Friday, August 16, 2013

Cryptic Tales from the ISI

In 1985, after three years of teaching and research in the United States, Jackie and I decided to move back to India and try our hand at living there. Jackie (a.k.a. Devaki Bhaya), a plant molecular biologist, was the prime architect of this move. I had wonderful colleagues at Stanford, and I was more selfish about my research, so I wasn't as convinced. Nor was my then-employer. ("You can't do this to us...we haven't had the chance to deny you tenure yet!"). But I'm glad Jackie pushed us to move. From being one more boring economic theorist I was transformed into ... well, another boring economic theorist. No, I de-xaggerate: I did get far more interested in real issues, and understood that it's hard to do good economics. But that's another story.

I joined as an Associate Professor at the Indian Statistical Institute in New Delhi. The ISI, but not to be confused with its homologue in a neighboring country! At that time the ISI didn't offer an Economics degree though we did start a masters' program in economics later, which is still going strong. 

Our arrival was slightly bizarre. The ISI had apartments on campus, but there was a temporary shortage while new apartments were being built, so we were housed in three adjacent rooms in the Guest House that were transformed into a makeshift apartment by some rare, farsighted magic of interconnecting doors. (See photograph of guest house on the right; we lived right below the arrow.) We stayed there quite happily for some years. Other, similarly hired colleagues were likewise given apartments at the Guest House, the number of rooms depending on one's conjugal status.

Maybe the passage of some thirty years has favorably colored my look backwards, but it seems astonishing to me how comfortably we lived with so little. Our salary was around USD 200 per month. You will say, ah, but the prices were also lower by US standards, and indeed they were. But not for stuff that was priced on world markets: international travel, submission fees to journals, scotch, and even several domestic amenities such as television and cars. We could afford none of these (so we drank what you see to your right and left, and indeed, we drank it right and left). We enjoyed the occasional trip that was paid for international hosts (but only with a delay, so friends such as Rajiv Vohra were constantly lending me money, largely paid back I think). Because a paper submitted to the American Economic Review would cost close to a quarter of our monthly salary, we were shut out of publishing in some journals, and so tried Econometrica or the Review of Economic Studies instead. Naturally, these august outlets were not exactly tripping over their feet to publish us, so it was hard going. But it was liberating. Get thee to a nunnery, tenure!

The ISI had a staff of "technical typists" who would use the same electric typewriters as we still had at Stanford (just about beginning to be phased out, I suppose). This was just before TeX and the PC arrived. My typist was a venerable gentleman named Meher Lal. Meher Lal-ji was a jat farmer. He wore white: a dhoti and a kurta. He had grey hair and an enormous mustache. He did not look like a technical typist, or like any typist I had ever seen at Stanford or anywhere else. When I first saw him, my heart sank with stereotypical predictability. It bobbed back up again when Meher Lal-ji asked of my handwritten script: "arré Doctor, is this an epsilon or an 'element-of' symbol?" For those of you who don't work with mathematical symbols, this story will mean nothing, but for those of you who do, you will see that Meher Lal-ji was no ordinary man.

Both Meher Lal-ji and V. P. Sharma switched to TeX with great alacrity when it appeared. It took them very little time to make the switch. I don't think Yevtushenko's great love poem was dedicated to TeX, but hey, why not: 

When your face appeared
over my crumpled life
at first I understood only the poverty of what I had.

TeX was a revelation, a revolution, a reverberation through the Indian Statistical Institute, a reminder to this day of the poverty of commercial alternatives such as Microsoft Word. We clung to Knuth like drowning souls. A typesetting program of the greatest beauty, delivered free, optimized for the dot matrix printers of our crumbly Institute: we all felt we were publishing, and that joy of typesetting entered me and never left, culminating in the typesetting of my own book on coalition formation, from start to finish, many years later. Three of us --- Shubhashis Gangopadhyay, Rajeeva Karandikar and I --- were put in charge of the ISI computer lab. We bought computers for the lab, IBM ATs, followed by 386s, by running TeX on them and seeing how long it would take to typeset a 20-page paper. I can still see Rajeeva holding a stopwatch as we sat at a Wipro outlet.

The ISI also had another rare breed: translators. They translated largely from Russian to English: mathematical papers for an India in the Soviet orbit. I knew one of them, perhaps the Last of the Translators.  I actually used his services; if you don't believe me; check out reference #10 here. He was a character. He told me the story of Kolmogorov's visit to India many years ago, when he received an honorary degree at the Institute. Kolmogorov spoke, and then there was awed silence during the Q&A, upon which my translator proceeded to deliver a homily in Russian. He spoke for a good minute or two. He told me that he said: "Professor Andrei Nikolaevich Kolmogorov, it is because of the presence of individuals such as yourself that I owe my livelihood. Not just mine, but that of my wife and children, who, but for the grace of your genius, would never have had the opportunities..." and so on and so forth, a homage if ever there was one. At the end of which Kolmogorov reportedly said: "Excuse me, but my English isn't very good. Could you repeat please?"

We also had Belgian students in large numbers, all seeking to escape compulsory military service. Among them was the indomitable Jean Drèze, affectionately known in the Institute as nanga sahib for his proclivity to sleep on the ISI lawn with very little on. There were mathematicians, of course (largely of the applied variety), and statisticians, and economists, and there were also sociologists and anthropologists, and we had lots of visitors from all over the world.

One thing we started was an annual conference. With some alterations and improvements, it continues to this day. But in 1986 there was no money to invite anyone. We then realized that the National Science Foundation actually had an allocation of PL480 rupee funds, a stash of nonconvertible rupees which the US government accepted as a payment for wheat dumping and used to fund local consulates. But there were rupees left over, and the irrepressible Dan Newlon at the NSF had got hold of some of them. He handed them over to us, and so funded a set of wonderful annual conferences.

Of course there were frustrating moments as well as happy ones. But apart from the abysymally low pay (which reflected nothing more or less than the poverty of India), life was happy. We felt looked after. There was meritocracy in the system. Young faculty members were put in positions of power. We had some very smart colleagues. Looking back, I cannot complain.

I can tell you many stories about the Indian Statistical Institute, but I'll settle for one last one in this post. There used to be one special seminar every so often at which joint attendance for all the faculty was encouraged: from the anthropologists to the probabilists. One of these seminars was given in 1986 by a Frenchman named Jean-Marc Deshouillers, who had solved Waring's problem for fourth powers. 

You're kidding me. You want to know what Waring's problem for fourth powers is? Oh well, if you insist. It isn't the sort of problem you and I have on an everyday basis, such as having to decide what to cook tonight for the kids. More beautifully, more surreally:

Every positive integer is the sum of at most nineteen fourth powers.

We sat in the audience, quite entranced as Deshouillers told us how he did it. (Actually, this post will tell you that even in 1999, there were top mathematicians who were not aware that the fourth power problem, which had remained open for many years, had been solved. See also the next two messages in that thread.)

After he finished, it was question time. Our resident anthropologist raised his hand:

"What is the practical application of all this?"

I could sense that most of the audience, like me, wanted to dig a deep hole and disappear into the ground. This cool mathematician comes to the ISI, tells us about a great open problem in number theory, and he gets this? Practical application?

But Deshouilliers had been apprised of the great reputation of the Indian Statistical Institute, and took the question very seriously indeed. His excited reply came immediately: "Mais oui, yes, of course!! ... This ... this ideé ...is connected with the application of ze Ardy-Little-Hamanujan's circle methode, and then.." and so on and so forth. By the end of this animated discourse our resident anthropologist was nodding his head in vigorous assent, and a near-disaster had been successfully averted.

Truly interdisciplinary, the Indian Statistical Institute!

Postscript for the true believer who wants to know how Waring got his bound of 19: Try and express 79 as the sum of fourth powers. And now believe that still larger numbers are easier to generate from fourth-power combinations (there are more combinations available). You will then have proved Waring's problem, barring what macroeconomists refer to as "technicalities."