The main theoretical approaches to linguistics are generative (derivational and non-derivational), functionalist, and cognitive. All modern theories of language are descriptivist, in that linguists attempt to describe and analyze the structure of a language, be it a formal, informal, or dialectal variety. That is, we aim for descriptive rules or principles of languages, rather than finding rules of "correct" usage to impose on speakers or by which to evaluate them or judge them.
The main theories that deal with language structure can be characterized as follows.
1 Generative linguistics
This paradigm was founded by Noam Chomsky, attempts to find the basic underlying rules that underlie any given language, and all languages. The term "generative" is a mathematical term that means to fully describe, account for, and explain a system (it has nothing to do with generating in a mechanical or mechanistic sense, like a machine cranking out grammar patterns or sentences). Chomsky's theories began with a complex set of transformations in the original Transformational Grammar framework, which have been simplified in later theories - Principles & Parameters, Government & Binding Theory, and the current Minimalist paradigm. These have been reduced to a smaller set of so-called movements, but to make this simplification possible, the syntactic structure has been made more complex by adding abstract categories in the syntactic trees or syntactic structure.
It rests on the following assumptions.
- Generative. Linguists aim to "generate" the grammar and structures of a language. This word is used in a technical, mathematical sense, meaning to analyze, describe, explain, and fully account for something. Thus, the aim is to find the abstract rules that govern a language, and that would describe and account for any possible utterance that a speaker of the language could say or understand. By extension, the aim is to find sets of rules that explain the structure of all the world's languages.
- Descriptivism. We are to find the rules or principles that govern the structure of a language, as native speakers actually use them, be it the grammar of a formal, informal, or dialectal variety of a language. We can investigate languages and grammars by finding out what native speakers would say or understand; as Chomsky said, the native speaker is always right.
- Performance versus competence. A native speaker has psychological knowledge of the rules of his/her language, and what one would or could say or understand in the language -- the native speaker's mental competence. The goal of generative linguistics is mainly to find the rules that comprise speakers' competence. In speaking, speakers might make occasional performance errors, due to other factors (stress, fatigue, memory load, etc.) that are not related to competence.
- Universalism. All languages developed from the same psychological stock, as language evolved in humans when or before humans began spreading over the globe. Thus, all languages make use of the same set of abstract principles and structures.
Thus, the theory consists of the following main components.
- Generative. The goal of linguistics is to find the rules of the grammar of languages that fully account for the structure of those languages.
- Universal grammar. Rules come from a set of universal parameters, which can be set in different ways in different languages. These rules are abstract, and most often speakers are not aware of them. Children are born with an innate repository of the general parameters, and in learning their first language, they deduce how these various parameters work in their language, and thus they subconsciously learn the rules of their first language.
- Child language acquisition. Another goal of linguistics is to understand how children learn their first language, e.g., how they make use of the universal grammar to learn a language naturally.
- This ability is constrained by the critical period, a period of time before puberty when a child must learn a language.
1.1 Constraint-based generative linguistics
Some generativists do not favor the use of transformations, derivations or movements, and in response, non-derivational constrain-based theories have emerged, particularly Optimality Theory in phonology. Instead of movements or derivations, universal constraints are proposed, which are present in all languages, but work differently in various languages. This has also been applied to syntax, e.g., Optimality Syntax.
This approach is an attempt to ground language on pragmatic or usage-based and communication-based principles and categories. It focuses on surface forms rather than abstract structure. In fact, functionalists reject many or all notions of generative theory, and do not concern themselves with deep or abstract grammatical structure, and only focus on the surface forms and the communicative functions of surface forms. It has advantages in application to language pedagogy and the study of pragmatics.
3 Cognitive linguistics
This paradigm began with the insights of Gestalt psychology, and later, schema theory in cognitive psychology. It assumes that language is a cognitive domain that is grounded in cognitive psychology and other mental faculties, and that language must have arisen from cognitive faculties rather than just on its own, and hence, language is crucially connected with cognition - be it natural semantic categories in our world, or human social cognition. It combines these psychological insights with the key insights from the study of metaphor - that metaphor is a key element of meaning in language, and hence, a whole field of cognitive semantics exists for exploring this.
This is the older paradigm in linguistics, which dominated the field in the 20th century until the generative revolution in the 1960s. It focuses on identifying and understanding the various types of structures in a language, such as grammatical and phonological structures. Techniques were developed for collecting data and comparing forms to deduce and analyze these grammatical, phonological, and communicative structures in languages. This was particularly fueled not only by work on known languages, but also by research by anthropologists, who studied less known or previously unknown cultures and languages; they collected data on and analyzed many new languages. Unlike the generativists, they did not focus on deeper, abstract grammatical rules at work.