Sebastian Shaumyan

From Wikipedia, the free encyclopedia

Theoretical Linguistics/Semiotics
Sebastian Shaumyan
Name	Sebastian Konstantinovich Shaumyan
Birth	February 27, 1916, Tbilisi, Georgia
Death	January 21, 2007, New Haven, Connecticut, United States
School/tradition	Structural Linguistics, Semiotics
Main interests	Theoretical Linguistics, Semiotics, Philosophy of Science
Influenced by	Ferdinand de Saussure, Georg Hegel

Sebastian Konstantinovich Shaumyan (February 27, 1916 - January 21, 2007) was a theoretician of linguistics and a passionate partisan of a structuralist analysis. He was born in Tbilisi, the polyglot capital of a soon-to-be independent Georgia, on February 14 (although the shift to the Gregorian calendar a couple of years later made his birthday February 27), 1916. A sickly child, he was mostly tutored at home until he took a course in chemistry at a vocational school.

Having learnt German and English in addition to his Armenian, Georgian and Russian, Shaumyan took his degree in philology at Tbilisi University. At some time in the late 1930s he came across Ferdinand de Saussure's Course in General Linguistics (1916) and, captivated, knew his academic course was set.

The Second World War briefly interrupted his scholarly aspirations, as he became embroiled in the battles for twice Nazi-occupied Kerch. He applied for a front-line posting, but instead he was sent to the Main Intelligence Unit in Moscow (GRU), where he was permitted to pursue his studies. He was a Party member and, with a post at Moscow State University, used his position to help, and sometimes to shelter, those who might be accused of the various crimes of formalism or idealism.

Shaumyan published Structural Linguistics in 1965 and founded the Section of Structural Linguistics at the Institute of Russian Language in Moscow, where he co-wrote many works with Polina Arkadievana Soboleva. He promoted the work of Roman Jakobson and Nikolai Trubetzkoy, both of whom were out of favour (one an émigré, the other a prince). He also defended the "formalist", Noam Chomsky, (whom later he vigorously assailed) in Fundamentals of the Generative Grammar of Russian (1958), and Applicational Generative Model and Transformational Calculus of Russian (1963), both written with Soboleva.

In 1968 Shaumyan spent a year in Edinburgh and in 1975 was able to join the wave of Jewish emigration permitted at that time, joining Yale’s faculty of linguistics.

As part of the procedure for granting tenure, the department solicited opinions about Professor Shaumyan’s strengths and his standing from a large number of academic linguists of varying ideologies around the world. "Brilliant world-famous linguist deserves tenure," came an admirably unambiguous cable from Belgrade. Others called his work "the cornerstone of modern linguistics" and him "one of the supreme masters of his subject".

Generous praise came from the universities of Eastern Europe. For some Cold War scorekeepers, "Russia's loss is America's gain". From MIT itself, the belly of the beast, there were generous words. Chomsky himself, from whose positions Shaumyan was to deviate ever further, writes that “there should be no question, as far as I can see, with regard to offering him a permanent appointment at the highest level".

Roman Jakobson elegantly praised his "genuine enthusiasm for inspired research and inspiring teaching"; while for Umberto Eco, Shaumyan’s model is the only alternative to Chomsky's.

Shaumyan's theory of applicative grammar was developed, reinforced, and extended in Applicational Grammar as a Semiotic Theory of Natural Language, (1977); in A Semiotic Theory of Language (1987); and finally in Signs, Mind, and Reality (2006, in the series Advances in Consciousness Research), with the intriguing subtitle A theory of language as the folk model of the world.

To his chagrin, he was superannuated by Yale in 1986, but maintained, as Emeritus, a vigorous and very productive retirement. His bibliography contains a dozen books, some two hundred papers, and he was active on the conference circuit. In 2005, approaching 90, he returned to Moscow as a Fulbright scholar (but was refused a visa for a longer stay.)

Shaumyan’s later work is marked by a broad interest in the philosophy of science, in foundational questions of linguistics and in related but separate studies of consciousness theory, and neurolinguistics. It is sharply critical of Chomsky, who Shaumyan saw as being unable to properly delineate what pertains to the study of linguistics proper. The list of languages cited in his last book gives evidence of the breadth and liveliness of his interests; they include Basque, the endangered Australian language of Dyirbal, and the extinct Oregon Indian Takelma.

Shaumyan's inspired revival of Saussure's ideas and his dexterous reintroduction of the "dialectic" method into linguistics gave him considerable persuasive power, and not merely among Chomsky-sceptics.

While some of his ideas at first blush may sound old-fashioned and downright Hegelian, each criticism was astutely and impressively parried in the language of contemporary scientific debate. Shaumyan's gentle manner cloaked a steely determination, and he was widely respected.

[edit] Applicative Universal Grammar (AUG)

The core of Shaumyan's linguistic theory is Applicative Universal Grammar or AUG. The theory was first introduced in his book Strukturnaja lingvistika (Structural Linguistics), published in Moscow in 1965. AUG is based on combinatorial logic and grammatical categories which are built from two primitive universal types, called a term (T) and a sentence (S), which exist in every language. A term represents a noun or a noun phrase: for example "dog", "a dog", "a big dog" would all be considered terms. "A dog runs" would be a complete sentence. The verb "runs" is an operator that acts upon the operand term "a dog" and transforms it into a complete sentence "a dog runs". In Shaumyan's operator notation the verb "runs" would be represented symbolically as OTS. Recently, AUG has been used in computational linguistics in the development of a natural language parsing program, using the programming language Haskell. Natural language parsing has important applications in machine translation of sentences from one language into another.

In the paper entitled Using Types to Parse Natural Language (In Proceedings of Glasgow Functional Programming Workshop, IFIP, Springer Verlag, 1995), Mark P. Jones, Paul Hudak and Shaumyan give a brief introduction to AUG:

To understand the way that AUG works, it is useful to think of words and phrases as atoms and expressions, respectively, in a typed language of combinators. For our simplified version of AUG, there are just two primitive types: T representing terms (for example, nouns such as "friend" and noun phrases such as "my friend"), and S representing complete sentences (such as "my friend runs"). The only non-primitive type is of the form Oxy, denoting phrases that transform phrases of type x to modified phrases of type y; this is the most important concept behind the AUG formalism. For example, the word "my" is treated as having type OTT since it is applied to a term of type T to obtain a modified term, also of type T (every word is pre-assigned one or more types in this way). Thus the construction of the noun phrase "my friend" can be described by an inference:

More generally, we can use the following rule to describe the application of one phrase, p of type Oxy, to another, q of type x:

Clearly, types of the form Oxy correspond to function types, written as (x --> y) in more conventional notation, while the typing rule above is the standard method for typing the application of a function p to an argument value q. The O for function types is used in the descriptions of AUG cited above, and for the most part we will continue to use the same notation here to avoid any confusion with type expressions in Haskell; in our program, the types of natural language phrases are represented by data values, not by Haskell types. Another advantage of the prefix O notation is that it avoids the need for parentheses and allows a more compact notation for types. The results of parsing a complete sentence can be described by a tree structure labelled with the types of the words and phrases that are used in its construction. The following example is produced directly by the program described later from the input string "my friend lives in Boston". The results of parsing a complete sentence can be described by a tree structure labelled with the types of the words and phrases that are used in its construction. The following example is produced directly by the program described later from the input string "my friend lives in Boston".

Notice that, to maintain the original word order, we have allowed both forward and backward application of functions to arguments. The first of these was described by the rule above, while the second is just:

For example, in the tree above, we have used this rule to apply the phrase in Boston to the intransitive verb lives; the function acts as a modifier, turning the action of "living" into the more specific action of "living in Boston". It is sometimes useful to rearrange the trees produced by parsing a phrase so that functions are always written to the left of the arguments to which they are applied. This reveals the applicative structure of a particular phrase and helps us to concentrate on underlying grammatical structure without being distracted by concerns about word order -- which vary considerably from one language to another. Rewriting the parse tree above in this way we obtain:

In situations where the types of subphrases are not required, we can use a flattened, curried form of these trees, such as in Boston lives (my friend), to describe the result of parsing a phrase. The two different ways of arranging a parse tree shown here correspond to the concepts of phenotype and genotype grammar, respectively, in AUG, but will not be discussed in any further detail here. One of the most important tasks in an application of AUG is to assign suitable types to each word in some given lexicon or dictionary. The type T is an obvious choice for simple nouns like "friend" and "Boston" in the example above. Possessive pronouns like "my" can be treated in the same way as adjectives using the type OTT. In a similar way, intransitive verbs, like "lives", can be described by the type OTS transforming a subject term of type T into a sentence phrase of type S. The word "in", with type OTOOTSOTS, in the example above deserves special attention. Motivated by the diagram above, we can think of "in" as a function that combines a place of type T (where?), an action of type OTS (what?), and a subject of type T (who?) to obtain a sentence phrase of type S.