play Room refusing to make his points in stodgy regimented form, tying everyone up in knots. What I eventually came to realize was that there was a sort of temperature inversion in the AI Lab. Many academic departments and labs divide into the young Turks, who boldly defy the establishment rules, and the old farts, who wring their hands (or shake their fists) about the erosion of standards, the lack of discipline, and so forth. But Marvin, playing the unusual role of old Turk, was leading a bunch of young farts on a merry chase. When I heard the standard gag about him—the world’s smartest five-year-old—I got it immediately, and never thought of it as negative. His playfulness was what I particularly adored, even when it backfired, as it sometimes did.

Who else had a life-sized papier mache moose in his living room and a trapeze over the dining table? Then there were the two grand pianos, loaded down with dozens of robotic toys from around the world. A museum of curiosities to ravish the imagination of any smart child. One time when I went to discuss something with him at his home, he was on the phone when I got there, so I looked around at all the marvels in the living/dining room while waiting for him to finish his call. I was struck by a new table lamp on one of the shelves, the base made out of some intricate piece of presumably antique brass machinery. I looked at it, studied it, prodded and poked it, pushing the levers and turning the knobs, examining it in every way I could without picking it up or taking it apart, trying unsuccessfully to deduce what the lamp base had originally been. (I’ve actually written about such puzzles, in an essay on “Artifact Hermeneutics” and in several book chapters. I rather pride myself in my abilities as a reverse engineer.) But this time, I was stumped. When he finished the call, and turned to me, I asked, “OK, Marvin, I give up. What is it?” His answer: “It’s a lamp.” It had been made for him by a sculptor/tinkerer/constructor friend—I wish I could remember who—and it was a hook designed to catch me and people like me. It worked to perfection. I wonder how many others he caught with his lure. 

I also remember a time when I invited him to Tufts for some sort of discussion with my students and they turned out to be entirely too docile and respectful in the company of such greatness, and they retreated into variations on deer-in-the-headlights. Marvin was never comfortable when everybody agrees and nobody puts up a fight, so he couldn’t just sit there and hold forth wisely. He started saying more and more outrageous things, trying to shock the students into some kind of rebuttal or challenge or even laughter. They sat stupefied, and possibly even alarmed. If I hadn’t finally intervened and challenged him on something (I had so wanted the students to do the disagreeing) I think he would have said that elephants fly and Mars is a movie set. Anything to cause a rumble.

Marvin Minsky’s many ways of thinking are an acquired taste, and lots of very smart people just don’t or won’t get it. I think those of us who learned so much from him should make sure we keep his books and articles in the syllabus, so that, although many of our students will turn away baffled, a few free spirits will be lifted out of their chairs, their heads swarming with new ideas, and carry on.

The last time I saw Marvin, just a few months ago, he was hanging out in his wonderful house, front door unlocked, students dropping by unannounced. One young MIT student had worked for a summer in a circus, and naturally a trapeze hung from the vaulted ceiling. She slid upwards like a cat and swung about as we all argued about AI, just like it was forty years ago.

I remembered when that trapeze was being installed, and I was the young student. Why was it hung there? I don't remember, but it was also when the tuba arrived in its place under a piano, now obscured by books, telescope parts, many wonderful things.

On my way to see Marvin that night I got a call from a mutual friend. "Don't argue with him, he's frail." I couldn't believe what I was hearing. "But Marvin thrives on arguments."

I was right. Marvin said, "What you're doing, criticizing AI, it's perfect. If you're wrong in the big picture, you'll make AI better. There's a lot of terrible work, after all. If you're right in the big picture, then you're right. So great!"

Marvin invented about half of the way we think about ourselves these days. His particular way of characterizing AI consumed a million imaginations. Marvin's narrative about the future of machines is the thing people are afraid of. But that's a sideshow. The main event is that Marvin's way of thinking about people and our emotions has more or less replaced Freud's mythology. Pixar's Inside Out, for instance, feels and even looks like Marvin's lectures from decades ago. (He used to ask that we imagine our brains painting colors on memories of things or events so that we might react to them with a given emotion, for instance.)

And all this might be taken as an aside from his work at the foundations of computer science. And his technical contributions to so many other fields. The latest Virtual Reality optics have been influenced by one of Marvin's inventions, for instance: the confocal microscope.

Why was Marvin so generous to me? I gave him grief. I disagreed with him at every turn. I wasn't ever his student, officially, and yet he mentored me, inspired me, put serious time into helping me. His kindness was total, a singularity of kindness.

He came out to visit in California in the 1980s, when I was in my twenties and Virtual Reality was getting tolerable. He sat in a headset—was it a simulation of being inside a hippocampus with neurons firing?—while at the same time playing a real physical grand piano, and somehow the two planes of reality became beautifully coordinated.

