Transformational Generative Grammar

UNIVERSITY HASSAN II FACULTY BEN M’SIK OF HUMAN SCIENCE AND LITERATURE ENGLISH DEPARTMENT CASABLANCA

RESEARCH PAPER FOR GRADUATION

Transformational Generative Grammar Some Aspects of Movement Transformation

Under the Supervision of

Pr. Mounia Amrani

A research by the undergraduate: Bouchaib Touilaat

01/06/2015

A CKNOWLEDGMENT

My gratitude to the supervisor of this research paper, Pr. Mounia Amrani, is beyond words. Not has she had to reiterate an explanation during the meetings for this project, but she has also been a source of fascination for four years which contributes to a better understanding of English syntax. Since the first semester till now, her lectures that analyse in depth the syntax of English grammar in a very interactive way have been always a challenge for me, a learner, to search and ask more questions. Every question that I ask Pr. Amrani has had always an answer with detailed explanation. I do remember her first lecture about tenses, and to my surprise I had no idea of what subject complement is. Since then by careful listening to her and many times by interrupting her, a sense of inspiration has developed to adopt her scientific and simplified method of analysis that I do not find in any textbook. In fact, she has guided me to choose the right textbook, although sometimes I do insist on other textbooks. However, after a self criticism I find her choice is very right indeed. It is impossible for a learner of syntax, if attending her lectures, to not interact. A total silence among the attendees can be noticed. If I compare myself now to the date before the first semester, it is a complete change that allows me to see the English syntax widely. This research paper will not be a lively one without her valuable suggestion, her correction, and her perceptive criticism. She has supervised my work step by step, page by page, and sentence by sentence. She spares no effort to explain again and again every time I handed her a chapter. Working under her was an extremely knowledgeable experience for me.

I do thank the librarians of the faculty for their help. I do thank my family who encourage me to finish my project. I do thank my sister Nadia Mendez. I do thank Pr. Jafari who has assisted me. I do thank Madam Amina, the responsible of English Department.

ii

CONTENTS

ACKNOWLEDGMENT ..................................................................................................................II INTRODUCTION ............................................................................................................................. 1 PLAN ................................................................................................................................................. 2 A MODEL OF TRANSFORMATIONAL GENERATIVE GRAMMAR: ..................................... 3 SECTION 1 .......................................................................................................................................... 1.

DEFINITIONS: BASE COMPONENT .................................................................................... 4 1.1

1.1

REWRITING RULES: ........................................................................................................................................ 4 1.1

Categorisation.......................................................................................................................................... 5

1.2

Subcategorisation ..................................................................................................................................... 7

1.3

Context-Free Subcategorisation................................................................................................................ 7

1.4

Context-sensitive subcategorisation: ........................................................................................................ 9

1.5

Level (1) : Context-sensitive Strict Subcategorisation .............................................................................. 10

1.6

Sub level: Context-sensitive Selectional Subacategorisation .................................................................... 12

1.7

Generation of the Surface Structure ........................................................................................................ 14 TREE DIAGRAM: .......................................................................................................................................... 15

SECTION 2 .......................................................................................................................................... I.

FORMALIZATION OF TRANSFORMATION : ANALYZABILITY ................................ 17

II. NP MOVEMENT ..................................................................................................................... 17 1.

PASSIVE TRANSFORMATION: ............................................................................................................................. 17

2.

DATIVE MOVEMENT .......................................................................................................................................... 22

III.

