Lecture 1 Introduction: Linguistic Theory and Theories

Course Outline Concentrating mainly on syntax But some of the discussion will be of relevance to other areas of linguistics We will look at phenomena and theories which were designed to them

Course Outline The order of presentation will be historically oriented: To understand current ideas, we need to know where they came from You will not end up an expert in any particular theory – but you will know something about all of them

Why do we need theory? Can’t we just observe language and describe it without bothering with difficult things like theories?

No matter what you do, there will always be a theory behind it The human mind always operates with some notion of the way the universe works this may not always be correct and it can be modified The human mind finds it easier to understand things if we break it down into smaller pieces But before we know something, how do we know where to place the divisions? We don’t – we just guess Guess = theory

We don’t have direct access to what we are studying Most theories of linguistics take a cognitive perspective: Language exists in people’s heads Language is ‘rules’ that we know: competence Linguistic knowledge is inaccessible to introspection and unobservable directly in others So how do we study stuff that we cannot observe?

Language is a ‘black box’ puzzle: You can observe what goes in and out of the box, but you can’t observe what goes on inside which changes the input to the output By studying the input and output we can imagine what must be going on inside Imagined processes = theory

A selective History of Linguistic Theory American structuralism (C1900 – 1960) Transformational Grammar (1957 – 1964) Standard Generative Theory (1964 – 1980) Government and Binding Theory Generalised Phrase Structure Grammar Lexical Functional Grammar (1980 – present) Minmalist Programme Optimality Theory (1993 – present)

American Structuralism Two characteristics: Very impressed by Behaviourism/Empiricism Studied Native American Languages

Behaviourism Strong empiricist view: Therefore: All knowledge comes from experience Therefore: Human behaviour is learned from the environment We can only study what we can observe We cannot observe the mind, so we must assume that it does not exist We can observe the environment and behaviour, so let’s assume this is all that exists Behaviourism was the only school of psychology to define away its subject matter

Structuralists and Empiricism We can only study what we can observe We can observe sound So we can study phonetics If we assume that phonology is based on phonetics, we can study phonology too If we assume that morphology is based on phonology, we can study morphology If we assume that syntax is based on morphology, we can study syntax We can’t study semantics – leave that to philosophers

Therefore: The unit of phonology (phoneme) is a collection of phones (observable) and distinguished in terms of the distribution of phones The unit of morphology (morpheme) is a combination of phonemes and distinguished in terms of distribution The unit of syntax (word) is a combination of morphemes, distinguished by distribution

Strucuralists and Native Americans There were not many native Americans left by 1900 and their languages were disappearing fast There was therefore and urgent need to record their grammars American languages were very different from Indo-European languages and it was thought that classical linguistic concepts did not apply

So what was needed was a set of ‘field tools’ that could be applied to any language which would accurately yield grammars As all linguistic units were identified by distribution patterns, distribution was the most important ‘discovery procedures’ to produce a grammar

Generative Linguistics 1952 Chomsky started to develop another view of language 1957 published first book – very influential But didn’t openly criticise Structuralism 1959 published a critical paper of Behaviourism Caused the collapse of Behaviourism Structuralism soon followed

Basic Ideas Rationalist approach The mind exists and can be studied Some knowledge comes from the mind itself Discovery procedures are foolish and limiting You get your data from wherever you can find it (not limited to observable data – i.e. Intuition also acceptable)

The notion of structure stays Words group into phrases Phrases group into sentences New type of rule for producing structure S  NP VP A set of such rules makes a Phrase Structure Grammar Grammar is a set of rules that are part of the mind

But phrase structure rules are not enough to describe human languages Discontinuous constituents [A man with blue eyes] walked into a shop [A man] walked into a shop [with blue eyes] To describe this phenomena we need transformations Rules which alter structures to form other structures (e.g. By moving things about)

Generative grammar and language acquisition From a rationalist perspective, knowledge of language could be innate But clearly this is not totally true as there is more than one language Chomsky supposed that some aspects of language are innate but some are learned

Language learning must be easy Children do it in about 5 years with no formal instruction But transformational grammars which are capable of describing even basic linguistic facts were very complicated How can we square these facts?

Developments of the 1970s The way to account for language acquisition is to assume that grammars are limited You can’t just suppose any kind of rule Therefore the focus was not on what we can do with transformation, but what we can’t Constraints

Developments of the 1980s Principles and Parameters Principle Universal Invariant Innate Parameter A limited number of choices Needs linguistic data to set

Modularity Principles and parameters seemed to group into a small number of components, each of which addresses specific linguistic phenomena Thus the grammatical system was seen as constructed from a set of simple modules which interact in complex ways Thus the grammar is simple, but capable of providing complex analyses

Developments of the 1990s Competition non-absolute grammaticality Grammaticality based on the ‘best’ of several options E.g. It is better to not move than to move It is better to move short distances than long ones

What does the grammar manipulate? From structuralism we have always assumed that words are the basis of syntax But transformational analyses started to discover that units smaller than words undergo syntactic processes: He is always sad He always phones his mother Some have suggested that this leads to a theory where syntax always manipulates items smaller than the word and that words are constructed by syntax