Translated by A. Shetsen

Sergei Anatolievich Starostin is dead. This was more than a merely famous man loved by many. Much is being written about him now, and more will yet be written, by those who knew him better, more closely, longer, than I did. His name will yet pop up in memoirs; people will reminisce about him, in forewords to scientific articles and on TV shows. He will be remembered as a true friend and as an inspired opponent, as a wonderful conversationalist and as a leader who even in the darkest days of Russian science knew how to infuse his colleagues with the quiet optimism of the consummate professional. His discoveries and his publications will be enumerated. Some will mention his passion for music; some will find solace in knowing that, thanks to him, in many universities around the world Russia is associated with something beyond vaguely anthropomorphic politicians and alcoholics wearing ear-flapped fur hats...

All of that will happen regardless. But I have my own reasons for writing this. Unfortunately, not so many people understand just what S. A.'s work and accomplishments were about. Anyone could see that he was an extraordinary person. But still. Though several days have passed since his death, serious, mature people, people who have seen a lot in their time, continue to exclaim: 'this cannot be! It boggles the mind! I can't believe that it's happened!' Why? Surely they, if anyone, know very well that even the most talented sometimes die very young, sometimes much younger than at 52? How to explain, in terms accessible to the general reading public, why Sergei Starostin's death is a catastrophe to those who knew what he had been doing for the last twenty years? For any serious researcher's work cannot be easily described in a few simple phrases. Now that great linguist can no longer say anything more for himself, a linguistic barrier divides him from the majority. And the professional instinct in me rebels (I used to be a translator), demands that I try to cut through this barrier; try to get something across of the man and his work, even if just to give an idea of its scale.

* * *

This cannot be done in one or two sentences. Forgive me, it will be a bit long. What can I say in a few words? Capture one shining detail, perhaps. We can talk, for example, of how the student -- on a dare -- once mastered in one night the dictionary of an exotic and complicated language: Gilyak, if that matters. So what? Like that guy who was dared to make bow-knots with... Or that the man could quote long passages in verse from Sanskrit, Heian Japanese, ancient Chinese, and with the authentic ancient accent to boot. Right, an ordinary Ph.D. (even in language or literature) will ooh and aah... But alas, the full scale of what the man did cannot be gauged from such easily understandable trivia. So we'll speak at some length, and of things that a non-specialist seldom comes across.

* * *

Starostin was a student of historical-comparative linguistics. He became especially renowned as a specialist in long-range comparison. What is that, exactly, and what is required to leave a mark in it?

Let's step back a little. Every literate person today will have encountered the term Indo-European languages, will have heard some echoes of the debates about the Indo-European homeland, mythology, etc. Indo-European studies is the most popular subfield in historical-comparative linguistics. It is a good starting point from which to approach the field of long-range comparison.

It would be nice to know Latin, Greek, Gothic, Sanskrit, Lithuanian and Old Icelandic, and of course some modern languages as well (just to read the literature, after all!) -- but that's just the beginning. Actually, it is necessary to become intimately acquainted with the inner workings of at least twenty languages, ancient and not so ancient, and to be able to navigate your way through the dialects of the major ones: it would not do, after all, to be condemned as an amateur just because you confused an Ionic form with an Attic one, or failed to detect Anglian levelling in an Old English word, as though it were the standard Wessex. Then you have to be able to gauge the enormously variable reliability of the vastly heterogeneous sources and descriptions; to learn to envision the dynamical evolution of a system so complex as language, to distinguish on sight the probable scenario from the improbable, the likely archaism from the result of more recent evolution... And when you have accomplished all that, and cognates in languages not very closely related (such as Sanskrit chakra and English wheel) are as obvious to you as the connexions in your native tongue (say ghost and aghast), only then are you competent to write your own theses, to hold your own opinion of where the Indo-European homeland was and whether or not their culture possessed the notion of freedom.

Indo-European languages are a classical field of study. There historical linguistics developed its methods, and very many comparative linguists there had their start. Starostin, too, left his mark in Indo-European studies. But to understand what he did for the greater part of his life, one point must be clearly made. In order to go a step further, to attempt to discern groups of languages more ancient than the Indo-European family, all of the above does not suffice. Direct comparison of data from various languages simply does not yield good results at the 'long range'.