S’ MOVEMENT .................................................................................................................. 24

S’ EXTRAPOSITION .................................................................................................................................................... 24

ORDERING TRANSFORMATIONS: .......................................................................................... 26 CONCLUSION:............................................................................................................................... 30 BIBLIOGRAPHY............................................................................................................................ 31

iii

I NTRODUCTION Every sentence in English has a deep structure and a surface structure. Based on (Chomsky), we define generative grammar as a system of rules that are applied successively in order to generate a deep structure of a sentence, interpret it, transform it to a surface structure, and eventually interpret this surface structure. This system of rules is analysed according to three major components: the syntactic component, the semantic component, and the phonological component. The syntactic component generates the deep structure of a sentence by applying categorial rules 1, and generates the surface structure by applying transformational rules. The semantic component operates within the base of the syntactic component and interprets the deep structure generated by categorial rules. The transformational subcomponent of the syntactic component transforms the deep structure to a surface structure. The phonological component interprets the latter. The syntactic component determines a set of abstract category and subcategory symbols, each of which incorporates all the information relevant to semantic and phonological interpretation of a particular sentence. That is, the syntactic component applies a system of phrase structure rules 2 that generate a string of terminal category and subcategory symbols. This string is terminal, if the information that is provided by its symbols is relevant for the insertion of the lexicon elements. The latter must follow the interpretive rules that are applied by the semantic component. Therefore, the semantic component relates a structure generated by the syntactic component to a semantic representation from the lexicon. Besides, the phonological component relates a structure generated by the syntactic component to a phonetically represented signal. In addition to its base, the syntactic component contains a transformational subcomponent that applies certain formal operations on the deep structure and transforms it to a surface structure. Such operations involve movement, insertion, deletion, and so on. Hence, every deep structure of a sentence must contain a transformational marker which indicates the transformation(s) to apply.

1

Chomsky suggests Categorial Component. See pages 67 of (Chomsky).

2

also Categorial rules

1

P LAN

This research paper is divided into two sections. The first section will discuss major definitions regarding the Generative Grammar, mainly subcategorisation rules, giving a general idea of how the base component operates. The latter will be explained systematically elaborating on Chomsky’s simple example 3. We start from the deep structure to the surface structure. That is, we will start from the abstract category symbols generated by the categorial component (or phrase structure rules), and stage by stage till the surface structure is generated. In every stage, there is a definition entry accompanied with examples and a continued interpretation. The insertion of lexicon will not be by convention but by the means of the interpretive rules the semantic component applies. Presumably, all the definitions given in this section are based on (Chomsky), Chapter 2. The second section will be indeed reversal to the first one. The primary concern of this section is to analyse some Movement Transformations. We will start from the surface structure, assuming a transformational rule, drawing tree diagrams of the deep structure, and analysing every diagram with respect to the structural description and structural change. This section is based on (Heny et Akmajian) 4, (Chapter 5), elaborating on different transformational rules. This is not an exhaustive investigation of all Movement Transformations, yet it is only an initiation. Eventually, it should be noted that the italicized typing of the lexicon and the surface structures represents phonological interpretation.

3

Sincerity may frighten the boy

4

See bibliography

2

A M ODEL

OF

T RANS FORMATIONAL G ENERATIVE G RAMMAR :

3

SECTION 1 • Major Definitions of Generative Grammar:  How the Base Component operates • Elaboration on a simple example

1. D EFINITIONS : B ASE C OMPONENT

1.1

R EWRIT ING

R ULES :

The mechanism of the base component is oriented with respect to the order of left and right. This mechanism categorises and subcategorises by the means of a system of rewriting rules. Based on Chomsky’s definition, a rewriting rule is a rule of the form A→Z / X─Y It is interpreted as the category A is realized as a string Z when it is in the environment consisting of X on the left and Y on the right. That is, a rewriting rule expands a single category symbol as one or two categories, or subcategory symbols (complex symbols) with respect to the left-and-right order. For instance, if A stands for NP (Noun Phrase); X for Det (Determiner); and Y for N (Noun), we say that the category NP is realised as the string Det─N in the environment where Det should be on the left and N on the right. More examples: Monotransitive Verbs is subcategorised as: Monotransitive V→ [+ V, +─NP] 5 ( in the environment where V should be on the left and NP on the right)

categorised as

Sentence      → the string (NP) (Predicate) / NP─Predicate Phrase ( in the environment where NP should be on the left and Predicate Phrase on the right) The arrow in a rewriting rule can be also read as “rewrites as” , “expands as”, or “categorised as”. However, “subcategorised as” is not synonymous to “categorised as” , while “rewrite as” or “expand as” means both of them.

