Dependency distance (DD) is a syntactic complexity measure proposed by Hudson (). With the development of annotated corpora, researchers have become able to examine the relation between DD and processing difficulty using large-scale data. Based on treebanks of twenty languages, Liu () identified the universal later called “dependency distance minimization”. While the mean DD of languages tends to be minimized as constrained by working memory, there also seem to be considerable differences among languages, particularly between Chinese and English (Hudson, ). To investigate whether this variation is induced by corpus-based factors (such as sentence length, genre and annotation scheme) or by deeper motivations (such as syntactic differences of languages), a series of studies were conducted. Their results indicate that while shorter sentences, informative and spoken texts, as well as annotations based on syntactic functions may contribute to shorter DDs, the DD variation between Chinese and English is more likely the result of different syntactic structures and processing mechanisms (such as the use of words rather than affixes to express tense in Chinese).

This chapter discusses the rise of Noam Chomsky’s formal, syntax centered, transformational grammar (T) and then ‘generative grammar’ (GG), chronicling its start in Syntactic Structures (1957), influenced by Zellig Harris, the Prague school, mathematics, set theory, Carnap and Goodman. In Aspects of the Theory of Syntax (1965, the ‘Standard Theory’), Chomsky defined competence (vs. performance), the creativity and infinitude of language, grammar as a theory of a language, sentence meaning (the Katz-Postal hypothesis), deep vs. surface structure, the formal ‘simplicity’ of the grammar, and Universal Grammar (UG).

After 1965, alternative models, such as Generative Semantics and later Case Grammar arose, as well as Construction Grammar and Cognitive Linguistics, while Chomsky developed Extended Standard Theory (EST), leading to Government-Binding (GB), Principles-and-Parameters (P&P), modularity, and in the 1990s the Minimalist Program (MP), with ‘minimalist’ syntax but ‘maximal’ UG.

Alternative approaches were varied, including Relational Grammar (RG), and also ‘superlexicalism,' with narrowly conceived T’s, constraint-based architecture, model-theoretic semantics, e.g., CF-PSG (Context-free Phrase Structure Grammar), Head-Driven PSG, and LFG (Lexical Functional Grammar), with no Ts -- and many others.

While few followed the MP in 2000 and after, Chomsky’s influence was incontestable worldwide, especially for those who take a formal approach to grammar.

Dependency syntax holds that the goal of syntactic analysis is to establish all the binary relations between words in a sentence. This process is closely related to working memory. In a sentence, the words between two syntactically related words make for dependency distance, which is an index of sentence comprehension difficulty because, when the two syntactically related words are combined to each other in working memory, the first of them may suffer from memory decay or memory interference caused by the intervening words. Thus, working memory and the least effort principle may dictate a universal tendency for syntactic structures to be organized in such ways as to reduce dependency distance. This tendency has great shaping effect on the patterns of word order in human languages and is potentially able to account for many linguistic universals in language typology. As for the sporadic long dependency structures motivated by communicative needs, specific syntactic patterns may have evolved, utilizing some cognitive mechanisms to lessen the memory load of these long dependencies. Syntactic structures, therefore, are probably the result of self-adaption of a language system to certain external human constraints and motivations, among which working memory is a very important one.