Ekaterina Lyutikova on the Architecture of Grammar

Nov 19, 2015 by

[Note to Readers: the following is my translation from Russian of Ekaterina Lyutikova’s video lecture posted on PostNauka.ru, posted here with her permission. Some Russian examples have been replaced by corresponding English examples.]


lyutikovaCurrently, linguistic theory assumes that a model of natural language includes several autonomous, independent units, such as phonetics, morphology, syntax, discourse analysis, linguistic semantics. The question arises: how are these units related and how do they interact? What is the architecture of grammar, the architecture of the models of natural language? There are two types of answers to this question. One type of response are the so-called multi-level models of language. The second type is the models of language, in which syntax occupies the central position. Let’s discuss them.

What is multi-level models of language? They are based on the principle “from simple to complex”, meaning they are based on increasing complexity of the basic unit of the appropriate level. We start from phonetics, which deals with the sound side of language, then proceed to morphology, which is the first level at which a two-sided linguistic unit—a morpheme—arises. Morphology builds words from morphemes. Then we turn to syntax, where phrases and sentences are built from words. And further sentences are combined into larger linguistic units at the level of discourse. But semantics is present at all levels. It interprets the appropriate units, because morphemes, words, phrases, and sentences all have meanings, as does the discourse.

Another type of model is a model with a central position of the syntax. The reason for this central positioning of syntax is the fact that it is there for the first time that recursive rules emerge. In 1957, a model with syntax in a central position was proposed by American scholar Noam Chomsky, and the last 50 years is the period when syntax develops the most rapidly among all areas of linguistics. Of course, such a model was proposed by Chomsky not because Chomsky was very fond of the syntax and decided that this is the main area of ​​linguistics, but he was lead to this model by his logical reasoning about the structure of language. Let’s try to follow Chomsky’s arguments.

Chomsky, for the first time, posed Plato’s problem in relation to natural language. What is Plato’s problem? It is the problem of inconsistency between our knowledge and our experience: we can not know as much as we do know, based on the experience that we have had. With regard to the problem of natural language, Plato’s problem is the problem of the gap between our linguistic knowledge, our capacity for language and the linguistic experience that we get in the process of language acquisition.

On the one hand, we are able to produce and understand an infinite number of linguistic expressions, including those that we have never heard before. On the other hand, it is clear that the experience of a child acquiring language, however great this experience may be, is finite. What solution does Chomsky offers to this dilemma? On the one hand, it is a particular model of language acquisition. As Plato suggests that throughout our experience we develop knowledge of things on the basis of in-born ideal ideas, so language acquisition takes place by applying the child’s linguistic experience to an innate universal grammar, and this dilemma has very specific consequences for the architecture of the model of language. What consequences?

Knowledge of the language is not the ability to perceive and produce a certain set of pre-set sentences, albeit a very large one. Grammatical sentences of a natural language cannot be listed.

Knowledge of the language is knowing a procedure that allows one to produce a potentially infinite number of linguistic products, even if one has a finite set of units and a finite set of rules.

Let’s try to imagine where this infinity comes from. What is the source of this infinity? After all, in principle, if we have a finite number of source units with a finite number of rules, the final product can also be finite. Imagine a grammar that says: if we combine a noun with an adjective, we get a grammatical linguistic expression, and if we combine a verb with a noun, we also get a grammatical linguistic expression, a sentence. What would such a grammar produce using the Russian dictionary? It would produce such expressions as mal’chik bežit (lit. ‘boy runs’), vesëlyj mal’chik bežit (lit. ‘cheerful boy runs’), krasivaja devočka guljaet (lit. ‘beautiful girl walks’), krasivyj mal’chik bežit (lit. ‘handsome boy runs’, and so on. It is clear that there would be more of these linguistic products than there are source lexical units, but nonetheless there would be a finite quantity. So where does the infinity come from? What rules of our grammar take it to a new level?

These rules are recursive. What is a recursive rule? These are rules or chains of rules where the end product of an application of the rule can serve as an input for the subsequent implementation of this rule. A simple recursive rule, for example, in the Russian language is the formation of phrases from two nouns. We can take the noun syn ‘son’ and učitel’ ‘teacher’ and form the phrase syn učitelja ‘son of a teacher’. Then we can take this phrase and feed it back as the input of the rule, to obtain, for example, the phrase drug syna učitelja ‘friend of the son of a teacher’. Next – otec druga syna učitelja ‘father of the friend of the son of a teacher’, and so on. It is clear that this procedure is in principle infinite.