5

According to Chomsky, this symbol between two brackests is called Complex Symbol, Cs.

4

1.1

C ATEGORISATION

In order to generate the deep structure of a sentence, the phrase structure rules 6 in the base of the syntactic component applies a system of categorial rewriting rules that take the form of a sequential derivation that results into terminal categories. This sequence of categorial rules refers to categorisation. Rules such as NP→Det─N, or VP→V─NP, are categorial rules. We define categorial rules as A→Z/ X─Y in which neither A nor Z involves any complex symbols, and we say that A can be categorised to another string of categories. Hence, a categorial rule is a rewriting rule that branches two categories with respect to the order of the constituents of the deep structure of the sentence. They have two seperate functions: a) they define the system of grammatical relations, and b) they determine the ordering in the deep structure. The essence of categorial rules is that they determine whether a constituent modifies a sentence, or it is dominated by another constituent within the deep structure of a sentence. However, they are not sufficient for the insertion of the lexicon. They should occur in accordance with the subcategorisation rules. In fact, subcategorisation rules operate within the syntactic context of categorial rules. Besides, categorial rules are applied in a sequential derivation. That is, in the deep structure of a sentence, they start initially from the symbol S (stands for sentence) that expands as the constituents NP (Noun Phrase), Aux (Auxiliary) and VP (Verb Phrase), to a terminal string in which category symbols can no longer expand, and therefore, the derivation is terminated. For instance, category symbols like N, V, Det, Prep, and Adj are terminal because they can no longer derive to other categories. The obligatory categorial rule, S→ NP─Aux─VP, is considered as the main construction from which all other rules derive. That is, every deep structure of a sentence in English should be generated from that rule. We say it dominates all other subrules in the deep structure. On the left, the NP 7 (the noun phrase, or an embedded sentence) refers to the Subject of the deep structure of

6

Chomsky suggests ‘Categorial Component’ instead of ‘Phrase Structure Rules’ Component. According to him ‘phrase structure rules’ are independent of order. 7

It can be substituted by Pronoun by the Pronomilisation transformation

5

the sentence. The NP can be categorised to other rules as in “NP→Det─N”, where Det stands for Determiner; and N for Noun. On the right, the Aux─ VP ( Predicative Phrase of the deep structure) is composed of two categories: Aux stands for Verbal Auxiliary and VP for Verb Phrase. Aux incorporates the information such as (Tense, Modal, Aspect), which is determined by subcategorisation rules. The VP contains the base of the Verb in the deep structure and other categories that relate to the Verb (Complement, Adverb, etc..). The VP dominates all other constituents which make part of the predicate of the deep structure of the sentence. The VP can be categorised to other rules, such as “VP→V─NP".

For instance, the base component applies the following sequential derivation of categorial rules 8: 1. S→ NP ─Predicate-Phrase 2. Predicate-Phrase →Aux─VP 3. VP→V─NP 4. NP→N 5. NP→Det─N Applying (2) into (1), we shall get: 6. S→ NP─Aux─VP

(non-terminal) (continue branching)

Applying (3) into (6): 7. S→ NP─Aux─V─NP (non-terminal)

(continue branching)

Applying (4) into (7): 8. S→ N─Aux─V─NP

(non-terminal) (continue branching)

Applying (5) into (8): 9. S→ N─Aux─V─Det─N

8

(terminal)

They can be also called phrase structure rules

6

(stop branching)

The string (9) is terminal. The categories N, Aux, V, Det, and N can no longer expand in terms of categorial rules. That is, there is no more branching. Moreover, still the insertion of the lexicon formatives is not possible. There should be other rules to apply.