The music! Everyone knows Marvin improvised at the piano in the approximate style of Bach—elaborate counterpoint—but he never ever fell into a rut. He was just as fascinated by the obscure musical instruments I brought by, from around the world. Since everything was always new to Marvin, even Bach's style was always brand new. Marvin lacked the capacity to become jaded, bored, or fall into any state of mind shy of being startled by the constant novelty of reality.

I remember Marvin talking to Margaret, his daughter, and me about his take on Alan Watts. It's hard to imagine a philosopher who might seem more distant from Marvin than the guru-like, Asian-leaning Watts, and yet Marvin thought Watts was remarkably wise about death. I recall Marvin discussing Watt's idea that reincarnation is the wave way of interpreting people instead of the particle way. (Not that Marvin, or Watts, for that matter, accepted the notion of individual survival through incarnations. Instead a person's properties or patterns would eventually reappear, approximately, in new combinations in fresh sets of people.)

I remember once we were walking near some shops on a spring day in Cambridge and we came upon an infant in a stroller. Marvin starting talking about "it" as if the baby were a device, a gadget, and I completely knew he was doing so to get a rise out of me. "It's able to track objects in the visual field, but with limited interaction capabilities; it has not yet built up a corpus of observed behavioral properties to correlate with visual stimuli." Oh, that sly smile. He guessed I was the one who'd get all huffy and thus prove that I was the slave to my ideas. Marvin's warmth shown through so radiantly that the rouse didn't work. We laughed.

Marvin linked humor with wisdom. Humor was his brain's way of noticing a hole to fill, a way to be wiser. I always think of him finding a way to make each moment a little funnier, a little wiser, a little warmer, a little kinder. He never failed at that, so far as I ever saw.

What was obvious for even those like myself who didn’t know Marvin well was his astounding capacity for delight and surprise. He had such an original way of seeing our possible futures. In one talk I heard him give, he attributed all the problems facing humanity to an excess of biomass. He proposed three ways technology could help: use genetic engineering to enable a child to have many parents; several people could share one body and keep their minds uploaded on the network when it wasn’t their turn to be embodied or, use genetic engineering and quantum computing to shrink people to the size of mice. The artist Vic Muniz, himself a master of surprise, responded instantly from the back of the room: “Marvin, you better shrink all the other animals first.”

Isn’t it delightfully ironic that Marvin, who had such a wonderfully playful intelligence, was a proponent of artificial intelligence? I would like to propose a second Turing test—the Minsky test—in which a machine would have to convince the respondent that it was an embodiment of Marvin.

About 30 years ago when I was a young graduate student, John Cocke (an IBM fellow and influential voice in computer architecture) took me to meet Minksky at his home. We walked in and it was the first time I’d encountered a mind that was fascinated by, well, everything. Minsky’s passions and preoccupations were seemingly scattered high and low throughout the house—musical instruments here, gadgets there, demonstrations over here. Perhaps the decades have exaggerated my memory of the scene, and no doubt those who knew him far better than I could paint a more accurate picture, but I left the visit with a sense that I’d met someone rare and deep.

I first met Marvin Minsky not as a participant in conversations with him but as an observer of the conversations he had with his longtime colleague and collaborator Seymour Papert. It was 1977. Seymour Papert and I had rented a bungalow on Plum Island. Marvin would visit, sometimes for a day, sometimes for two or three, sometimes for a few hours. It depended on how the conversation went. Marvin’s wife, the brilliant, wonderful, Gloria Rudisch, would sometimes come along.

Our bungalow had one large work room. A table was pushed up against a bank of windows that looked out to a porch that faced the sea. I’m allergic to the sun so during the work day I stayed inside, at that table. Seymour and Marvin usually worked by talking on the porch outside my window. I had a book deadline to meet, but when Marvin visited, I found it hard to not eavesdrop.

They were talking about an idea they called the “society of mind”—an idea for how the mind might do very smart things while being made up of agents that were quite dumb and had no idea what the mind as a whole was up to. It was the interaction of the agents that made the difference. Emergence: It was a way to think about mind as a machine that didn’t require the invention of a brilliant machine, only the collaboration of many simple-minded ones.

The idea was compelling. But equally compelling was how Marvin and Seymour worked.

Here is how I remember it: Marvin would declare an idea as though it were a fact.

But the tone was deceptive. The declaration was really an invitation to have the idea talked through. Otherwise put, it was ready for Seymour. So Seymour would start to talk and then Marvin would pick it up again.

After a while, Seymour would cook and everyone would eat. And then, the improved idea would come out again for more polishing. Finally, there came a moment, as though by the mutual decision of the two friends, when the carefully nurtured idea was up for deconstruction. Then, together, Seymour and Marvin would cart away the weak parts. Or discard the whole. Or salvage a small bit of writing for another bit of theory-making.

