On Psychological Reality Of Phrase-Structure Grammar And Dependency Grammar

Ever since the 1950s, Chomsky's generative grammar shifted the focus of linguistic research onto the systems of linguistic knowledge possessed by individuals. Linguistic studies are concerned with human mind and language is considered as a property of the individual. A very basic fact about language is that we can produce and understand unlimited numbers of sentences without any effort even though we have never encountered them in our previous linguistic experience. So we do have something in our mind– some knowledge about the language we speak; and the objectives of grammar are to display and explain that knowledge.Within the framework of phrase-structure grammar,"immediate constituency"or"phrase structure"representation is employed in syntax. It has dominated the linguistic field since it was first formulated by Bloomfield in the 1930s and then strongly promoted by the Structural School in the following decades. It is the sole syntactic representation ever seriously discussed in the work of Chomsky and his followers. Dependency grammar, on the other hand, exploits dependency (i.e. the relationship between a word and its dependent) as the syntactic representation. It has been the principal means for traditional syntacticians for centuries, especially in Europe, but it has been overshadowed by the predominance of phrase-structure syntax and become unknown to many linguists.This dissertation aims to compare the psychological reality of phrase-structure grammar and dependency grammar. First, the issue of psychological reality of grammars is addressed and it is argued that to claim a grammar is psychologically real is simply to claim it is true of the speaker's knowledge of language. Then the syntactic representation of phrase-structure grammar and dependency grammar is discussed in detail. Based on that discussion, the similarities and differences of phrase structure and dependency structure are presented. It is pointed out that dependency grammar is a better model representing sentences in our minds than phrase-structure grammar in the following ways: phrase structure in itself lacks the information about the asymmetrical relation between the head and its modifiers which is available in dependency structure; phrase structure provides extra nodes, the necessity of which is doubtful, for the syntactic features on the head now have to be just the same as those on the phrasal node and the latter cannot be used to carry extra features; phrase structure cannot provide the kind of analysis like modifier sharing as dependency structure. Further evidence from the following perspectives also proves that dependency grammar is psychologically real: dependency distance is a more natural way to measure processing load than any measure based on phrases; consistent direction (head-initial or -final) is cognitively simpler but harder to process, so mixed directions are functionally motivated; grammatical functions are sub-types of dependency, and require the same kinds of cognitive machinery as non-linguistic relations such as kinship relations; dependency types are prototypes which combine observable and unobservable features in bundles which allow exceptions; dependencies are added one at a time to the head word, rather than by the addition of extra 'mother' nodes; dependencies are stored as relations between lexical items; and every dependency can be learned by induction from adjacent word pairs, and its properties can be elaborated through experience. It is noteworthy that this psychological reality of dependency grammar may explain why it is a comparatively suitable means for natural language processing tasks such as authentic text processing in corpus and machine translation, etc.