1.2

S UBCATEGORISATION According to (Chomsky) 9, we define subcategorisation rules as the system of interpretive

rewriting rules that the semantic component apply to subcategorise a syntactic or a lexical category into a Complex Symbol, Cs, that contains positively or /and negatively specified syntactic features inside two brackets, [+Category, ±Feature1, ±Feature2,...]. Those Features can be either inherent or contextual features. For example: Aux

as Subcategor   ised  → [+tense, Modal, have+en→ , be+ing→]

Collective Nouns → [+Common, -Count, -Abstract, +Animate, +Humane, +Plural] ({the elderly, kith and kin, committee...}, [+N, +Common, +Collective→, +Animate, +Humane])

1.3

C ONTEX T -F REE S UBCATEGORISATION

In order the semantic component relates the terminal string (9) to a lexical representation, it is necessary that there should be prior inherent features, for every single lexical category, introduced under the form of complex symbol for V, N, and Det. That is, prior context-free subacategorisation for V ,N, and Det. For every lexical representation, there should be inherent features, and therefore the lexicon is subcategorised under the sequence of (D,C), where D stands for a lexical category 10 or formatives, and C stands for complex symbol which contains specified inherent features. For example (D,C) = ({the elderly, kith and kin, committee...}, [+N, +Common, +Collective, +Animate, +Human])

List of Lexicon

9

See bibliography

10

Copula is a lexical category, whereas {taste, feel,...} are lexical subcategory within the Copula.

7

Some 11 Context-free subcategorisations of V as enlisted in the lexicon, (Wekker et Haegeman):

Copula (Lexical category)

Monotransitive

Lexical Representation

Inherent features 12 (complex symbol)

{be, become,...}

[+V,+ ─NP]

{taste, feel, be, ...}

[+V,+ ─AdjP]

{be}

[+V, + ─PP]

{ like, believe …}

[+V, + ─NP] (where NP→(Det) N (S’) )

{think, show...}

[+V, +─S’]

{eat, read, like ,believe,

[+V, +─N]

/

[+V,+ ─N]

*

*

complete,...} Middle Verbs

{have, resemble, ....}

[+V, +─N]

{have}

[+V,+─(Det)N(S’)]

Context-free subcategorisation of N : The lexicon of N is mapped according to the following diagram: (next page)

11

We did not include Ditransitive, Complex Transitive, Transitive+PP, Prepositional Verbs, Phrasal Verbs

12

Chomsky suggests that there could be more features included, such as phonological features.

8

*

However, the above diagram does not cover all Nouns, such as Collective Nouns with the entry {The elderly, kith and kin...}.We also notice that {the} can be part of the Proper N, [-Common] , such as {the U.S}. ({The elderly, kith and kin...}, [+N, +Common, +Collective, +Animate, +Human] ) We can expand the tree above, by adding the feature [±Plural], at the bold points where indicated, which will duplicates for example the entry {ant, lion, mouse..} as {ants, lions, mice...}. Context-free subcategorisation of Det : the lexicon of Det looks like this Lexical representation {the}

Inherent features [+Definite,+Main, +─Common, ±─Plural]

(It occurs with

Common Singular, [-Plural], and Plural, [+Plural], Nouns

The semantic component will start the interpretation of the string.

1.4

C ONTEX T - SENSITIVE

SUBCATEGORISATION :

We define context-sensitive subcategorisation rules as a system of interpretive rules that are applied by the semantic component. First, they determine the whole terminal string of categories 9

of the deep structure into possible sections of substrings with respect to the interpretation of the position of V in the terminal string of categories, allowing only the insertion of V formatives from the lexicon. Second, they permit in every section the Sub-Vs to select on the left and on the right (in terms of the N’s inherent features, for string (9) . The following figure illustrates the two types of context-sensitive subcategorisation respectively.

1.5

L EVEL (1) : C ONTEX T - SENSITIVE S TRICT S UBCATEGORISATION

Considering string (9): N─Aux─V─Det─N Semantic component interpreting the position of V: Subject─ V─Subject Complement OR

Subject─V─Direct Object

OR

Subject─V─Object 10

The string (9) implies a syntactic context for V. That is, at the first level of interpretation the semantic component interprets V, using syntactic information, in the following way: Since V in string (9) is set between two Nouns, that is, immediately a N precedes the V and a N follows the V, and since the first N on the left refers to the Subject of the sentence, and the second N on the right can be either Subject Complement, Object of the Verb, or a Direct Object, there are three possibilities for V to be strictly sucategorised. V is either a Copula, Middle Verb, or Monotransitive. That is, there are three possible lexicon entries for V. We call this type of rules that determine the possible lexicon entries for V with respect to the position of V in the terminal string of categories in the deep structure of a sentence as context-sensitive strict subcategorisation rules.

A.

[+Copula, + N - Subject Complement]    V      →[+ Monotransitive, + N - Od]  [+ Middle Verbs, + N - Object]    Strictly subcategorised

This rule interprets the lexicon as the following:

Not all Copula Verbs only those that dominate immediately Subject Complement Noun: ({be, become,...}, [+V, +─N] ) Not all Monotransitives only those that dominate immediately a Direct Object Noun: ({eat, read, believe, complete, finish, accomplish, buy, sell, see, hear,...}, [+V, +─N]) Not all Middle Verbs only those that dominate immediately an Object Noun: (excluding the causative) ({have, resemble, ...},[+V, +─N] ) Considering string (9):

11

N─Aux─V─Det─N Applying strict subcategorisation rule A, will insert the lexicon of V, and the string (9) is converted to three substrings:

(12) Substring 1

Substring 2

Substring 3

N─Aux─[+Copula]─Det─N

N─Aux─[+Mono]─Det─N

N─Aux─[+Middle]─Det─N

 be     become    .  .    .

 eat  read    frighten     finish  complete    chart  ....  

( Strict subcategorisation rules are terminated )

have        resemble   .... 

(V insertion is terminated).

Still N, Det, and N are not inserted. There should be other rules to apply

1.6

S UB

LEVEL :

C ONTEXT - SENSITIVE S ELECTIONAL 13 S UBACATEGORISATION

We define context-sensitive selectional subcategorisation rules as subtype rules of strict subcategorisation rules that permit the Sub-Vs to select its Subject-Noun and its ComplementNouns in terms of the inherent features of Nouns, assigning to the Subject Noun the features [±pers3rd, ±Plural]. The latter is necessary for Subject-Verb Agreement. The lexicon items {be, eat,

13

Chomsky call them «Selectional » and they can be also called “Selective”

12

have} inserted in (12) determine what features should be assigned to their Subject-Nouns and their Complement-Nouns. Moreover, this subtype of selection must operate in a way that does not allow deviation, such as, {John is a table*}, {Egypt eats the dog* }, or {Lions have The Hague*}. Furthermore, once right N is selected in every substring in (12), the right N will also select the Det. It should be noted that there could be other specified features for both N and Det in the lexicon, so their insertion is not deviant such as {his metropolis*}. Besides, deviation can be accepted in the metaphoric and poetic English, such as {Sun is the Queen}, which implies that the complex symbol can include more other semantic features to generate such methaphor. The following rules are examples of subtype rules applied to the three substrings in (12) respectively: Substring 1:

N─Aux─[+Copula]─Det─N

 + Subject [+N, - Common, - Animate, + persrd, , − Plural] {be} →   + Subject Complement [+N, + Common, + Count, - Animate, − Plural] 

Substring 2:

    

N-Aux-[+Monotrans]-Det-N

 + Subject [+ N; - Common, + Collective, + Animate, + Human, + pers3, + Plural ] {chart} →     + Od [+ N; + Common, + Count, − Animate, + Plural ]

Substring 3:

N─Aux─[+Middle]─Det─N

 + Subject [+ N, − Common, + Animate, + Human, 3 pers, − Plural ]  {have} →    + Object [+ N, + Common ,+ Count ,− Animate, + Plural  ]   

13

1.7

G ENERATION

OF THE

S UR FACE S TR U CTURE

Supposing that Aux is selected by a semantic context, then subcategorising Aux as, in substring 1: Aux → [+present] (where other features of Aux are not occurring) Applying the selective rule in substring 1, it converts the latter to: N[ + Subject [+N, - Common, - Animate, + persrd  , , − Plural  ] ]-Aux[+pres→]be-Det 14 [ + Definite , + Main, ∩ ± Plural] ─ N[+N, + Subject Complement [+N, + Common, + Count, - Animate, − Plural] ]