It was a way of working that struck me as exactly the opposite of the kind of collaboration I was most familiar with from graduate school. There, you tried out an idea on a colleague and the colleague’s role was to tear it down—the idea being that this criticism would lead you to make the idea better. I certainly learned a lot in graduate school, but that summer, listening to Marvin and Seymour nurture ideas in conversation, I know I learned how to think about thinking.

Marvin loved music and it was a central part of his life. He grew up as a piano prodigy, but especially as a “thinker” about music. He understood music’s great pull on our emotions, and this made him curious, even suspicious. (He once asked me: “What right does Schumann have to order us around?”)

Why do we like music?

Why do we spend so much time listening to it?

Should we be feeling the music or thinking about it, or are these the same thing?

What actually goes on in our minds when we listen to a piece of music?

These sound like obvious questions, but you won’t find answers to them in any traditional book of music history or theory. Marvin is the only person I know who seriously asked these questions, and he spent much of his life trying to answer them.

What process did he use to do that? Marvin always said that the world was divided into two kinds of people, composers and improvisers. He was definitely an improviser. This showed in the very unconventional way he presented lectures and public discussions, never prepared in any conventional way, slides and transparencies seemingly placed in random and surprising order, but always brilliantly spontaneous and coherently profound.

And this was one of the great pleasures of being around Marvin. He lived to think, did much of that thinking out loud, and a lot of that thinking was around music. He’d sit at a piano—any piano, anywhere, always best if there were many other things going on in the same room—and improvise, talking while he played. He often started in the style of Bach or Beethoven, but branched out in ways that were purely Minsky-esque. Unlike anyone else I’ve ever known, Marvin was able to think and talk about his improvisations while playing. He’d say out loud:

Well, this chord could go here or here, why not there?

Boy, I’m stuck; maybe this inner melody needs to lead the way out.

Gee, there are a lot of different layers going on now; what would Bach have done.

And he’d do all of that without skipping a beat, literally.

Marvin improvised in similarly creative fashion when thinking about why music matters and what it means. Standing around the piano in our research area at the MIT Media Lab, Marvin would ask:

Does music use more different parts of our brain than any other activity?

Could one of the functions of music be to coordinate these different parts of the brain?

Do music’s harmony and counterpoint allow us to experiment with entertaining multiple thoughts at the same time?

Does music allow us to practice modulating our emotions on an intensified, compressed time scale?

Does music allow us to practice thinking itself—comparing, evaluating, following, deciding—while liberating us from the real world consequences of being wrong?

Is music in fact the best medium for showing us that emotion is just another form of thinking?

Being around Marvin meant being invited into this playful world of audacious exploration, sharing his ego-less passion to understand, witnessing his courage at realizing that with each answer there would only be more questions.

Marvin loved music, but he sure did take it seriously. And although he was the gentlest and kindest of men, certain things did make him angry. Wasting time without thinking was especially unforgivable. As he once said to me: “Thinking about the mind might sensitize people to how silly it is to spend a day thinking nothing at all”), and listening to insubstantial music was one of the paramount sins (except perhaps for sports, which he truly could not understand why anyone would play or watch).

Marvin looked for music that made him think, that connected with bigger ideas about life, and that opened up new horizons. He could find this by going back to Bach, Beethoven and Brahms, but also in John Cage and Frederic Rzewski.

Generally he did not find this in popular, commercial music, although there were exceptions such as The Beatles and U2. He got to know and admire Bono and The Edge and attended many of their concerts, always intrigued with how—as Bono had told him—the group carefully linked songs through common tempi and rhythmic units, to create continuity and to build tension over an entire live show.

One of my most powerful memories of Marvin is from a birthday party of mine that Marvin attended two years ago when his health was just starting to fail. It was one of those big birthdays, so our 18^th century wooden barn just outside of Boston was filled with friends and family, and there was much hubbub.

About halfway through the gathering, Marvin asked if he could play me something and I said of course; the whole group crammed into the smaller part of my barn, where my keyboard is located. Unlike his usual improvisations, which often took place in parallel with multiple other activities, this one silenced the room and kept us riveted for 20 minutes or so.

Marvin started out in a musical world that conjured Beethoven’s late piano sonatas, especially the slow, fugal movements, then built harmonically and motivically into something original, powerful, and surprising at every turn.

When Marvin finished, we all knew that we had heard something very special, full of joyful, inventive spontaneity but also—I suspect—having been carefully planned by Marvin ahead of time to produce the coherent, probing, moving result.Improvisation and composition, emotion and thinking, perfectly united. Beethoven could not have done it better.