Generative Grammar and Modeling Syntax

In chapters one through three, we are introduced to syntax – the liaison between morphology and semantic meaning – parts of speech, and the hierarchical structure used to represent the syntax of a phrase or sentence. The author begins with an introduction to syntactical rules, predictions about the grammatical correctness of any given sentence, from which she launches into a higher level discussion about the foundations of the theory surrounding syntax: the origin of these rules, the merits of judgment driven science, learning versus acquisition, and Language as an innate instinct (Steven Pinker’s The Language Instinct). She follows the theory of generative grammar, which aims to model the generative processes by which our minds produce construct grammatical sequences, presuming that there exists some sort of Universal Grammar, some sort of biologically inherent “mentalese.” Her most interesting claim comes as a justification for variation across languages, which she explains by suggesting that the human brain is programmed to be able to process the language resulting from a number of different parameters (e.g. topic vs subject, SOV vs SVO, etc.).  The existence of a Universal Grammar, it seems, hinges upon the argument that Language is not learned but acquired, and so it is fitting that a grammar (a collection of syntactical rules) which aims to explain how children acquire language is ideal.

She also introduces the variation seen over words through part of speech, an attribute which determines how any one word might be used syntactically. She makes a point to note that part of speech is not determined by the meaning of a word, but by the word’s distribution – which affixes attach to the word as well as which words systematically precede or follow the word. Her presentation of part of speech reads like a second syntactical layer, still very distinct from semantic meaning, but which edges closer toward building and diagramming grammatically sound speech segments.

I have actually read Steven Pinker’s The Language Instinct (though I may remember very little), so what I found most interesting was when the author briefly discussed the child’s ability to distinguish a grammatically correct speech segment with a finite resource from which to draw. In particular, I was fond of her choice of example, the question word “who.” I can’t quite remember, but the idea of an implied subject or object which takes root in the middle of a sentence while being introduced by some sort of pronoun seems like a fascinating problem in computational linguistics, and one which I would love to learn more about.