The semantic component relates this structure to the lexicon, including other semantic features that will exclude {house} from the right: {Casablanca, Istanbul, Egypt, The Hague... }- Aux[+pres] {be}-{my, his, the,…}-{city, country, court, ..}

Now, the transformational subcomponent will transform the substring above, using the transformational marker where the arrows are indicated, generating the surface structure:

Casablanca is my city. The same process will be followed for other two substrings. We summarize the findings in a table. N─Aux─V─Det─N N─Aux─[+Copula]─Det─N

N─Aux─[+Monotr]─Det─N

-{be}-

- {chart}-

Applying Sub 1 Aux[+pres]

Transformation:

N─Aux─[+Middle]─Det─N -{have}-

Applying Sub2

Applying Sub3

Aux[+Pres, +Have+en]

Aux[+Pres, +Modal]

Affix hoping for V

Casablanca is my city.

The House of Commons have John might have two caps. charted several laws.

14

All the Main Determiners that occur with either Plural or Singular [-Plural] Nouns, {a, several,..} are excluded.

14

1.1

T REE D I AGR AM :

The categorial rule A→Z/ X─Y can be converted to the tree: A X

Y

This diagram is called a tree diagram that consists of a node A and has two branching. The two categories X and Y are sisters, as they are dominated by one node. X and Y can be subject to derivation, and therefore they can form both of them two nodes. Returning now to (chapter 1.1), the sequential derivation rules (1), (2), (3), (4), (5), unlike the linear application of those categorial rules, the tree diagram simplifies them indeed. We shall draw a tree diagram of the string (9), including temporarily the complex symbols of all categories.

15

16

S ECTION 2 •

NP Movement o Passive Transformation:  (Heny et Akmajian)’s argument o Dative Movement

• Extraposition: S’ Movement • Ordering Transformations

I.

F ORMALIZATION

OF

T RANS FORMATION : A NALYZABILITY

Our analysis of Movement Transformation will be in terms of structural description (SD) and structural change (SC), comparing it with a tree diagram. The SD describes the order of constituents in the input tree (tree diagram of the deep structure) by assigning for every constituent a number term. The SC describes the order of the constituents in the output tree (tree diagram of the surface structure), and consequently it will determine which transformation occurs. In , the statement of a transformational rule abstractly would look like : Structural Description:

Structural Change :

X

A

B

C

Y

1

2

3

4

5

1

4

ϕ

2

5

where X and Y are variables that do not change when a transformation apply, and A, B, C are phrase-nodes (we will deal only with phrase-nodes) on which a rule can apply.

A transformation can apply to an input tree if and only if the input tree is analyzable in accordance with the structural description. That is, if the input tree does not satisfy the conditions of the SD, the transformational rule does not apply.

II.

NP

MOVEMENT

1. P ASSI VE T R ANSFORM AT ION : The following surface structures are generated by a Passive Transformation: a. Germany was defeated by England. b. Her baby has been called George.

17

We will adopt (Heny et Akmajian) ‘s argument.

Deep Structure of (a): Passive Transformation Optional Akmajian suggests the deep structure of (a) to be an active sentence :

18

Analysis of the underlying structure of (a): The base component generates a Transformational marker, the Transitivity Property of V. That is, the Phrase Structure Rules generate the structure (S→NP-Aux-V-NP), where NP is a Direct Object. According to Akmajian, since Passive Transformation applies only on Transitive Verbs, so the Transitivity Property of V is in itself a Transformational Marker. Passive Transformation applies only and only if the PSR generate the structure (S→NP- X -Aux-V-NP-Y), where X and Y are variables. If the PSR generate the structures such as (S→NP-Aux-V) or (S→NP-Aux-V-PP), the Passive Transformation does not apply. The Transformations that should apply to generate the sentence (a) are Movement and Insertion, according to the following rule:

Structural Description :

Structural Change:

NP Aux 1

2

4

2 > be + en  ↑

V

NP

3

4

3 + by  ≠1 ↑

In order to generate the surface structure (a), simultaneously, the NP, Direct Object, in the fourth term preposes to the position of the NP in the first term, and the NP in first term postposes to the position of the NP in the fourth term. This shifting of positions requires obligatorily the insertion of two elements in the SC, be+en and by. The element be+en must be inserted to the rightmost ( > ) of the node Aux. That is, be+en is daughter-adjoined as the rightmost daughters of Aux. The element by is inserted in the SC adjoining the NP node in the term 1. Such insertion does not build any new structure, such as PP, but rather by is sister-adjoined of V and Chomsky-adjoined of NP term 1 without affecting the structure of NP term 1.

It should be noted that the Passive

Transformation in case (a) is optional. The following Tree Diagram of (a) elucidates the Passive Transformations that occurred: (Next Page)

19

Deep Structure of (b): By using Akmajian’s argument, the deep structure of (b) is roughly:

20

Analysis: The base component generates a Transformational Marker, the incorporated structure (S→NP-AuxV-NP) where the second NP is specified as a direct object. However, this time the subject in the deep structure is an indefinite pronoun which will require an agent deletion transformation. Two transformations should occur, Movement and Insertion according to the following rule.

Structural Description :

Structural Change:

NP

Aux

V

NP

NP

1

2

3

4

5

4

2 > be + en  ↑

3

1

5

+ by  ↑

In order to generate the surface structure (b), the NP in the fourth term preposes to the position of term 1 and the NP of term 1 postposes to the position of term 4 . This preposing and postposing requires the insertion of one element, be+en. The latter is adjoined to the Aux node and becomes its rightmost daughter. The fifth term does not change. Eventually the following structure is generated: Her baby has been called George by someone. Another transformation should apply to generate the structure (b). it is agent deletion by which the phrase (by someone) is deleted.

21

2. D AT IVE

MO VEMENT

Dative Movement rules relate pairs of sentences such as: (1) a. They refuse permission to enter to visitors. b. They refuse visitors permission to enter. (2) a. They built a park for the residents. b. They built the residents a park. (3) a. They built it for the residents. b. *They built the residents it

We assume that Dative Movement transforms the structures 1a and 2a to the structures 1b and 2b respectively. Considering the following input tree of 1a and 2a (limiting the tree diagram only on the VP

node) :

22

The base component generates a transformational marker, the preposition to or for . That is, the PS rules generates the expanding rule, VP→ (V) (NP) to/for (NP) in which the Ditransitive V selects either to or for, and selects the first NP as a direct object and the second NP as indirect object. Once those conditions are met as well as the fact that the lexical verbs such as {refuse, build} are specified as [+Dative Movement], Dative Movement applies optionally. This transformational rule is a governed rule, that is, it depends on the specific main verb of the sentence. Accordingly, Dative Movement has the effect of moving the indirect object (the second NP, the prepositional complement) to the position immediately following the verb, thus making that NP the new direct object, and deleting the preposition to or for. We can state the rule as follows: SD:

SC:

V

NP

for    to 

NP

1

2

3

4

1+4

2

ϕ

In the structure 3b, Dative Movement does not apply due to the fact that the first NP is pronominalized with the pronoun it .

23

III.

S’ M OVEMENT

S’ E XT RAPOSIT ION

The following structures are related by an optional Transformational rule, S’ Extraposition : i. ii.

For the house to be painted would irritate John (Extraposition is not applied) It would irritate John for the house to be painted ( Extraposition is applied)

iii.

Everyone believes it that the world is round.

According to Akmajian, in the Extraposition analysis, the NP can be expanded as the following : NP N

S’

it

That is the head of NP in the deep structure can be it, and the latter is dominated by N. Basing on this NP expantion, the input tree of the sentence (ii) would be :

The base component generates a Transformational Marker, the head Noun it and an optional S’ which is empty at the end of the sentence. It is a Trace in the deep structure, according to

24

Chomsky, to where the S’ should move. However, the empty S’ is optional, that is, if the Extraposition does not apply, it-deletion must apply instead. In order to generate the sentence (ii), the S’ that follows the head Noun in deep structure should extrapose to the position of the empty position S’ at the end of the sentence. In this transformation, S’ becomes dominated by the highest S . Eventually the following rule applies on the input tree (page 24): SD :

SC :

it

S’

Aux

V

NP

S’

1

2

3

4

5

6

3

4

5

2

1

Sentence (iii): The input Tree of the sentence (iii):

25

O RDERING T RANS FORMATIONS : The following structure is generated by ordering transformational rules: i)

It was suggested by the psychologists that children should be told a story before bedtime.

The transformational rules that are applied to generate this structure are according to this order: 1. Passive Transformation in the main sentence 2. Extraposition of the embedded sentence 3. Dative Movement in embedded sentence 4. Passive Transformation in the embedded sentence 5. Agent Deletion in the embedded sentence

The underlying structure of the sentence (i) is: The psychologists suggested it that ( someone) should tell a story to children before bedtime which has the following input tree:

26

1. By the application of Passive Transformation rule:

SD : NP 1 SC : 4

Aux

V

NP

2

3

4

2 > be + en 

3+ by ≠ 1

↑

↑

We shall get the following output tree:

27

2. By the application of Extraposition rule:

SD:

SC:

N

S’

X

1

2

3

1

3

2

We shall get the following output tree:

28

3. At the level of the embedded sentence S’ respectively Dative Movement, Passive Transformation, and finally Agent Deletion apply.

29

C ONCLUSION :

We have seen that in order the surface structure of a sentence to generate, the deep structure of the sentence undergoes several stages. First, in the base component, the categorial component generates abstract categorial symbols that should undergo a sequential derivation by application of rewriting rules till those category symbols are terminal. Then , Once they are terminal, the semantic component, which operates within the base, relates every category symbol to its lexical representation

according to the interpretation of the position of the verb with respect of what it

immediately follows and what it immediately precedes. The latter are called selective rules which interpret using complex symbols allowing the lexical items to be inserted. This generated deep structure contains always a marker by which the transformational rules apply. This transformational marker should be indicated in the structural description. In the study of transformational rules, we have concluded the following points: • Any transformation or sequence of transformations when applied does not build new structures. That is, if a structure is NP in the SD, it will definitely remain an NP in SC. •

Movement transformation changes the order of phrase-nodes or non-phrase nodes to other position in the tree, sometimes deleting or /and inserting new elements.

• Transformationalists can have different arguments on the way the deep structure of a sentence would look like. That is, any sentence can have different deep structures. •

There is a relationship between the deep structure and transformation. Once the transformational marker is defined, the SD is stated and then it becomes easy to assume the deep structure of a given sentence.

•

Transformational rules have certain constraints which indicate either those rules apply or not.

Nevertheless, any transformational rule in English can be subject to acceptance or rejection when they are compared cross-linguistically, that is, whether those rules are analyzable in other languages or not.

30

B IBLIOGRAPHY

Chomsky, Noam. Aspects of The Theory of Syntax. Massachusetts: The Massachusetts Institute of Technology, 1965. Emonds, Joseph E. Transformational Approach to English Syntax: Root, Structure Preserving and Local Transformations. Michigan: Academic Press Inc , avril 1976. Heny, Frank and Adrian Akmajian. An Introduction To The Principles Of Transformational Syntax. Cambridge, Massachusetts and London, England: The MIT Press, 1975, 1980, 1981. Perlmutter, David M. and Scott Soames. Syntactic argumentation and the structure of English. Berkeley, Los Angeles London: University of California press, 1979. Wekker, Herman and Liliane Haegeman. A Modern Course in English Syntax. London, New York Guilford and King's Lynn: Croom Helm Ltd, Biddles Ltd, 1985, 1989, 1991, 1992, 1993, 1996.

31