It must be noted that native speakers consciously or subconsciously use this infinite nature of this procedure in different language games. Recall, for example, the English comic poem “The House That Jack Built”, which we know in translation by Marshak, or, for instance, Arkady Raykin’s monologue “Certificate”: “Give me a certificate that you need a certificate that I need a certificate”, and so on. These are endless linguistic products. Thus, if the capacity for infinite complexity is a basic, fundamental property of language structures (of course, in morphology too there are recursive rules, but in syntax recursivity is the key property), it turns out that it is in syntax that the creative character of language emerges, so syntax is put at the forefront.

Syntax produces an infinite number of linguistic products. Other language modules perform in relation to the syntactic representations the so-called interpretative function. They either help create a phonetic form of a linguistic expression, as do morphology and phonetics, or provide interpretation of the sentence, as semantics does.

Why do morphology and semantics need access to the syntactic representation?

Let’s start with morphology. We know that words of different parts of speech are characterized by different grammatical categories. For example, in Russian verbs are characterized by tense, mood, voice, person, number, and in the forms of the past tense by gender. Nouns and adjectives have forms for number, case, gender and so on.

Let’s discuss in more detail how the values of these grammatical categories emerge. There are cases when the values ​​of the respective grammatical categories are recorded in the dictionary: kniga ‘book’ is feminine, žurnal ‘magazine’ is masculine, and izdanie ‘publication’ is neuter. These are so-called interpretable [or constant] grammatical features. But there are variable grammatical features. How does a speaker choose the value of a variable grammatical feature? Some values ​​he chooses in accordance with what he wants to say, according to the semantics. For example, if he wants to talk about a book or books in general, he will choose the singular form, and if he wants to talk about a plurality of books, he chooses the plural form. But there are cases where the choice of the value ​​of a grammatical category of a given word is determined by its neighbor (which may be a close neighbor, or a very distant neighbor) in the syntactic structure of the sentence. Therefore, in order to determine in what form the adjective interesnyj ‘interesting’ should appear, we need to know what it modifies: is it interesnyj žurnal ‘an interesting magazine’, or interesnaya kniga ‘an interesting book’, or interesnoe izdanie ‘an interesting publication’?

Thus, morphology implements, in particular, syntactically defined features of words, so it only interprets the syntax representation of a sentence.

Let us now turn to the semantics. Is access to the syntactic structure of a sentence needed for a correct interpretation? Of course, yes. And we see it very well when there is semantic ambiguity caused by differences in the syntactic structure. Great Russian syntactician Alexander Peshkovsky gives the following example of an ambiguous sentence: Veli emu pomočliterally ‘Tell him to-help’. It can mean two things: either ‘tell him that he helped (someone)’ or ‘order that (someone) helped him’. The differences in the meanings ​​are tied to the syntactic structure surrounding the word emu ‘him’: does it depend on veli ‘tell/order’ or pomoč‘help’. This principle in the relationship of semantics and syntax, the principle that the interpretation depends on the syntactic representation, is known as the principle of compositionality, or Frege’s principle—named after its creator. It sounds as follows: the meaning of a complex linguistic expression is a function, in the mathematical sense, of the meanings of its parts and how they are combined syntactically. Thus, if we want to know the meaning of a linguistic expression abc, we need to know not only the meanings ​​of a, b and c individually, but also the order in which a, b and c were joined together. In fact, it is easy to see. Imagine the two different meanings of the phrase interesting books and magazines. What has been combined first books and magazines or interesting books? The fateful nature of the comma in Eats(,) shoots, leaves is ultimately determined by the fact that a comma imposes on us a certain syntactic structure of the sentence.

So we see that syntax, on the one hand, is the only component of language providing a creative character of language, and on the other hand, it creates a representation that is necessary for the proper phonetic and semantic representations of sentences. In this context, generative grammar theory believes that syntax is the central, essential component of language. In a sense, syntax for them is language.

Related Posts

Subscribe For Updates

We would love to have you back on Languages Of The World in the future. If you would like to receive updates of our newest posts, feel free to do so using any of your favorite methods below: