Evolution of Chomsky's Transformational Grammar E-Book

El Mouatamid Ben Rochd

• Herausgeber: Books on Demand
• Kategorie: Geisteswissenschaft
• Sprache: Englisch
• Veröffentlichungsjahr: 2020

Seitenzahl: 147

Beschreibung

In his early plastic age, the child presents an asymmetrical growth in what concerns his motor system, when compared with his cognitive system. The latter seems to be far ahead of the former. It consists of many sub-systems; the most outstanding of which being language. TRANSFORMATIONAL GRAMMAR has been striving for more than half a century to try and deal with this phenomenon. It has evolved from a model of rules to a model of principles, to end up with the Minimalist Program. This book presents the successive approximations that Transformational Grammar has gone through since its early insemination in1957 up to the Minimalist Program. After which, there is a critical chapter called "Occam's Razor". Finally, there is a chapter on "Logical Form" and an appendix that deals with the major transformations of early transformational grammar.

Leseprobe

CONTENTS

Acknowledgment

Introduction

CHOMSKY'S T.G.

1.1. First Version

1.1.1. Phrase Structure Rules

1.1.2. Transformations

1.2. Standard Theory

1.2.1. Phrase Structure Rules & Lexicon

1.2.2. Transformations

1.3. Extended Standard Theory (E.S.T.)

1.3.1. X-bar theory

2

1.3.2. Lexicon

1.3.3. Movements

1.3.4. Constraints

1.4. Government and Binding

1.4.1. GB Principles and Parameters1

1.4.2. Empty Categories

1.4.3. CP/IP Barriers

1.4.4. Cliticization

1.5. Chomsky’s Minimalist Program & Arabic Non-Concatenative Morphology

1.5.1. PHILOSOPHY

1.5.2. MINIMALIST GUIDELINES

1.5.3. MINIMALIST MODEL:

1.5.4. LANGUAGE TYPOLOGY (SVO/VSO)

1.5.5. TENSE OR ASPECT?

1.5.6. NON-CONCATENATIVE MORPHOLOGY

1.5.7. PERFECTIVE TRANSFORMATION

1.5.8. CASE FILTER

1.5.9. 4/10 AGREEMENT

1.5.10. CHOMSKY’GREED

1.5.11. PHONETIC FORM IN ARABIC

1.5.12. LOGICAL FORM

1.5.13. LCA SUPERSEDES X-BAR

OCCAM’s Razor

1

2.1. Parametric theory

2.1.1. X- bar Theory

2.1.2. Move Alpha (Bounding theory)

2.1.3. Case theory

2.1.4. Binding Theory

2.1.5. Control Theory

2.2. Occam’s Razor

2.2.1. DP Hypothesis

2.2.2. Headship

2.3. Case

2.4. Noun Phrase or Small Clause?

5

2.5. QP hypothesis (and demonstratives)

2.6. Bound Variables

Logical Form

3.1. Word Meaning

3.1.1. HISTORICAL PERSPECTIVE

3.1.2. SOCIAL PERSPECTIVE

3.1.3. DISTINCTIVE FEATURES

3.1.4. MORPHOLOGICAL PERSPECTIVE

3.2. Sentence Meaning (LF)

3.2.1. Predicate Calculus

3.2.2. Binding Theory

3.3. Idiomatics

3.4. Contradiction

3.4.1. PREDICATIVES

3.4.2. CONDITIONALS

3.4.3. DISJUNCTIVE

3.4.4. Contradiction

Appendix

: TRANSFORMATIONS

REFERENCES

Acknowledgment

Many thanks are due to my students, relatives, friends and colleagues in particular Dr. Khalid Chaouch, Dr. Ahmed Elheggach, Dr. Amr Sellam, Dr. Mostafa Shoul, Dr. Reitha Ben Rochd, for their invaluable help. Above all our thanks must go to Professor Avram Noam Chomsky, from whom we learnt this most elegant theory of language, which filled an infinity of hours of teaching – it is indeed GENERATIVE & TRANSFORMATIONAL.

“Money takes you nowhere,

Friends take places, indeed!”

Introduction

The material dealt with in linguistics is so vast and complicated. Languages1 are said to be inflectional, concatenative... They belong to families: Indo-European, Semitic.., sub-families: Romance, Germanic..etc. There are old and modern, spoken and written, standard and colloquial versions of every language. In syntax - let alone the levels of linguistics - many problems exist.

The linguistic investigation of sentence structure is, in a way, similar to what happens in arithmetic theory. When a set is discovered to be inadequate to cope with a mathematical item or mathematical class, the mathematicians opt for a new set (that includes the preceding one). The subtraction 2-3 is an impossible operation within the framework of N2 so Z is established to cope with it, but within the latter 50/3 is impossible so Q is created. √2 does not exist in Q so R is set forth. √-2 is not allowed in R so C is created and so on and so forth.

In linguistics, there are many old and new schools. Traditional grammar, the Neo-grammarians, the Tagmemic school (the followers of which believe that linguistic structure can be exhaustively accounted for in terms of SLOTS and FILLERS) and the systemic linguists who work by RANK, SYSTEM and DELICACY.

For Bloomfieldians (Bloomfield et al.) the sentence is the syntactic unit because it shows greater freedom of occurrence in speech. Every sentence is marked by a particular intonation. In writing, it starts with a capital letter and ends with a full stop, a question or an exclamation mark. They deal with the surface structure as a sequence of adjacent items, but do recognize hierarchies of immediate constituents by opposition to Ferdinand de Saussure3 who is a 'dualist'. He distinguishes LANGUE (i.e., language) from PAROLE (i.e., speech).

Chomsky is a still more explicit 'dualist' than Saussure .He distinguishes between COMPETENCE4 and PERFORMANCE as well as between DEEP and SURFACE STRUCTURE. His approach - a computer-like grammar (the word GENERATIVISM5 itself is taken from mathematics) - gives us syntactic models that can predict and recognize, more or less explicitly, the sentences of natural languages.

Some psychologists assume, against Chomsky's point of view, that the deep structure matches what actually exists within the speaker's mind. Furthermore, Chomskian linguists consider the actual utterances as incomplete and non-representative, because they are continually broken by slips of tongue, self-corrections and interruptions. So the actual performance is to be replaced, in the linguistic inquiry, by the competence of an imagined ideal native speaker in an ideal speech environment (cf. Labov's Functional Competence). This process is known as "scientific idealization".

Linguistics is meant to be scientific. For the linguist's attitude is just like that of a physicist, who wants to describe, for instance, the motion of a shell. He first limits himself to four variables - say the length of the cannon, the weight of the shell, the distance of the target at and the angle made by the cannon and earth. Then, if there's and outside interference at the time of the firing like - say wind - he/she, or the engineer, will have to assess the force and the direction of the interference and correct the first equation accordingly.

Likewise in linguistics, we must limit our scope to a small area of language at the beginning. Our method must be as systematic and economic as possible. As for any empirical attempt, we must proceed from a hypothesis towards a conclusion by means of observation and testing. Ultimately, we can build a model of what natural language patterns might be. We must confine our analysis to language structure, not refer to the outside world, and express the results of our discoveries in purely formal terms.

The segmentation of language remains, however, only a matter of trial and error, far more than what happens in Physics. An unknown language can only be identified phonetically and transcribed in an impressionistic fashion. But neither phonetics nor phonology can catch other patterns bigger than the sound system. So the linguist must learn the language he wants to investigate (or, at least, use the services of informants), notice the regular occurrences of items bigger than the sound segment and ask native speakers about the relevance of the samples he may come to describe.

The linguist can also start with his own language (idiolect) and proceed gradually to the investigation of his language family, then to human language (i.e. UNIVERSAL GRAMMAR). He can also proceed from level one of linguistics to the other levels in the same language. He can use the most practical grammar (simple and accurate) available to him, try it and eventually replace it, when he comes to its limitations and discovers a better one.

After (or apart from) Phonology and Morphology, Syntax is proposed to cope with the structure of a given language. It is claimed that syntax is the key the process of learning6 and understanding a language. The sentence is generally considered the upper unit, but syntax deals also with the clause, the phrase and the word. Syntax deals with word-relations and the structure of the sentence in the most systematic way possible.

Years ago, as I was writing at the railway station of York city, I met an English student of history. He tried to tell me about his sister who had gone sightseeing to "Tangiers in Tunisia" (sic!). Then he said that he would be delighted to come and see my country. I told him that he was "most welcome to my home in Tokyo!) We laughed a lot and he concluded saying; "Actually, you're right, I might go to a travel agency, book for a country and end up in another!"

This student of history had a good approximation (knowledge) about the geography of Africa. At least, he knew that Tangiers is a city (and not "something to eat"!). He had a second approximation which told him that it is a city in Africa and not in Asia or Australia. He even knew that it is situated in North Africa (which is a third approximation). Of course, he needed two more approximations: he needed to shift Tangiers to the Moroccan territory and finally, spot it in the north of Morocco (fifth approximation).

Although this student may have seemed silly at first, in fact his approach is perfectly valid. It is actually used by all researchers and working scientists who are "silly", in that sense. They work by successive approximations. Likewise, the linguist works step-by-step towards a system of rules and principles that are part of the native speaker's competence. He formulates a RULE, comes across a counter-example, formulates a new rule and so on and so forth.

In syntax, this approach is applicable. Consider the rule known as QUESTION TAG:

It has a bell, hasn't it?

1st approximation:

a) locate the verb and copy it to the right of the sentence. If the original verb is positive, make the copy negative, and vice versa.

b) locate the subject of the sentence and copy it to the right of the verb.

* The boys are in town, aren't the boys?

2nd approximation:

a) locate the verb and copy it to the right of the sentence

b) if the original verb is positive, make it negative and vice versa

c) insert to the right of the verb the pronoun that agrees with the subject in person, number and gender

* John could have arrived, arrivedn't he?

* John could have arrived, hasn't he?

3rd approximation

a) copy the first verb to the right of the sentence, making the copy negative if the original verb is positive and vice versa.

b) insert to the right of the copied verb the pronoun that corresponds to the subject in person, number and gender. (Akmajian & Heny 1975)

John could have arrived, couldn't he?

In Semantics, the same scenario applies. Consider the BINDING of reflexives:

John1 saw John1 ==> John1 saw himself1

*Mary1 never talks to us about themselves1

Mary1 never talks to us about yourself1

Mary1 never talks to us about himself1

Mary1 never talks to us about herself1

1st approximation:

a reflexive is bound to any expression having the same number, gender, and person (AGREEMENT).

What do your friends think about Fred?

John doesn't like himself very much,

2nd approximation:

a reflexive is bound to a noun phrase in the same sentence

[s John never talks to himself about Fred]

3rd approximation:

a reflexive is bound to a noun phrase in the same clause

John thinks [c that Fred hates himself]

4th approximation:

A reflexive is bound to a noun phrase in the same clause that preceeds it. We say that they are CO-REFERENTIAL. (Radford 1981)

This book aims at presenting the successive approximations7 that TRANSFORMATIONAL GRAMMAR has gone through since its early insemination in 1957 up to the "Minimalist program3 of the late nineties. After which, there is a critical chapter I call "Occam's Razor". Finally there is a chapter on "Logical Form". (For "Phonetics Form" see Ben Rochd 1994). The appendix deals with the major early transformations of the '60s.

1 The amazing organization of the bee colony is obviously based on their no less amazing "language". The latter system that challenged the researchers for a long time, until German Karl Von Frisch of Munich University found the solution to that mystery after spending thirty years of patient (read painful!) research and observation. (Benveniste 1966)

2 N, R.. stand for Natural Numbers, Real Numbers, etc..

3 Last century's Philology and Evolutionist Historicism were brought to a halt when they were faced by the stumbling block of Saussurean Synchronic Structuralism.

4 Thanks to our Linguistic Competence we know that certain pairs of words rhyme, that some phrases mean the opposite of other phrases, that some sentences are ungrammatical.

5 Daring Chomskian rationalism, mentalism, and generativism superseded hard-line Bloomfieldian empiricism, behaviorism and descriptivism respectively. This revolution started in the United States and the rest of the world followed.

6 "The fact about languages... we don't really learn anymore than we learn to have arms. It just grows. You can't help it. You speak to your child in an environment where language is being used and that child has no more choice whether to weight five pounds or something or decide not to eat. It can't decide not to learn that language. Because learning language is just a process that the brain goes through under certain conditions. It is a process of growth not a process... what people call learning. The system grows in certain ways because of the way it is built." (Chomsky 1972)

7 "(there is) a very striking difference between the natural sciences and the humanities. In the natural sciences, it is taken for granted that theories are constantly changing. We do not understand enough, even in the most intensively studied areas, to be at all confident that we know "the truth". In fact, it is taken for granted that we do not. Therefore, if theories are not changing, the field is probably dead or not worth pursuing because it is too trivial and boring." (Chomsky 1995)

1. CHOMSKY'S T.G.

1.1. First Version

1.1.1. Phrase Structure Rules

The Bloomfieldian linguists dealt somehow with syntax; their approach was known as the Immediate Constituent Analysis (or IC analysis for short). For instance, in an "Indian" utterance like the boyhittheball we would recognize only the upper unit (which is the sentence) and the lower unit (which is the sound), but in between we would be left with a problem, if not many. Fortunately, the scholars of English (as it happens to be English) have left spaces between the words. So we have the sequence: the boy hit the ball. To move further, we can use SUBSTITUTION, as in the sentence John hit Mary. To prove that the boy (and a ball) form one unit. Substitution again shows that hit a ball is also a unit (cf. John came). Finally, the upper unit as we know, is the whole sentence the boy hit a ball. This can be represented diagrammatically as follows:

(1)

We can say that although IC analysis is very explicit, it is rather limited as it does not draw any GENERALIZATION about language. (Soams & Perlmutter, 1979) Suppose now we started from top to bottom instead of bottom upwards. We would obtain what follows:

(2)

Which will generate the boy hit a ball and

the boy hit the ball and

a boy hit a ball and

a boy hit the ball and

a boy hit a boy etc.

So we can see the merits of this new approach known as PHRASE STRUCTURE GRAMMAR (or the mathematical way of dealing with sentences which is much richer).

Now a piece of terminology seems necessary:

(3)

a. ( ): the item inside the parentheses is optional, as in: the (old) man - old is optional. This is in terms of acceptability.

b. [ ]: the items included in the square brackets form a unit like [the old man] which is a noun phrase.

c. { }: the items included in the braces form a paradigm or table of substitution as in:

d. stands for a phrase structure rule.

e. ===> stands for a transformation.

What we have established so far is the internal structure of the sentence the boy hit a ball, or indeed the generative system for all the simple, declarative affirmative (KERNEL) sentences of English (at least). It is called a TREE diagram, or PHRASE MARKER (P-MARKER for short). It actually has two alternatives, the LABELLED BRACKETING as in (4):

(4) [s[NP[Det the][N boy]]]

[VP[V hit][NP[Det a][N ball]]]]

and PHRASE STRUCTURE RULES1:

1.1.2. Transformations

Now phrase structure grammar (PS grammar for short) itself is limited. It cannot possibly generate/link all the sentences of English. It especially faces problems such as DISCONTINUITY as in the sentence: In God, we trust. (in the Dollar). We know intuitively that in fact it is: we trust in God and likewise in the sentence: what are you looking at that woman like that for, boy? (Griffiths, Black Like Me) we know intuitively that what... for form one unit which is why.

PS grammar also faces the problems of DIFFERENCE between sentences (known as PARAPHRASE) as in the following paradigm:

(6)

a. The Moors defeated the Spaniards

b. The Spaniards were defeated by the Moors

c. It was the Moors who defeated the Spaniards

d. The Moors' defeat of the Spaniards (was total)

e. I expected the Moors to defeat the Spaniards

f. They defeated them.

Another problem faced by PS grammar (a semantic one this time) is AMBIGUITY2. The noun phrase old men and women has two meanings either old men and old women (the second occurrence of the adjective old having been deleted) or old men and any women.

Finally, PS grammar faces the problem of SIMILARITY. Some sentences like John is easy to please and John is eager to please seem to have exactly the same surface structure namely NP, V, Adj and S. But we know intuitively that the first sentence means something toughly like [it is easy - someone pleases John], whereas the second one means [John is eager - John pleases someone].

All these problems are solved by another grammatical rule called TRANSFORMATION. There were indeed many transformations in the early TG literature.

To take but one transformation, PASSIVE for instance, we need two main stages known as STRUCTURAL DESCRIPTION and STRUCTURAL CHANGE, as follows:

(7)

Three conclusions can be drawn about the first version of T.G.:

PS rules generate KERNEL sentences which are by definition simple, declarative and affirmative.