Any serious linguist knows, after all, that it is meaningless to compare, let us say, Abkhaz and modern Mandarin Chinese: any similarities can be written off a priori as wholesale coincidence. Comparison at such a distance must be performed in stages (a staged 'step-by-step' reconstruction), approximately as follows.

From one side, the 'Chinese', you must start by significantly enhancing our understanding of 7th century A.D. Chinese pronunciation. You do this by analysing the ancient rhyming dictionaries, the modern Chinese dialects, and ancient borrowings from Chinese in the languages of China's neighbours: Japanese, Korean, Vietnamese. (And, by the way, this Middle Chinese sounded as different from any of its modern descendants as Latin does from French.) By burrowing further into still more ancient borrowings and poetic rhymes, and by closely examining the way in which Chinese characters are composed, you extend your reconstruction, step by step, to the language as it was spoken during the Zhou dynasty. With Chinese alone you can go no further, but by correlating with the data from other known languages -- Classical Tibetan, Lushei, Kuki-Chin and others -- you revise and enhance the reconstruction of the Proto-Sino-Tibetan.

Going about it from the 'Caucasian' side, you first reconstruct the original proto-languages of every East Caucasian group: Nakh, Avar-Andic, Dargi, Lezgic... In other words, you must accomplish more or less the same amount of work as all the Indo-European scholars from various countries together required a century and a half to complete. After all, since the ancient languages are missing, the temporal distance between their various modern descendants are great: even languages in the same group can differ from each other more than Sanskrit differs from Greek. (It is true that Troubetzkoi and others worked with this language family before Nikolaev and Starostin: in exile, Troubetzkoi restored from memory the field notes he had lost during the Revolution. But any total revision of this kind of comparison is little different from starting from scratch.) Operating with these intermediate proto-languages, you then reconstruct Proto-East-Caucasian, a language, it seems, somewhat more ancient than Common Indo-European. After that, you repeat the same work with the West Caucasian languages (Abkhaz is one of these): since these have not become so divergent, you can do it in one step, more or less. Correlating the East Caucasian reconstruction with the West Caucasian, you convince yourself that the languages are in fact cognate, and restore their original common form, the Proto-North-Caucasian.

What is left? To note the systematic correspondences between Proto-Sino-Tibetan and Proto-North-Caucasian (which might have remained unnoticed decades longer, had not one man been responsible for much of the preliminary work) -- and to formulate the sensational Sino-Caucasian hypothesis. But, as everyone knows, binary juxtapositions aren't quite comme il faut; so your erudition and breadth of perception must be great enough that you can point out how suspiciously similar the reconstruction looks to the basic vocabulary of the Ket language; squeeze out all you can from the scant materials offered by the Yeniseian family; convince yourself that the data from all three groups (North Caucasian, Yeniseian, and Sino-Tibetan) fit together perfectly; and only then offer to the scientific world not merely a hypothesis, but a fully thought-out and developed comparison. In other words, you must embark on a decades-long argument against innumerable opponents and skeptics, figuring out along the way whether your details can be made more precise by enhancing the comparison with material from Old Burmese or the Bodo-Garo group.

This is, more or less, what is meant by a step-by-step reconstruction -- as performed by a virtuoso, a master of the genre. The comparison is not of immediately known languages, but of intermediate proto-languages, here in three or four layers, each of which subsumes dozens of ordinary languages. But one must be on intimate terms with all of the latter, or be acquainted with them, at the least: for the reliability of this or that detail must ever be judged, and the probability of any given solution crucially hangs on the details... How much effort is demanded of an ordinary adult, or even of an ordinary trained linguist, to become acquainted with just one such ordinary language?

* * *

Had Starostin published nothing but his Reconstruction of the Old Chinese Phonological System, he would have been a scientist of note. His accomplishment would have been fundamental, had he rested upon completing (with S. L. Nikolaev) the reconstruction of Proto-North-Caucasian. If he had had enough after his Sino-Caucasian reconstruction, we should have to speak of him as a scientist of global magnitude. But how are suitable epithets to be found for a man who undertook in succession several projects of this scale?

For the Sino-Caucasian reconstruction was followed by the Altaic (he headed a research group that included A. Dybo, O. Mudrak and several others). Then, in collaboration with I. Peiros, he reconstructed the proto-language of the Austric macrofamily. And, along the way, proposed numerous revisions the reconstructed protolanguage of another macrofamily, Nostratic. After that, Starostin dared to attempt a full classification of all the languages of Eurasia: to be more precise, he practically completed this work. He was able to put together a large international group of researchers to work on this problem. Were any serious linguists before him capable of thinking at such a scale? Only Greenberg, perhaps, with his theory of 'mass comparison'. Which is, today, still torn to shreds for the crudity of its methods -- and altogether undeservedly so, for Greenberg himself understood very well, and wrote, that his methodology was intended for the preliminary classification of a large number of yet poorly described languages, and was never meant even to approach a level of critical precision equal to Starostin's. Starostin's comparisons will also be attacked, probably for as long as comparative linguistics itself is not dead. But the objections to Starostin's work from 'narrow' specialists (those who work, that is, within one language family) are often expressed in a different way. The usual thought goes something like this: 'I know he's made several brilliant contributions in my field, but it's inconceivable that he's capable of the same level of work with all these other languages, and those ones as well... that can't be serious.' Different people attest that such comments have come from specialists in several language families.

Over time, his critics changed their tone. Skepticism gave way to attempts at correction, demands for greater precision, a desire to come to grips with his work; then, to cautious praise. More and more often international authorities cited him sympathetically, or even enthusiastically. It is not always the case that a scholar lives to see himself granted international recognition and fame for work destined to change his peers' perception of where the bounds of the possible, of the scientifically decidable, lie. Starostin almost made it.

Yes, but... He wrote computer programs, too. And taught. Wrote textbooks. Improved the computational algorithms of lexical statistics. And in comparative linguistics, he made a dozen or so 'minor' (worthy of an ordinary doctoral dissertation) discoveries -- like that Yeniseian reconstruction or his proof that the Hurrian and Urartian are North Caucasian. Not to mention... no, I think I'll stop there. It would be a very long list, I know I'd forget something. Really, I just wanted to give a hint of the scale of his work.

I'm not sure whether I've got it across. I'll end with just one small observation. If you've ever opened the covers of a book about comparative linguistics, you may have come across such terms as 'Verner's law', 'De Saussure's law'... But there is no 'Starostin's Law', nor can there be one. The term would be meaningless, for to which of the hundreds of systematic relationships discovered by Starostin would it refer?

* * *

The preceding text is confused and pointillistic. Obituaries cannot be written like that. For some reason hurry overwhelmed me as I wrote, impelled to express it right now, today. This text is devoid of real emotions; there was no room for them; except, by some chance, between the lines. I will not presume to write about Sergei Starostin, the human being. But... the work is the man, in large part he lived indeed for his work.

Comparative linguistics has developed explosively in recent decades. No less than astrophysics and the like, this science has shifted the paradigm of our perception of reality. For if we consider human history as a whole -- not restricting ourselves in space and in time to the history encompassed by written records -- the illumination of this span must fall to the linguists. The materials of archaelogy, after all, of palaeoanthropology, even of genetics, are in themselves mute. But language gives history its truly human expression: culture, religion, the work of the mind. This is what made Starostin indispensable. As he plumbed the previously inaccessible spoken depths of our past, he transformed our mental image of the world, and we felt the universe around us shift.

And so he leaves us the issues of science. Five, seven years at most, and he would have resolved them; but now they are left to a new generation. Oh, they will be resolved, without a doubt: thanks in no small part to Starostin, we now have the strongest school of comparative linguistics in the world. This school will survive, of necessity, the inevitable next wave of meaningless total reform, as it has survived one or two such already. But...

So it is that a heart attack has ended -- an age. Yes, we can see now: a whole age has passed. Never have I seen a death so clearly expose the abyss. Ahead, in the void left by Starostin, lies a different world, to be inhabited and explored by other people. And no death can ever be more strongly compelling: for us to measure our work and our deeds on that scale of eternity to which, while he lived, Sergei Starostin approached more closely than any of us.

06.10.2005