Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Recursive grammars and the creative aspect of language use Angel, Jay Leonard 1974

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata

Download

Media
831-UBC_1974 A1 A549.pdf [ 10.91MB ]
Metadata
JSON: 831-1.0106613.json
JSON-LD: 831-1.0106613-ld.json
RDF/XML (Pretty): 831-1.0106613-rdf.xml
RDF/JSON: 831-1.0106613-rdf.json
Turtle: 831-1.0106613-turtle.txt
N-Triples: 831-1.0106613-rdf-ntriples.txt
Original Record: 831-1.0106613-source.json
Full Text
831-1.0106613-fulltext.txt
Citation
831-1.0106613.ris

Full Text

RECURSIVE GRAMMARS AND THE CREATIVE ASPECT OF LANGUAGE USE by JAY LEONARD ANGEL A t h e s i s submitted i n p a r t i a l f u l f i l m e n t of the requirements f o r the degree of Doctor of Philosophy i n the Department of Philosophy at the U n i v e r s i t y of B r i t i s h Columbia December, 1973 We accept t h i s t h e s i s as conforming to the requ i r e d standard^ In p r e s e n t i n g t h i s t h e s i s i n p a r t i a l f u l f i l m e n t o f the requirements f o r an advanced degree a t the U n i v e r s i t y o f B r i t i s h Columbia, I agree t h a t the L i b r a r y s h a l l make i t f r e e l y a v a i l a b l e f o r r e f e r e n c e and study. I f u r t h e r agree t h a t p e r m i s s i o n f o r e x t e n s i v e c o p y i n g o f t h i s t h e s i s f o r s c h o l a r l y purposes may be g r a n t e d by the Head o f my Department o r by h i s r e p r e s e n t a t i v e s . I t i s understood t h a t c o p y i n g or p u b l i c a t i o n of t h i s t h e s i s f o r f i n a n c i a l g a i n s h a l l not be a l l o w e d w i t h o u t my w r i t t e n p e r m i s s i o n . Department of ?ki 10$° y The U n i v e r s i t y of B r i t i s h Columbia Vancouver 8, Canada Date / 0 ABSTRACT The aim of t h i s study i s to discover the r e l a t i o n s h i p between the a b i l i t y of a human being to use language to express ever new thoughts, on the one hand, and the presence i n n a t u r a l languages of devices f o r the d e r i v a t i o n of an i n f i n i t y of deep s t r u c t u r e types f o r sentences, on the other. The c o n c l u s i o n reached i s that the f u l l expression of thought can be c a r r i e d on i n a language with a f i n i t e and small number of deep s t r u c t u r e types f o r sentences. This apparently c o n f l i c t s with the claims of Noam Chomsky and other contemporary l i n g u i s t s to the reverse e f f e c t . The s i g n i f i c a n c e of t h i s r e s u l t i s f u r t h e r explored w i t h i n the o v e r a l l context o f generative l i n g u i s t i c theory. The m a t e r i a l i s organized i n the f o l l o w i n g way. The f i r s t chapter begins with an e x p o s i t i o n of the b a s i c elements of generative t r a n s f o r m a t i o n a l theory, and focuses on the d e f i n i t i o n of r e c u r s i v e sets of r u l e s and r e c u r s i v e d e r i v a t i o n s of sentences. I t continues with an e x p o s i t i o n of the theory of generative l i n g u i s t i c s as a branch of human psychology, and r a i s e s , and b r i e f l y d i s c u s s e s / the question of the extent to which progress i n d e s c r i p t i v e l i n g u i s t i c s a u t o m a t i c a l l y c o n s t i t u t e s progress i n that p o r t i o n of explanatory l i n g u i s t i c s which attempts to exp l a i n the p o s s i b i l i t y of there being r a t i o n a l as opposed to 1.1 n o n - r a t i o n a l creatures, creatures who use language c r e a t i v e l y as opposed to creatures which are incapable of such a use of language. The question i s not pursued very f a r , but i t i s c l e a r that Chomsky and other generative t r a n s f o r m a t i o n a l l i n g u i s t s such as Fodor and Katz take i t that the r e c u r s i v e generation of sentences i s a necessary feature of the language of a r a t i o n a l c r e a t u r e , or a creature whose thought processes and whose expressive c a p a c i t i e s are on a par with those of humans. This c l a i m i s an i n t e r e s t i n g one, but i t i s never supported by a n a l y s i s . A c c o r d i n g l y the chapter c l o s e s w i t h a d i s c u s s i o n of the means whereby the c l a i m may be checked. The second chapter performs a t e s t on a c r u c i a l p o r t i o n of the claim, and y i e l d s the f o l l o w i n g negative r e s u l t : the r e c u r s i v e generation of deep s t r u c t u r e types f o r sentences i s not necessary f o r the f u l l expression of thought. The chapter i s long and tangled, but u n i f i e d around the demonstration of t h i s p o i n t . In the t h i r d chapter, the s i g n i f i c a n c e of t h i s r e s u l t i s explored. I argue that the l e x i c o n contains some u n d e r l y i n g recursions — a p o i n t which i s apparently denied by Chomsky, Fodor, Katz and P o s t a l . I argue f u r t h e r than an extended study would be r e q u i r e d to determine, whether or not these recursions are necessary for thought expression. The question of 'recursive generation of paragraphs' i s a l s o r a i s e d , and some co n s i d e r a t i o n s are brought f o r t h suggesting that r e g a r d l e s s of whether or not l e x i c a l r e c u r s i o n s are necessary f o r thought expression, a c l a i m that they are, and a c l a i m that the r e c u r s i v e generation of paragraphs i s necessary fo r thought expression would have q u i t e d i f f e r e n t i m p l i c a t i o n s f o r semantic theory and p s y c h o l i n g u i s t i c s as compared with the c l a i m that the r e c u r s i v e generation of deep s t r u c t u r e types f o r sentences i s necessary f o r thought expression. F i n a l l y , the r e l a t i o n -s h i p between the a n a l y s i s of Chapter Two and the assumption i n semantic theory that the semantic content of a complex sentence i s a f u n c t i o n of the semantic content of i t s elementary p r o p o s i t i o n s and t h e i r s t r u c t u r a l r e l a t i o n s i s touched upon. I t i s suggested that a demonstration of the non-necessity of r e c u r s i v e generation of deep s t r u c t u r e types, supplemented i n c e r t a i n ways, provides a j u s t i f i c a t i o n f o r what i s now often assumed but not j u s t i f i e d -- that the semantic content of complex p r o p o s i t i o n s can be r e c u r s i v e l y s p e c i f i e d i n terms of the semantic content of i t s elementary p r o p o s i t i o n s and t h e i r s t r u c t u r a l r e l a t i o n s . TABLE OF CONTENTS Abs t r a c t , i Acknowledgements v Preface v i Chapter One A. Recursion i n Generative Transformational Grammars.. 1 B. Some Problems i n Generative L i n g u i s t i c s 23 C. On Checking the Nece s s i t y of Recursive Generation of Sentences f o r the F u l l Expression of Thought 62 Chapter Two A. P r e l i m i n a r y 69 B. R e l a t i v e Clauses 74 C. Noun and Verb Phrase Complements 94 Noun Phrase Complements 104 'That + Sentence' Noun Phrase Complements 105 'What* Clauses 144 'Whether' and ' I f Noun Phrase Complements 150 'Why' and 'How' Noun Phrase Complements 154 Noun Phrase Complements as Subjects of Verbs....... 159 Noun Phrase Complements Follo w i n g Nominalized Verbs 161 'For..to..' Noun Phrase Complements 164 'Poss...ing' Noun Phrase Complements 172 Verb Phrase Complements 174 Mixed M u l t i p l e Embeddings 181 D. S e n t e n t i a l Co-ordinations 183 E. Non-Sentential R e c u r s i v e l y Derived S t r u c t u r e s . . . . . . 196 Co-ordinated Noun Phrases 197 Nested Possessives 198 Compounded A d j e c t i v e s 199 Ad v e r b i a l Sequences 203 F. F i n a l Remarks on Chapter Two 203 Chapter Three A. A T e c h n i c a l Problem 211 B. Some Results From Chapter Two 216 C. The S i z e of the Lexicon i n Synchronic D e s c r i p t i o n . . 219 D. The P l a u s a b i l i t y of E l i m i n a t i n g L e x i c a l Recursions. 225 E. Implications of Chapter Two for Semantic and P s y c h o l i n g u i s t i c theory 229 L i s t of Works C i t e d 249 L i s t of Abbreviations 251 Appendix: Recursion and the Predicate Calculus 252 V ACKNOWLEDGEMENTS I am most g r a t e f u l f o r the generous a s s i s t a n c e I r e c e i v e d while working on t h i s p r o j e c t ; p a r t i c u l a r y , I would l i k e to thank Thomas Patton, my t h e s i s s u p e r v i s o r , f o r h i s p a t i e n t reviews of the d r a f t s of the t h e s i s and h i s numerous comments and suggestions f o r r e v i s i o n ; I a l s o warit to thank Howard Jackson Jonathan Bennett, and Sam Coval f o r t h e i r c r i t i c i s m s , suggestions and f r u i t f u l d i s c u s s i o n s on both the l a r g e r and smaller issues r a i s e d i n the study. N a t u r a l l y , while acknowledging t h e i r i n s i g h t s and suggestions, I assume f u l l r e s p o n s i b i l i t y f o r any and a l l e r r o r s which t h i s work may co n t a i n . I a l s o wish to thank the members of the Canada C o u n c i l f o r t h e i r k i n d f i n a n c i a l support during the years 1972-'73 and 1973-'74. v i PREFACE Language i s a m i r r o r of mind; hence i t s study has always h e l d a f a s c i n a t i o n f o r p h i l o s o p h e r s , and a c c o r d i n g l y there i s a long h i s t o r y of the philosophy of language. However, i n a d d i t i o n to a p h i l o s o p h i c a l account of language some h o l d that there must be, or that we ought to t r y to construct, an explanatory and t h e r e f o r e p s y c h o l o g i c a l account of language. In p a r t i c u l a r , accounting i n p s y c h o l o g i c a l terms f o r the o r d i n a r y and ' c r e a t i v e ' human use of language as an instrument f o r the f r e e expression of thought has been s e t f o r t h as the c e n t r a l goal of l i n g u i s t i c study i n the w r i t i n g s of Noam Chomsky, and other prominent contemporary t h e o r i s t s i n l i n g u i s t i c s . Of course, a statement of t h i s goal remains d i s t i n c t from the progress that has been r e c e n t l y made w i t h i n the f i e l d of d e s c r i p t i v e l i n g u i s t i c s . Chomsky, f o r example, des p i t e having made s u b s t a n t i a l c o n t r i b u t i o n s to the f i e l d of d e s c r i p t i v e l i n g u i s t i c s , i s on the whole p e s s i m i s t i c about the relevance of these c o n t r i b u t i o n s towards ac h i e v i n g the goal which he him s e l f defines to be the c e n t r a l one of l i n g u i s t i c study. In Form and Meaning i n Natural Languages, for example, he says, "with each advance i n our understanding of the mechanisms of language, thought, and behavior, comes a tendency to b e l i e v e that we v i i have found the key to understanding man's apparently unique q u a l i t i e s of mind. These advances are r e a l , but an honest a p p r a i s a l w i l l show, I think, that they are f a r from p r o v i d i n g such a key. We do not understand, and f o r a l l we know, we may never come to understand what makes i t p o s s i b l e f o r a normal human i n t e l l i g e n c e to use language as an instrument f o r the f r e e expression of thought and f e e l i n g ; or, f o r that matter, what q u a l i t i e s of mind are i n v o l v e d i n the c r e a t i v e acts of i n t e l l i g e n c e that are c h a r a c t e r i s t i c , not unique and e x c e p t i o n a l , i n a t r u l y human exis t e n c e . " (Language and Mind, enlarged e d i t i o n page 1 0 1 ) . Nevertheless, w i t h i n the w r i t i n g s of generative t r a n s f o r m a t i o n a l t h e o r i s t s , Chomsky's works i n p a r t i c u l a r , there are a number of suggestive and s u b s t a n t i a l remarks about the formal requirements f o r a language which i s to be used c r e a t i v e l y In t h i s study some problems concerning the r e l a t i o n s h i p between the s p e c i f i c theory of generative t r a n s f o r m a t i o n a l l i n g u i s t i c s and the accounting i n p s y c h o l o g i c a l terms f o r the ordinary, ' c r e a t i v e ' use of language as an instrument f o r the f r e e expression of thought w i l l be d i s c u s s e d . The bulk of the study i s devoted to checking the a s s e r t i o n s made i n various forms by Chomsky, Fodor, Katz, P o s t a l and other generative l i n g u i s t s that the s y n t a c t i c component of a grammar i n v i r t u e of i t s r e c u r s i v e generation of sentence s t r u c t u r e s i s the c r e a t i v e component of a grammar, or mo.e s p e c i f i c a l l y , that i f a language i s to accommodate the c r e a t i v e aspect of language use, i t s grammar must contain devices f o r the r e c u r s i v e generation of sentences. v i i i The c o n c l u s i o n reached i s that a sub-grammar of E n g l i s h with no r e c u r s i v e generation of deep s t r u c t u r e types f o r sentences i s r i c h enough to provide f o r the f u l l expression of thought. In p a r t i c u l a r , i t i s argued that u s i n g a small number of sentence s t r u c t u r e types one could express any thought that can be expressed i n E n g l i s h . This r e s u l t has p o t e n t i a l s i g n i f i c a n c e f o r p h i l o s o p h i c a l i n v e s t i g a t i o n s i n t o the l o g i c a l s t r u c t u r e of thought, and f o r the c o n s t r u c t i o n of semantic theory; i n a d d i t i o n , i t i s of i n t e r e s t with regard to the c o n s t r u c t i o n of the p s y c h o l o g i c a l account of the human use of language as c u r r e n t l y pursued by generative t r a n s f o r m a t i o n a l l i n g u i s t s and other t h e o r e t i c a l l y motivated p s y c h o l i n g u i s t s . 1 CHAPTER ONE A. Recursion i n Generative Transformational Grammars. The f i r s t p a r t of t h i s chapter contains an e x p o s i t i o n of some of the key terms of t r a n s f o r m a t i o n a l a n a l y s i s . The reader who i s a l t o g e t h e r u n f a m i l i a r with the views, methods, and terms of the t r a n s f o r m a t i o n a l grammarians i s advised not to t r y and piece the s t o r y together from the f o l l o w i n g account, but to go d i r e c t l y to the sources: Chomsky's Aspects of the Theory of Syntax, f o r example, or, f o r a l e s s t e c h n i c a l i n t r o d u c t i o n , Chomsky by John Lyons. On the other hand, there are many v a r i a t i o n s of terminology, angles of approach, and so on«, w i t h i n generative t r a n s f o r m a t i o n a l l i n g u i s t i c s , and the f o l l o -wing m a t e r i a l i t i s hoped, w i l l f i x to some extent what i s being meant by what term as i t occurs i n t h i s study. In general I adopt Chomsky's terminology and b a s i c framework, si n c e i t i s , i n l a r g e measure, h i s view about l i n g u i s t i c ' c r e a t i v i t y ' which I am examining. The concept of r e c u r s i o n plays a l a r g e r o l e i n the development of at l e a s t one p o i n t i n the 'Chomskyian r e v o l u t i o n ' i n d e s c r i p t i v e l i n g u i s t i c s , and i t should be mentioned at the outset that t h i s study does not remove, or attempt to remove, any of the grounds for t h i s r e v o l u t i o n . The r e v o l u t i o n i s 2 based on a twin demonstration: the n e c e s s i t y of regarding a grammar f o r a n a t u r a l language as a set of r e c u r s i v e generative r u l e s , and the n e c e s s i t y of p o s t u l a t i n g two l e v e l s of gramma-t i c a l s t r u c t u r e f o r sentences. To s t a t e the argument f o r regarding a grammar as a s e t of r e c u r s i v e generative r u l e s i n i t s simplest form, consider the f o l l o w i n g p a i r of sentences: 1.1 The man who saw the boy who saw the boy who chased the dog had gone. 1.2 The man who saw the boy who saw the boy who saw the boy who chased the dog had gone. From such sentences we can e a s i l y r e a l i z e that we can't p l a c e l i m i t s on the numbers of grammatical sentences i n E n g l i s h . This i s t r ue because f o r one th i n g , we can't p l a c e l i m i t s on the numbers of nested clauses i n a given sentence; and i f f o r any such sentence as 1.1 or 1.2 there i s another, l a r g e r one, then there are an i n f i n i t y of sentences. But, i f there are an i n f i n i t y of grammatical sentences i n E n g l i s h , then we cannot have an adequate grammar f o r E n g l i s h which i s based on the c o n s t r u c t i o n of l i s t s of grammatical s t r u c t u r e s of sentences, but must regard a grammar as a set of r e c u r s i v e l y generative r u l e s . For a more d e t a i l e d account of t h i s p oint, see Chapter Two i n Chomsky, Topics i n the Theory'of Generative Grammar. Secondly, a generative grammar must pos t u l a t e two l e v e l s of grammatical s t r u c t u r e s : deep s t r u c t u r e and surface s t r u c t u r e . 3 I f we consider two sentences l i k e : 1.3 John persuaded the doctor to examine B i l l . 1.4 John expected the doctor to examine B i l l . we can see that, d e s p i t e an apparent s i m i l a r i t y of grammatical s t r u c t u r e , i n the f i r s t case 'doctor' i s an object of the verb 'persuade', whereas i n the second case 'doctor' i s not an ob j e c t of the verb 'expect'. This can be seen i f we p a s s i v i s e the embedded clauses so as to o b t a i n the sentences: 1.31 John persuaded B i l l to be examined by the doctor. 1.41 John expected B i l l to be examined by the doctor. We q u i c k l y r e a l i s e here that 1.3 and 1.31 are not synonymous whereas 1.4 and 1.41 are synonymous. This i s explained i f we p o s t u l a t e such s t r u c t u r e s as: 1.32 (John persuaded the d o c t o r ) ( t h e doctor w i l l examine B i l l ) 1.42 (John expected) (the doctor w i l l examine B i l l ) to u n d e r l i e 1.3 and 1.4 r e s p e c t i v e l y , and: 1.33 (John persuaded B i l l ) (the doctor w i l l examine B i l l ) 1.43 (John expected) (the doctor w i l l examine B i l l ) to u n d e r l i e 1.31 and 1.41 r e s p e c t i v e l y . The explanation i s given i n the i d e n t i t y of 1.42 and 1.43 and the n o n - i d e n t i t y of 1.32 and 1.33, corresponding to the synonymy of 1.4 and 1.41 and the non-synonymy of 1.3 and 1.31. The underlying s t r u c t u r e s 1.32, 1.42, 1.33, and 1.43 can be c a l l e d deep s t r u c t u r e s . (For f u r t h e r d i s c u s s i o n , see Chapter Twenty Four i n Jacobs and Rosenbaum, E n g l i s h Transformational Grammar; see a l s o page twenty two i n Chomsky, Aspects of the 4 Theory of Syntax.) P o s t u l a t i n g a l e v e l of deep s t r u c t u r e means that we must a l s o p o s t u l a t e a s u r f a c e l e v e l of s t r u c t u r e i n order to provide a l e v e l of s t r u c t u r e to determine the p h o n o l o g i c a l i n t e r p r e t a t i o n of a sentence; c l e a r l y , deep and surface s t r u c t u r e , as the example j u s t given i l l u s t r a t e s , do not always c o i n c i d e . The surface l e v e l i s c a l l e d s urface s t r u c t u r e . The f u l l grammatical d e s c r i p t i o n of a sentence, then, includes (among other things) both deep and s u r f a c e s t r u c t u r e s . This s t i l l leaves open the question of how the deep and surface s t r u c t u r e s of a p a r t i c u l a r sentence w i l l be r e l a t e d to each other. Now corresponding to the d i s t i n c t i o n between deep and surface s t r u c t u r e , generative r u l e s of a grammar are of two kinds: base r u l e s and t r a n s f o r m a t i o n a l r u l e s . The base r u l e s , or ' c a t e g o r i a l ' r u l e s generate deep s t r u c t u r e s , and the t r a n s f o r m a t i o n a l r u l e s transform deep s t r u c t u r e s i n t o surface s t r u c t u r e s . i n what follows I w i l l be concerned almost e x c l u s i v e l y with the base r u l e s , s i n c e i t i s the base r u l e s which co n t a i n the r e c u r s i v e property to be d e s c r i b e d _ s h o r t l y . Base r u l e s are of the f o l l o w i n g form: 1 - 5 C p > C q + C r + • • • + C x where Cp to C x are grammatical categories l i k e 'sentence 1, 'noun', 'noun phrase', 'verb phrase', ' a r t i c l e ' , 'verbal', and the arrow i s an i n s t r u c t i o n to rewrite the s t r i n g of categories on the 5 r i g h t hand s i d e i n place of the s i n g l e category on the l e f t hand s i d e . The dots i n 1.5 may be f i l l e d i n with a p o s s i b l y n u l l s t r i n g of c a t e g o r i e s . Thus, base r u l e s may be c a l l e d c a t e g o r i a l r e w r i t e ru1es. For example: 1.6 S entence— —^ Noun phrase + Verb phrase Noun phrase — ^ A r t i c l e + Noun phrase Verb phrase* } Verb + Noun phrase i s a set of base or c a t e g o r i a l r e w r i t e r u l e s . Through a sequence of a p p l i c a t i o n s of these r u l e s , a s t r u c t u r e l i k e : 1.7 ( A r t i c l e + Noun) + (Verb + ( A r t i c l e + Noun)) might be generated. Such a generative process i s c a l l e d a d e r i v a t i o n . More e x p l i c i t l y , a d e r i v a t i o n of a s t r i n g i s an ordered set of r e w r i t e r u l e s . Not any ordered set, however, can c o n s t i t u t e a d e r i v a t i o n ; to s t a t e the two important r e s t r i c t i o n s , the l e f t hand s i d e of each nth r u l e must con t a i n a category which appears on the r i g h t hand s i d e of some mth r u l e f o r a l l ( i n t e g r a l and p o s i t i v e ) values of n greater than 1, and m l e s s than n; second, the ordered set must contain s p e c i f i c a t i o n s as to which l i n e s are to be regarded as s e q u e n t i a l p a i r s . These r e s t r i c t i o n s make d e r i v a t i o n s best represented by t r e e diagrams; i n the r e l a t i o n s h i p between nodes and branches, a t r e e diagram can represent the complete d e r i v a t i o n a l h i s t o r y of a given deep s t r u c t u r e type. A t r e e diagram for the d e r i v a t i o n of 1.7 from the s e t of r u l e s 1.6, f o r example, might look as f o l l o w s : 6 1.8 Sentence ^ \ Noun phrase Verb phrase / \ y \ a r t i c l e Noun Verb Noun,phrase a r t i c l e ^Noun And represented as an ordered set of r e w r i t e r u l e s , i t might be w r i t t e n as f o l l o w s : 1. Sentence > N o u n p h r a s e + V e r b p h r a s e 2. Noun phrase A r t i c l e •+ N o u n (1) 3 . Verb phrase » Verb + Noun p h r a s e (1) 4. Noun phrase > A r t i c l e + Noun (3) In 1.8athe numbers i n parentheses o n t h e r i g h t r e f e r t o t h e previous r e w r i t e r u l e s , thus i n d i c a t i n g t h e s e q u e n t i a l p a i r s , and corresponding to the second r e s t r i c t i o n r e f e r r e d t o a b o v e . . Comparison of 1.8 and 1.8a m a k e s t h e f u n c t i o n o f t h e s e p a r e -n t h e s i s e d numbers c l e a r ; i n a d d i t i o n , i t m a k e s i t c l e a r t h a t t r e e diagrams are p r e f e r a b l e over o r d e r e d s e t s o f r e w r i t e r u l e s as repres e n t a t i o n s of d e r i v a t i o n s f r o m t h e p o i n t o f view of graphics although the two kinds o f r e p r e s e n t a t i o n s c a n s e r v e the same f u n c t i o n at the a b s t r a c t o r m a t h e m a t i c a l l e v e l , . T o make the set of rewrite r u l e s p e r f e c t l y c o r r e s p o n d t o t h e t r e e diagram, more n o t a t i o n a l devices w o u l d h a v e t o b e e m p l o y e d f o r the ordered s e t s . We need not discuss t h e s e p r o b l e m s h e r e ' as henceforth d e r i v a t i o n s w i l l b e represented by t r e e diagrams. The base r u l e s , then, generate s t r i n g s o f c a t e g o r i e s i n t o which l e x i c a l items can be i n s e r t e d . We s h a l l say t h a t t h e 7 c a t e g o r i a l or base r u l e s , by themselves, generate deep s t r u c t u r e types. Thus, u n d e r l y i n g the sentences: 1.9 The boy saw the dog. 1.10 The boy saw the c a t . i s the same deep s t r u c t u r e type, which might i n f a c t be represented by the deep s t r u c t u r e type given i n 1.8. I t should be noted that i n tr a n s f o r m a t i o n a l l i t e r a t u r e , the term 'deep s t r u c t u r e ' i s sometimes used i n the same way that 'deep s t r u c t u r e type* w i l l be used here, whereas on other occasions i n t r a n s f o r m a t i o n a l l i t e r a t u r e the term 'deep s t r u c t u r e ' r e f e r s to what I am c a l l i n g the 'deep s t r u c t u r e type' plus the l e x i c a l items. A deep s t r u c t u r e type with l e x i c a l items added can be represented by a t r e e diagram l i k e 1.11: 1.11 Sentence Noun phrase Verb phrase A r t i c l e Noun Verb Noun phrase / \ A r t i c l e Noun I I the boy saw the dog In t r a n s f o r m a t i o n a l l i t e r a t u r e both 1.8 and 1.11 would be s a i d to provide the deep s t r u c t u r e f o r the sentence 1.9, "The boy saw the dog." But f o r our purposes to c a l l 1.8 a 'deep s t r u c t u r e ' would be to use dangerously ambiguous terminology. I w i l l c a l l a s t r u c t u r e such as i s represented i n 1.8 a 'deep s t r u c t u r e type', and a s t r u c t u r e such as i s represented i n 1.11 8 a 'deep s t r u c t u r e ' . This d i s t i n c t i o n i s necessary here because i n Chapter Two i t i s claimed that a l l our thoughts can be expressed using a f i n i t e number of deep s t r u c t u r e types, but s i n c e Chapter Two makes no claims about the minimum s i z e of an adequate l e x i c o n , i t cannot c l a i m that we can express a l l our thoughts using only a f i n i t e number of deep s t r u c t u r e s f o r sentences. I continue now to d i s t i n g u i s h between r e c u r s i v e and non-r e c u r s i v e sets of base r u l e s . In the course of t h i s chapter, two d e f i n i t i o n s f o r 're c u r s i v e f o r a s e t of base r u l e s ' w i l l be given — a general one, and a s p e c i f i c f o r m u l a t i o n . The more important of the two i s the general one. A" set of base r u l e s i s r e c u r s i v e i f and only i f i t i s capable o f generating an i n f i n i t y of deep s t r u c t u r e types. L a t e r a s p e c i f i c , or ' c l e r i c a l ' d e f i n i t i o n w i l l be presented, and some problems surrounding i t discussed; i n the meanwhile i t w i l l be u s e f u l to consider some examples of r e c u r s i v e and non-recursive sets of base r u l e s . The set of base r u l e s c o n s i s t i n g of the one member. 1.12 Sentence- • ^ Noun + verb i s non-recursive. This base r u l e i s capable of generating only the one deep s t r u c t u r e type: 1.13 Sentence 9 Note, though, that i f t h i s base belonged to a language which had an i n f i n i t e vocabulary of nouns or verbs, or both, then the one base r u l e , given appropriate other r u l e s f o r l e x i c a l i n s e r t i o n , etc., would be s u f f i c i e n t to generate an i n f i n i t e number of d i f f e r e n t sentences. Yet the source of the i n f i n i t y would not be the base r u l e ; each of these sentences would have the same deep s t r u c t u r e type. On the other hand, the set of base r u l e s : 1.14a S e n t e n c e — } Sentence + Sentence which might be used i n the d e r i v a t i o n of sentences which contain two or more independent clauses, can generate an i n f i n i t y of deep s t r u c t u r e types. 1.15 i s an example of one of the deep s t r u c t u r e types which 1.14 can generate. Noun Verb Noun Verb I t i s c l e a r t h at there are no l i m i t s to the number of deep s t r u c t u r e types which can be generated by a set of r u l e s l i k e 1.14. I f t h i s base r u l e belonged to the grammar of a language which had only a f i n i t e vocabulary, the grammar would s t i l l be able to generate an i n f i n i t y of d i f f e r e n t sentences, as the b. Sentence } Noun + Verb 1.15 Sentence L O set can generate an i n f i n i t y of deep s t r u c t u r e types. We have, then, an example of a r e c u r s i v e set of r u l e s . The f o l l o w i n g are f u r t h e r examples of r e c u r s i v e sets of base r u l e s : 1.16 Noun phrase ^ Sentence Sentence ) Noun phrase + Sentence 1.17 Verb phrase — ^ Sentence Sentence- ^ Noun phrase + Verb phrase Also, there could be r e c u r s i v e sets of r u l e s with many r u l e s i n each s e t . The f o l l o w i n g i s a r e c u r s i v e s e t with four r u l e s : (C]_...Cg are grammatical categories.) 1.18 C x- -> C 2 + C 3 c 3 — ~ * c 4 + c 5 c 5 > c 6 + c 7 c 7 — - » c 8 + C n F i n a l l y , here i s an example of a deep s t r u c t u r e type generated by one of these r e c u r s i v e sets of base r u l e s , namely 1.17: 1.19 Sentence Noun phrase Verb phrase sentence Noun phrase Verb phrase I sentence Noun''phrase Verb phrase In a base r u l e there i s only one category on the l e f t ; t h i s category can be s a i d to dominate the category or categories on the r i g h t hand sid e of the rewrite r u l e . Domination i s t r a n s i t i v e , 11 so that i f (as i n 1.18 above,) C T dominates C 3 and C 3 dominates C 5 , then C]_ dominates C 5 . I can now give a second d e f i n i t i o n of ' r e c u r s i v e ' f o r a set of base r u l e s . A set of base r u l e s (comprising one or more members) i s r e c u r s i v e i f and only i f at l e a s t one of the categories contained i n the r u l e s dominates i t s e l f . I t should be mentioned that t h i s i s only one of the p o s s i b l e s p e c i f i c , o r ' c l e r i c a l ' formulations of ' r e c u r s i v e ' f o r base r u l e s . One could a l s o , f o r example, formulate base r u l e s i n which there i s no l i m i t to the number of categories on the r i g h t hand s i d e of the r u l e s . Thus: 1.20 Sentence — ——$ Sentence -1- + Sentence might be an example of such a r u l e . I ' l l c a l l such a r u l e with an i n d e f i n i t e number of categories on the r i g h t hand s i d e an i n d e f i n i t e r e c u r s i v e r u l e . (Langendoen i n The Study of Syntax, pages 31 and 32 c a l l s i t a 'rule schema'.) The advantage of using an i n d e f i n i t e r e c u r s i v e r u l e would be to avoid unnecessary and inelegant m u l t i p l i c a t i o n of nodes i n some deep s t r u c t u r e t r e e r e p r e s e n t a t i o n s . Using the d e f i n i t e r e c u r s i v e r u l e 1.14a: 1.14a Sentence — ^ Sentence + Sentence one could generate deep s t r u c t u r e types f o r complicated s e q u e n t i a l conjunctions; but only by using a l a r g e number of nodes. To generate a deep s t r u c t u r e type f o r a sentence l i k e : 1.21 Mary saw B i l l , and Tom saw Mark, and Harry came home, and John was l a t e . a deep s t r u c t u r e type l i k e the f o l l o w i n g would be r e q u i r e d : 12 1.22 Sentence Sentence Sentence .^entence Sentz Sentence Sentenc Sen ence Sentence The second t i e r of nodes i s in e l e g a n t as the b i n a r y d i v i s i o n does not seem warranted by the sentence s t r u c t u r e . Using an i n d e f i n i t e r e c u r s i v e r u l e such as 1.20: 1.20 Sentence — ^ Sentence + ....+ Sentence a deep s t r u c t u r e type such as the f o l l o w i n g could be generated to b e t t e r represent the clause s t r u c t u r e of the same sentence: 1.23 Sentence Sentence Sentence Sentence Sentence Since i t i s not a necessary feature of an i n d e f i n i t e r e c u r s i v e r u l e that the l e f t hand sid e c o n t a i n a category which dominates i t s e l f ( i . e . , i n t h i s case, which i s i d e n t i c a l with a category on the r i g h t hand s i d e ) , t e c h n i c a l l y speaking, i n d e f i n i t e r e c u r s i v e r u l e s do not f a l l under the second d e f i n i t i o n of r e c u r s i v e sets of r u l e s . However, f o r a l l p r a c t i c a l purposes, i n d e f i n i t e r e c u r s i v e r u l e s such as: ! . 2 4 C m - V C n + + C n where C n and C m are d i s t i n c t grammatical c a t e g o r i e s , do not appear to be r e q u i r e d . Moreover, such i n d e f i n i t e r e c u r s i v e r u l e s with no category domination can be r e w r i t t e n i n the form of a set of (usually two) r u l e s which f a l l under the second d e f i n i t i o n f o r 'recursive' set of r u l e s provided above. 13 Thus an i n d e f i n i t e r e c u r s i v e r u l e such as: 1.25 Noun phrase • • ^ Sentence + ...... + Sentence could be r e w r i t t e n as: 1.25a Noun phrase ) Sentence b Sentence ^ Sentence + ........ + Sentence or i n general, a r u l e of the form: 1.24 C — > C + + C m n n could Tbe r e w r i t t e n as a s e t of r u l e s of the form: 1.24a C m > C n b C n ^ C n + .......... + C n Thus, the in e l e g a n c i e s created by the d e f i n i t e r e c u r s i v e r u l e f o r s e n t e n t i a l conjunction (Sentence y Sentence + Sentence) would be avoided, while at the same time, the c l e r i c a l d e f i n i t i o n f o r r e c u r s i v e sets of rewr i t e r u l e s could be kept u n i f i e d through the concept of category domination. The second d e f i n i t i o n of r e c u r s i v e f o r sets of re w r i t e r u l e s given above w i l l , then, be adequate f o r our purposes. J u s t f o r the sake of thoroughness, though, consider the question, could we ever need r e c u r s i v e r u l e s with more than one category type on the r i g h t hand side? Imagine that we had a r u l e l i k e : 1.26 Sentence > Noun phrase + Noun + ...... + Verb where the dots could be f i l l e d i n by any number of categories drawn from a c l a s s of categories none of which i s 'Sentence', the category on the l e f t hand s i d e . The r u l e would s t i l l be 14 r e c u r s i v e under the f i r s t d e f i n i t i o n , s i n c e i t could be used to generate i n f i n i t e l y many deep s t r u c t u r e types. But i t would not be r e c u r s i v e under the second d e f i n i t i o n , s i n c e no category dominates i t s e l f . In general such 'doubly i n d e f i n i t e ' r e c u r s i v e r u l e s might be represented as: ! . 2 7 C m • C x + + C y where C m i s a category, C x and Cy d e f i n e some p o s s i b l y n u l l set of c a t e g o r i e s , and the dots can be f i l l e d i n with any number of c a t e g o r i e s drawn from a set of categories C ...C J = n q such that C m i s not a member of that s e t . I f anyone were to present c o n s i d e r a t i o n s suggesting that r u l e s of the form given i n 1.27 might be necessary f o r the d e s c r i p t i o n of our languages, the second d e f i n i t i o n of r e c u r s i v e would have to be r e v i s e d . But, f i r s t l y , i t has never been suggested that r u l e s of such i n d e f i n i t e n e s s are necessary f o r the d e s c r i p t i o n of our languages; and secondly, they are so u n s p e c i f i c that i t i s d i f f i c u l t to e n t e r t a i n the n o t i o n of t h e i r being rewrite r u l e s of a grammar. The more open the r u l e s , the l e s s d e s c r i p t i v e content they c o n t a i n . For example, i n the case where C x and Cy are n u l l , and C n...Cg contains a l l categories i n the grammar except C m, then a r u l e of the form given i n 1.27 becomes a r u l e which s u f f i c e s by i t s e l f to generate any deep s t r u c t u r e type fo r any sentence of the language. In general, i t i s implausible to suggest that 15 a grammarian would be s a t i s f i e d that there are no more d e f i n i t e u n d e r l y i n g r u l e s f o r the language being d e s c r i b e d . To put i t another way, as rul e s become more inde f i n i t e , , they have le s s d e s c r i p t i v e value; the more d e f i n i t e they are, the more they can be r e w r i t t e n i n the form of sets of r e c u r s i v e r u l e s with category domination. These sp e c u l a t i o n s on the permissable i n d e f i n i t e n e s s of r e w r i t e r u l e s have some i n t r i n s i c i n t e r e s t , but c o n s i d e r i n g that no grammarians have ever considered doubly i n d e f i n i t e r e w r i t e r u l e s to be u s e f u l — l e t alone necessary -- f o r the d e s c r i p t i o n of our languages, the study of the permissable i n d e f i n i t e n e s s of a grammatical r u l e from a p u r e l y t h e o r e t i c a l p o i n t of view i s perhaps best set aside f o r f u t u r e i n v e s t i g a t i o n . So f a r , then, we have base c a t e g o r i a l r e w r i t e r u l e s with which s t r i n g s of c a t e g o r i e s , or deep s t r u c t u r e types, may be der i v e d . The deep s t r u c t u r e types can have l e x i c a l items i n s e r t e d so as to y i e l d deep s t r u c t u r e s . The deep s t r u c t u r e s may be transformed i n t o surface s t r u c t u r e s by means of t r a n s -formational r u l e s . Furthermore, we have a d i s t i n c t i o n between r e c u r s i v e and non-recursive sets of base r u l e s . Now the notion of ' d e r i v a t i o n ' as i t a p p l i e s to sentences or deep s t r u c t u r e s or deep s t r u c t u r e types i s f a i r l y b a s i c i n transformational l i t e r a t u r e (see Chomsky, Aspects, pages 65,66). Less b a s i c , and therefore sometimes undefined, i s the term ' d e r i v a t i o n ' as i t app l i e s to a co n s t i t u e n t of a sentence or 16 deep s t r u c t u r e or deep s t r u c t u r e type. The term ' d e r i v a t i o n of a c o n s t i t u e n t ' i s necessary i n t h i s study as the concept of ' r e c u r s i v e l y d e r i v e d c o n s t i t u e n t ' plays an important r o l e l a t e r on. The d e r i v a t i o n of a c o n s t i t u e n t C m i n the deep s t r u c t u r e type DS m i s d e f i n e d here as the ordered s e r i e s of categories Cp, Cg, C r, C m, such that i n the d e r i v a t i o n of DS m, Cp i s the i n i t i a l category ( i . e . the category on the l e f t hand s i d e of the f i r s t r e w r i t e r u l e used), Cp immediately dominates Cg, and i n general each nth entry i n the set immediately dominates the n + 1st entry, and C m i s the l a s t category i n the s e t . (Notice that c o n s t i t u e n t s are being r e f e r r e d to here by i n d i v i d u a l c a t e g o r i e s ; t h i s i s i n accordance with usual custom.) For example, i n the d e r i v a t i o n of: 1.11 Sentence. Noun phrase a r t i c l e * " * ^ " " noun J J the boy Verb phrase verb saw noun phrase a r t i c l e n t i e oun dog the d e r i v a t i o n of the noun c o n s t i t u e n t 'boy' i s Sentence, Noun phrase, Noun, and the d e r i v a t i o n of the noun phrase c o n s t i t u e n t 'the boy' i s : Sentence, Noun phrase. Based on t h i s d e f i n i t i o n , we can construct a d e f i n i t i o n 17 f o r ' r e c u r s i v e l y derived c o n s t i t u e n t ' as i t a p p l i e s to con-s t i t u e n t s i n deep s t r u c t u r e . A c o n s t i t u e n t C m of a deep s t r u c t u r e type DS m i s r e c u r s i v e l y derived i f and only i f some category C x appears more than once i n the ordered category s e r i e s Cp...C m which c o n s t i t u t e s the d e r i v a t i o n of the con-s t i t u e n t C m. The r a t i o n a l e here should be f a i r l y c l e a r : i f some category C x appears more than once i n the d e r i v a t i o n of a c o n s t i t u e n t , then there i s a category which dominates i t s e l f i n the d e r i v a t i o n of C m, and the set of r u l e s used i n the d e r i v a t i o n of DS m must be a r e c u r s i v e set; i t s d e r i v a t i o n must in v o l v e a p p l i c a t i o n of a r e c u r s i v e set of r u l e s , and so i t may be c a l l e d a r e c u r s i v e l y derived c o n s t i t u e n t . (For f u r t h e r d i s c u s s i o n see Chomsky, On the Notion "Rule of Grammar", s e c t i o n 2 . ) We now have the terms ' d e r i v a t i o n of a c o n s t i t u e n t C m i n deep s t r u c t u r e type DS m' and 'a c o n s t i t u e n t C m of deep s t r u c t u r e type DS m such that C m i s r e c u r s i v e l y d e r i v e d ' . I t would be convenient at t h i s p o i n t i f i t were easy to provide a d e f i n i t i o n f o r 'some c o n s t i t u e n t C m of surface s t r u c t u r e type S S m such that C i s r e c u r s i v e l y d e r i v e d . ' But i t i s not easy i n general m J to provide c r i t e r i a f o r i d e n t i f y i n g c o n s t i t u e n t s through the transformation from deep to surface s t r u c t u r e . Given a p a r t i c u l a r noun phrase, f o r example, as i t occurs i n some surface s t r u c t u r e , i t might be d i f f i c u l t to i d e n t i f y that p a r t i c u l a r noun phrase i n deep s t r u c t u r e s i n c e the s t r u c t u r e s undergo var i o u s changes i n the t r a n s i t i o n from deep to s u r f a c e structure? sundry c o n s t i t u e n t s , a l l dominated by 'noun phrase' i n the surface s t r u c t u r e , may appear s c a t t e r e d under d i f f e r e n t c o n s t i t u e n t s (not n e c e s s a r i l y a l l 'noun phrase') i n the deep s t r u c t u r e . For t h i s reason, no attempt w i l l be made here to define ' d e r i v a t i o n of a c o n s t i t u e n t C m of surface s t r u c t u r e SS m'»for the purposes of t h i s study i t w i l l be s u f f i c i e n t to define 'recursive d e r i v a t i o n of a sentence': A sentence i s r e c u r s i v e l y d e rived i f and only i f i t s deep s t r u c t u r e contains some c o n s t i t u e n t C m which i s r e c u r s i v e l y d e r i v e d . Chapter Two attempts to show that any sentence which i s r e c u r s i v e l y d e r i v e d as def i n e d above can be paraphrased by a s e r i e s of sentences none of which are r e c u r s i v e l y d e r i v e d . O c c a s i o n a l l y , d e s p i t e not having defined ' r e c u r s i v e l y derived' f o r a c o n s t i t u e n t of surface s t r u c t u r e , I w i l l r e f e r to some constituent! as i d e n t i f i e d through surface s t r u c t u r e as a r e c u r s i v e l y ^ d e r i v e d c o n s t i t u e n t ; the i m p l i c a t i o n i s i n such cases that the surface c o n s t i t u e n t being i d e n t i f i e d as r e c u r s i v e l y derived corresponds f a i r l y d i r e c t l y with some r e c u r s i v e l y d e rived c o n s t i t u e n t i n deep s t r u c t u r e . The category dominating i t s e l f i n almost a l l r e c u r s i v e d e r i v a t i o n s i s the i n i t i a l , or 'sentence' category, and i n these cases the surface s t r u c t u r e s contain ' s e n t e n t i a l clauses' bearing strong l i n e s 19 of c o n t i n u i t y with s e n t e n t i a l clauses i n deep s t r u c t u r e , so that the problem of i d e n t i f y i n g c o n s t i t u e n t s through deep and surface s t r u c t u r e l e v e l s i s at a minimum i n the case of r e c u r s i v e l y d e r i v e d c o n s t i t u e n t s . More important, the term ' r e c u r s i v e l y d e r i v e d sentence' i s the key one, and i t s d e f i n i t i o n does not depend on the d e f i n i t i o n of d e r i v a t i o n of a c o n s t i t u e n t of surface s t r u c t u r e . Once again, the term ' r e c u r s i v e l y derived sentence' i s the key one because the attempt of Chapter Two i s to show that any r e c u r s i v e l y derived sentence as defined above can be paraphrased by a s e r i e s of sentences none of which i s a r e c u r s i v e l y derived sentence. Two things must be further mentioned with regard to the d e f i n i t i o n , the f i r s t t e r m i n o l o g i c a l , the second explanatory. F i r s t , the term ' r e c u r s i v e l y derived sentence* w i l l be used i n t h i s study i n j u s t t h i s way, to r e f e r to sentences with r e c u r s i v e l y d e r ived c o n s t i t u e n t s , as defined above. In some broad context however, one might want to think of a sentence l i k e 1.28 B r a k t i l o s t i p a l t i m a n i s a h i l e x returned l a t e , as being r e c u r s i v e l y generated since i t contains a proper name, and there are no p r i n c i p l e d l i m i t s to the number of proper names i n a l e x i c o n . (For that matter, i t ' s not c l e a r that there are any p r i n c i p l e d l i m i t s to the number of verbs i n a l e x i c o n e i t h e r . But our i n i t i a l i n t e r e s t jn t h i s study i s on the semantic e f f e c t s of the recursions which Chomsky postulates and discusses; 20 Chomsky does not discuss the question of the i n f i n i t y of l e x i c a l items p o s s i b l e , and obviously, a proper name i n a sentence l i k e 1.28 would notbea r e c u r s i v e l y derived c o n s t i t u e n t as def i n e d above. Of course, some people might dispute whether sentences l i k e 1.28 are p r o p e r l y regarded as ' r e c u r s i v e l y generated'. I should add, then, that my point here i s not to advance the c l a i m that they are, but simply to make i t c l e a r that i n making s t a t e -ments about " a l l r e c u r s i v e l y derived sentences" (e.g., to a n t i c i p a t e the t h e s i s of Chapter Two, " a l l r e c u r s i v e l y d e r i v e d sentences are e l i m i n a b l e i n paraphrase") I do not regard myself as ipso f a c t o making claims about a l l r e c u r s i v e l y generated sentences. To put t h i s s t i l l another way, 'generation' of sentences i n t h i s study remains a more a b s t r a c t notion than ' d e r i v a t i o n ' of sentences, the l a t t e r being a s s o c i a t e d with the s t r i c t l y d e f i n e d d e r i v a t i o n a l procedure w i t h i n the framework developed i n , say, Aspects, or On the Notion 'Rule of Grammar'. Second, s i n c e Topics and Aspects, there are only s i n g u l a r y transformations, i . e . transformations operating on i n d i v i d u a l deep s t r u c t u r e s , so that r e c u r s i o n i s introduced only by the rewrite r u l e s ; t h i s makes i t p o s s i b l e f o r us to define the presence of a r e c u r s i v e l y derived c o n s t i t u e n t i n the deep s t r u c t u r e as being necessary -- as w e l l as s u f f i c i e n t — f o r a sentence's being ' r e c u r s i v e l y derived'. For f u r t h e r explanation of t h i s , see Aspects pages 132 - 136, Topics pages 63 - 67. 2 1 Also, but more marginally, the f o l l o w i n g ambiguity i s worthy of mention: the term ' r e c u r s i v e l y derived c o n s t i t u e n t ' can be used to r e f e r both to p a r t i c u l a r c o n s t i t u e n t s and a l s o to types of c o n s t i t u e n t s . Thus, with regard to the sentence: 1.29 John b e l i e v e s that Harry returned, we could say that the noun clause, or noun phrase complement, "that Harry returned" i s a r e c u r s i v e l y derived c o n s t i t u e n t . We can a l s o say, speaking of noun clauses or noun phrase complements i n general, that- noun clauses or noun phrase complements are. r e c u r s i v e l y derived c o n s t i t u e n t s . For the purposes of t h i s study I b e l i e v e that t h i s ambiguity i s not dangerous. To say t h a t the noun clause 'that Harry returned' i s a r e c u r s i v e l y derived c o n s t i t u e n t i s to say that i n the d e r i v a t i o n of the noun phrase c o n s t i t u e n t some category appears twice (here 'Sentence'). To say that noun clauses i n general are r e c u r s i v e l y d e r i v e d c o n s t i t u e n t s i s to say that i n the d e r i v a t i o n of any noun clause ( i . e . , noun phrase which dominates a 'sentence' constituent) some category w i l l appear twice (here 'Sentence', and f o r obvious reasons as well.) To summarize the h i g h l i g h t s of what has been s a i d so f a r , a grammar of a language i s a set of generative r u l e s . i t must have two l e v e l s of grammatical s t r u c t u r e , deep and s u r f a c e . Deep s t r u c t u r e types are generated by base rewrite r u l e s , or by c a t e g o r i a l rewrite r u l e s ; and deep s t r u c t u r e types with l e x i c a l items i n s e r t e d are c a l l e d deep s t r u c t u r e s . The deep 22 s t r u c t u r e s are transformed i n t o surface s t r u c t u r e s by means of trans formational r u l e s . An ordered set of rewri t e r u l e s as representable i n tre e diagrams c o n s t i t u t e s or represents the d e r i v a t i o n of a deep s t r u c t u r e or deep s t r u c t u r e type. The sequence of category dominations ending i n the c o n s t i t u e n t C n and beginning with the i n i t i a l category i s the d e r i v a t i o n of Ci;, i n a deep s t r u c t u r e type. I f a category i n a d e r i v a t i o n of C n appears twice then C n i s a r e c u r s i v e l y derived constituent„ I f a sentence contains a r e c u r s i v e l y derived c o n s t i t u e n t then the sentence i s a recurs i v e l y derived sentence. Up u n t i l t h i s p o int, no mention has been made of the semantic and pho n o l o g i c a l components of a grammar. In t h i s study I am concerned with the r e l a t i o n s h i p between s y n t a c t i c devices and the expressive c a p a c i t y of languages, and so the phonolog i c a l component i s not of primary i n t e r e s t . The semantic component, on the other hand, i s of primary i n t e r e s t . However, I w i l l make no attempt to review the recent semantic th e o r i e s of t r a n s f o r m a t i o n a l grammarians; the only notion of d i r e c t relevance t o t h i s study i s that of "semantic i n t e r p r e t a t i o n 1 . The base r u l e s generate deep s t r u c t u r e types, which r e c e i v e l e x i c a l items, and are then c a l l e d deep s t r u c t u r e s . The deep s t r u c t u r e s are transformed i n t o surface s t r u c t u r e s . Surface s t r u c t u r e s r e c e i v e phonological i n t e r p r e t a t i o n , and the deep str u c t u r e s wholly, (with rare exceptions) determine semantic i n t e r p r e t a t i o n . J u s t i n what terms, or i n what way, the notion 23 of semantic i n t e r p r e t a t i o n can be f i l l e d i n , Chomsky does not make very c l e a r - From the p o i n t of view of the scope of t h i s study, the controversy about whether deep s t r u c t u r e wholly, with no exceptions, determines semantic i n t e r p r e t a t i o n , or whether surface s t r u c t u r e plays a r o l e i n shaping semantic i n t e r p r e t a t i o n i s of l i t t l e or no concern. Neither i s i t necessary here to discuss the s p e c i f i c c o n t r i b u t i o n s to the f i e l d of semantics by t r a n s f o r m a t i o n a l grammarians. The only t h i n g that counts at t h i s p o i n t i s that the s y n t a c t i c component generates s t r u c t u r e s which determine semantic i n t e r p r e t a t i o n . I turn now to discuss Chomsky's theory of the r e l a t i o n s h i p between the study of d e s c r i p t i v e l i n g u i s t i c s and our under-standing of i n t e l l e c t u a l processes. B. Some Problems i n Generative L i n g u i s t i c s . On many occasions Chomsky expresses the hope that the study of l i n g u i s t i c s w i l l i l l u m i n a t e and therefore enable us to understand b e t t e r the nature of the human mind. In a r a t h e r general way i t might be taken f o r granted that the study of any aspect of human language use can be of value f o r our under-standing of the nature of the human mind. But more than t h i s i s intended. Chomsky's lectures" published under the t i t l e Language and Mind are s p e c i f i c a l l y addressed to the question, "What c o n t r i b u t i o n can the study of language make to our understanding of human nature?" (L & M, page 1) During the course of these l e c t u r e s he says, r e f e r r i n g to the possession 24 of language (L & M page 62): . . . i t seems to me that today there i s no b e t t e r or more promising way to explore the e s s e n t i a l and d i s t i n c t i v e p r o p e r t i e s of human i n t e l l i g e n c e than through the d e t a i l e d i n v e s t i g a t i o n of the s t r u c t u r e of t h i s unique human pos s e s s i o n . The view that the study of language w i t h i n the general framework of transformational l i n g u i s t i c s may c o n t r i b u t e to our under-standing of the d i s t i n c t i v e p r o p e r t i e s of human i n t e l l e c t u a l pr-ocesses i s , of course, shared by many working w i t h i n that framework. Further, A l l e n and Van Buren say, speaking i n general of the goals of generative t r a n s f o r m a t i o n a l l i n g u i s t i c s , i n t h e i r i n t r o d u c t i o n to some readings from the works of Chomsky, ( S e l . Read. page v i i i ) : I t must be c l e a r l y understood that current work i n generative l i n g u i s t i c s i s e s s e n t i a l l y t h e o r e t i c a l i n nature. I t i s not motivated by an i n t e r e s t i n computers or any kind of engineering, nor does i t c o n s t i t u t e an attempt to propogate a new and obscure branch of mathe-* matics . The main purpose of the research i s to suggest an explanatory hypothesis concerning the nature of language and u l t i m a t e l y of human thought. Further, the theory of language presented i n t h i s book i s 'quite e x p l i c i t l y and s e l f - c o n s c i o u s l y m e n t a l i s t i c ' ... i n the sense that a theory of language i s to be regarded as a p a r t i a l theory of the human mind. The concern then, i s to give us an understanding of human i n t e l l e c t u a l processes not only i n a rather general way, simply i n s o f a r as we use a very complex system of l i n g u i s t i c communication, but a l s o i n s o f a r as the study of language can contribut e to our understanding of the very s p e c i a l k i n d of mental processes which humans are capable of manifesting. And according to the theory of generative transformational l i n g u i s t i c s , the study of l i n g u i s t i c s may be able to provide 25 us with an explanatory hypothesis concerning the nature and s t r u c t u r e of these d i s t i n c t i v e processes. Moreover, many tra n s f o r m a t i o n a l l i n g u i s t s consider t h e i r study t o be a r a t i o n a l i s t , or 'Cartesian' l i n g u i s t i c s . They do so (as expounded by Chomsky) f o r two p r i n c i p a l reasons: the r a t i o n a l i s t s observed that our o r d i n a r y use of language gives evidence of r a t i o n a l i t y and i n t e l l i g e n c e of a kind markedly set apart from any i n t e l l i g e n c e manifested i n the behaviour of a l l other known species of animals; and they took upon themselves the task of p r o v i d i n g an explanatory hypo-t h e s i s concerning the p o s s i b i l i t y of there being such a d i s t i n c t i o n . Likewise, as has j u s t been observed, p r o v i d i n g an explanatory hypothesis concerning the c r e a t i v i t y of the human i n t e l l e c t with respect to the use of language i s the c e n t r a l goal of generative t r a n s f o r m a t i o n a l l i n g u i s t i c s , as expounded by Noam Chomsky, J . Fodor, J e r r o l d Katz, and other prominent t h e o r i s t s of tra n s f o r m a t i o n a l l i n g u i s t i c s . To t h i s extent, generative t r a n s -formational l i n g u i s t i c s might be described as r a t i o n a l i s t i n i t s a s p i r a t i o n : i t shares the explanatory goals of C a r t e s i a n r a t i o n a l i s m with regard to the use of language/ and c e r t a i n l y would consider these goals to be the appropriate ones f o r any p r o p e r l y conceived l i n g u i s t i c theory, that i s , f o r any l i n g u i s t i c theory which does not a r t i f i c i a l l y and a r b i t r a r i l y set i t s s i g h t s low. Secondly, l i k e Descartes (or apparently l i k e Descartes), 26 Chomsky, and other generative t r a n s f o r m a t i o n a l t h e o r i s t s b e l i e v e that the explanatory account w i l l come through the hypothesis of c e r t a i n 'innate p r i n c i p l e s ' of mental s t r u c t u r e . However here the r a t i o n a l i s m must be a reconstructed one. The p o s t u l a t e of a ' c r e a t i v e p r i n c i p l e ' of 'mind' f o r Chomsky i s not a s u b s t a n t i a l explanatory p o s t u l a t e ; the p o s t u l a t e of a U n i v e r s a l Grammar — an innate language determining s t r u c t u r e , however, i s . On Chomsky's account, a d i s t i n c t i o n i s recognized between the o r i g i n a l C a r t e s i a n p o s t u l a t e and the modern one — the o r i g i n a l p o s t u l a t e l e a v i n g no clues as to d e t a i l , the modern one s e t t i n g the stage f o r a d e t a i l e d proposal -- and at t h i s p o i n t an important h i s t o r i c a l controversy gets o f f the ground. Chomsky i s r e c o n s t r u c t i n g r a t i o n a l i s m , and, i n part, changing the explanatory v e h i c l e ; i s i t s t i l l a r a t i o n -9 a l i s t as opposed to an e m p i r i c i s t programme? Chomsky b e l i e v e s t h at f o r h e u r i s t i c purposes a d i s t i n c t i o n of t h i s k i n d i s i n order, t h i n k i n g no doubt of the h i s t o r i c a l l i n e s which l i n k the o r i g i n a l e m p i r i c i s t s with h a r d - b o i l e d psycholo-g i s t s of S k i n n e r i a n persuasions. However, t h i s i s almost e x c l u s i v e l y a h i s t o r i c a l matter, and one which l i e s wholly outside the scope of t h i s study. The p o i n t of departure f o r this" study i s the question: to what extent does Chomsky say that the d e t a i l s of generative transformational l i n g u i s t i c theory as they are c u r r e n t l y being developed have begun to provide us with the d e s i r e d exolanation? 27 As phrased, the question i s simply an i n q u i r y i n t o the t e x t , and c o n s i d e r i n g the a l l e g e d c e n t r a l i t y of the subject, one would l i k e to imagine that, whatever allowances f o r te n t a t i v e n e s s one would be i n c l i n e d to make on the grounds of 'theory i n progress', at l e a s t the b a s i c d r i f t of the answer to t h i s question has been i n d i c a t e d . Unfortunately, the connection between progress i n d e s c r i p t i v e l i n g u i s t i c s and progress towards the c o n s t r u c t i o n of a p s y c h o l o g i c a l account of the l i n g u i s t i c f a c u l t y has been inadequately t r e a t e d . In Aspects of the Theory of Syntax Chomsky wr i t e s , (page 6): Within t r a d i t i o n a l l i n g u i s t i c theory... i t was c l e a r l y understood that one of the q u a l i t i e s that a l l languages have i n common i s t h e i r ' c r e a t i v e ' aspect. Thus an e s s e n t i a l property of language i s that i t provides the means f o r expressing i n d e f i n i t e l y many thoughts and f o r r e a c t i n g a p p r o p r i a t e l y i n an i n d e f i n i t e range of new s i t u a t i o n s . The grammar of a p a r t i c u l a r language, then i s to be supplemented by a u n i v e r s a l grammar that accommodates the c r e a t i v e aspect of language use and expresses the deep-seated r e g u l a r i t i e s which, being u n i v e r s a l , are omitted from the grammar i t s e l f . He goes on to discuss the f a i l u r e of the t r a d i t i o n a l l i n g u i s t s to 'attempt a p r e c i s e statement of re g u l a r processes of sentence formation and sentence i n t e r p r e t a t i o n ' , s t a t i n g the fundamental reason for t h i s inadequacy of t r a d i t i o n a l grammars i n the f o l l o w i n g terms (Aspects, page 8 ) : Although i t was w e l l understood that l i n g u i s t i c processes are i n some sense ' c r e a t i v e ' , the t e c h n i c a l devices f o r expressing a system of r e c u r s i v e processes were simply not a v a i l a b l e u n t i l much more r e c e n t l y . i n f a c t , a r e a l understanding of how a language can (in Humboldt's 28 words) "make i n f i n i t e use of f i n i t e means" has developed only w i t h i n the l a s t t h i r t y years, i n the course of studies i n the foundations of mathematics. Now that these i n s i g h t s are r e a d i l y a v a i l a b l e , i t i s p o s s i b l e to return to the problems that were r a i s e d , but not solved, i n t r a d i t i o n a l l i n g u i s t i c theory, and to attempt an e x p l i c i t f ormulation of the ' c r e a t i v e ' processes of language. There i s , i n short, no longer a t e c h n i c a l b a r r i e r to the f u l l - s c a l e study of generative grammars. On the other hand, we f i n d the remarks, i n Language and Mind i n the c r u c i a l concluding passages, (L & M page 84): S t i l l , i n many respects we have not made the f i r s t approach to the r e a l answer to the c l a s s i c a l problems. For example, the c e n t r a l problems r e l a t i n g to the c r e a t i v e aspect of language use remain as i n a c c e s s i b l e as they have always been. This i s p u z z l i n g , as i t would appear that the ' c l a s s i c a l problems' are those of how to provide explanatory accounts of the c r e a t i v e workings of the mind with regard to language, and from the quoted passage from Aspects above, i t appears that recent work i n the e x p l i c i t formulation of s y n t a c t i c r e c u r s i v e processes f o r languages has an important c o n t r i b u t i o n to make i n t h i s r e s p e c t , Even more p u z z l i n g i s the statement j u s t a few sentences f u r t h e r on i n the co n c l u s i o n to Language & Mind (L & M page 84); Real progress had been made i n the study of the mechanisms of language, the formal p r i n c i p l e s that make p o s s i b l e the c r e a t i v e aspect of language use and that determine the phonetic form and semantic content of u t t e r a n c e s . From t h i s we can a b s t r a c t the statement that r e a l progress 29 has been made i n the study of the formal p r i n c i p l e s that make p o s s i b l e the c r e a t i v e aspect of language use. I f t h i s i s to be c o n s i s t e n t with the almost adjacent remark (quoted above) that "the c e n t r a l problems r e l a t i n g to the c r e a t i v e aspect of language use remain as i n a c c e s s i b l e as they have always been', then i t follows that the c e n t r a l problems r e l a t i n g to the c r e a t i v e aspect of language use cannot be to di s c o v e r the p r i n c i p l e s that make p o s s i b l e the c r e a t i v e aspect of language use — f o r a l l e g e d l y , r e a l progress has been made i n that area. However, i n A l l e n and Van Buren's e x p l i c i t and s u c c i n c t e x p o s i t i o n of the goals of generative transform-a t i o n a l l i n g u i s t i c s , we f i n d the f o l l o w i n g p r e s e n t a t i o n ( S e l . Read., page v i i i ) : . . . i n the context of generative l i n g u i s t i c s we are concerned with data not f o r i t s own sake but as evidence f o r the existence of c e r t a i n o r g a n i z i n g p r i n c i p l e s i n the mind which make i t p o s s i b l e f o r a speaker to use language c r e a t i v e l y . On t h i s account, d i s c o v e r i n g the p r i n c i p l e s that make p o s s i b l e the c r e a t i v e aspect of language use i s taken to be the c e n t r a l goal of l i n g u i s t i c theory. Furthermore, despite remarks such as the f o l l o w i n g i n Cart e s i a n L i n g u i s t i c s (CL page 12): It can h a r d l y be claimed that we have advanced s i g n i f i c a n t l y beyond the seventeenth century i n determining the c h a r a c t e r i s t i c s of i n t e l l i g e n t behaviour, the means by which i t i s acquired, the p r i n c i p l e s that govern i t , or the nature of the str u c t u r e s that u n d e r l i e i t . . . (the l a s t clause of which - at l e a s t - i s i n c o n s i s t e n t with the 30 passage quoted above beginning "Real progress...') Chomsky nevertheless s t a t e s i n Aspects that i n v i r t u e of i t s r e c u r s i o n s , the syntax i s the s o l e ' c r e a t i v e ' p a r t of a grammar (Aspects, page 135, 136): The f i n a l e f f e c t of a grammar, then, i s to r e l a t e a semantic i n t e r p r e t a t i o n to a phonetic r e p r e s e n t a t i o n — that i s , to s t a t e how a sentence i s i n t e r p r e t e d . This r e l a t i o n i s mediated by the s y n t a c t i c component of the grammar, which c o n s t i t u t e s i t s sole ' c r e a t i v e ' p a r t . The phrasing here i s a l i t t l e obscure; but i n Topics, i t i s c l e a r (Topics, page 66): The r e c u r s i v e property of the grammar ( i t s ' c r e a t i v e aspect' to r e t u r n to terminology used above) i s r e s t r i c t e d to the base component. In f a c t the r e s t r i c t i o n may be s t i l l h eavier than t h i s , since r e c u r s i o n may be l i m i t e d to i n t r o d u c t i o n of the symbol S, that i s , to i n t r o d u c t i o n of ' p r o p o s i t i o n a l content'. Focusing on the p a r e n t h e t i c a l phrase " i t s ' c r e a t i v e aspect', to r e t u r n to terminology used above", the "terminology used above" must r e f e r (for l a c k of other candidates) to the d i s c u s s i o n i n Topics on pages 11 and 12 about the ' c r e a t i v e aspect of language use'. Yet the pronoun ' i t s ' i n " i t s ' c r e a t i v e aspect'..." r e f e r s obviously to the grammar, and not to language use. This s h i f t i s s u r p r i s i n g and c l e a r l y r e q u i r e s explanation. Chomsky elsewhere wr i t e s (Preface to 2nd e d i t i o n of Language and Mind) ; ...a number of p r o f e s s i o n a l l i n g u i s t s have repeatedly confused what I r e f e r to here as "the c r e a t i v e aspect of language use" with the r e c u r s i v e property of generative grammars, a very d i f f e r e n t matter. 31 This makes i t impossible for the term 'crea t i v e aspect' on page 66 of Topics to be r e t u r n i n g to the terminology used p r e v i o u s l y . More important, i t leaves us i n the dark as to the contents of the term ' c r e a t i v e aspect of a grammar'. Obviously t h i s i s no t r i v i a l or t e r m i n o l o g i c a l matter. Chomsky equates ' c r e a t i v e aspect of a grammar' with 'recursive property of the grammar', and one i s free to make any such equations one pleases, but t h i s s u r e l y i n v i t e s the question, What i s the r e l a t i o n s h i p between 'the c r e a t i v e aspect of a grammer1 so defined, and 'the c r e a t i v e aspect of language use'? More simply (and more p e r s p i c u o u s l y ) , What i s the r e l a t i o n s h i p between r e c u r s i v e generation of sentence s t r u c t u r e s and the c r e a t i v e aspect of language use? To set t h i s question i n the broadest context, the theory of generative t r a n s f o r m a t i o n a l l i n g u i s t i c s focuses on e x p l a i n i n g the s p e c i a l c h a r a c t e r i s t i c s of the human use of language; i t has a l s o shown that no grammar which does not contain operations on two s t r u c t u r a l l e v e l s and a s e t of s y n t a c t i c r e c u r s i o n s can be d e s c r i p t i v e l y adequate. But there i s a s u b s t a n t i a l and i n t r i g u i n g question with regard to the r e l a t i o n s h i p between a demonstration of the n e c e s s i t y of a grammar's containing r e c u r s i v e devices f o r the generation of sentences f o r the purposes of d e s c r i p t i v e adequacy, and p r o v i d i n g a p s y c h o l o g i c a l account of the c r e a t i v e processes of language. I should s t r e s s as e a r l y as p o s s i b l e that t h i s p o i n t i s r a i s e d not to show that there i s 32 an u n f i l l e d gap i n the theory. i f we di g , we can f i n d i n the w r i t i n g s of Chomsky, Katz, Lyons and other t r a n s f o r m a t i o n a l l i n g u i s t s , a s u b s t a n t i a l theory about the r e l a t i o n s h i p between r e c u r s i v e generation of sentences and the c r e a t i v e aspect of language use. The theory presented, however, i s never supported, and i t w i l l be the attempt of t h i s study to check f o r the t r u t h of the theory. The eventual t o p i c , then, i s simply the r e l a t i o n s h i p between r e c u r s i v e generation of sentences and the c r e a t i v e aspect of language use. In t h i s chapter however, the t o p i c i s s e t i n the broader context of the r e l a t i o n s h i p between d e s c r i p t i v e and explanatory l i n g u i s t i c s . The connection between the two issues should be apparent: p s y c h o l o g i c a l l i n g u i s t i c s attempts to 'explain' how humans can use language c r e a t i v e l y , and so the question i s immediately r a i s e d , as to how the s t r u c t u r e s p o s t u l a t e d i n d e s c r i p t i v e l i n g u i s t i c s r e l a t e to the explanatory task. This l a t t e r problem, I think, has been inadequately t r e a t e d , and the f o l l o w i n g observations are r e l e v a n t i n t h i s regard. In order to q u a l i f y as the r i g h t kind of explanation of our c r e a t i v e use of language, the account must be both non-mechanical and at the same time d i s t i n c t from a statement of the c r i t e r i a according to which we d i s t i n g u i s h creatures which use language c r e a t i v e l y from those which do not. (This i s the 33 foundation of Chomsky's theory, e s p e c i a l l y from the philosopher's standpoint. For the p r e s e n t a t i o n of these views, see C a r t e s i a n L i n g u i s t i c s , pages 9 - 13). P o s t u l a t i n g explanations of t h i s k i nd immediately r a i s e s many problems, l a r g e l y because the paradigms of explanation are mechanical. J e r r y Fodor*s book, P s y c h o l o g i c a l Explanation t a c k l e s a number of these problems i n a f a i r l y s t r a i g h t f o r w a r d way; however there i s nothing i n that work which would enable us to i n t e r p r e t the d i f f i c u l t passages quoted above. The d i f f i c u l t y of the subject matter i s evident i n the one treatment of a ' m e n t a l i s t i c ' explanation of l i n g u i s t i c behaviour which i s discussed i n Fodor's book. To be explained i s the f a c t that people r e g u l a r l y tend to re p o r t the existence of pauses i n l i n g u i s t i c samples of languages f a m i l i a r to them, pauses which do not correspond to the f l u c t u a t i o n of the a c o u s t i c s i g n a l s as measured by spectographs. The 'pause perceptions', moreover, although they do not d i r e c t l y correspond to spectographic a n a l y s i s of the a c o u s t i c s i g n a l , tend to correspond to the s t r u c t u r a l a n a l y s i s of the sentence or l i n g u i s t i c s t r i n g i n question. A simple example given by Fodor i l l u s t r a t e s : i f the name 'Bob Lees' i s heard, there i s a tendency on the part of speakers of E n g l i s h to l o c a t e a pause between the two words, thus: 'Bob#Lees". However the ac o u s t i c pattern as s p e c i f i e d by spectographic a n a l y s i s i s u s u a l l y c l o s e r to 'Bo#bLees'. And the explanation i s that there are unconscious mental operations 34 which a s s i g n pauses according to s t r u c t u r e independently of the f l u c t u a t i o n s i n the a c o u s t i c s i g n a l . There are two not u n r e l a t e d problems here. One i s very b r i e f l y d iscussed i n Fodor's book, and the other not at a l l . I f one grants that there are unconscious operations which e x p l a i n the pause perceptions, one i s not o b v i o u s l y committed to regarding these operations as 'mental'. This question i s , fo r Fodor, ' p a r t l y t e r m i n o l o g i c a l ' ; a f t e r s a ying t h i s much, he argues that the unconscious operations should be regarded as mental because " i f they were conscious one would not h e s i t a t e to c a l l them i n t e l l i g e n t performances; they would indeed be paradigms of the s o r t of process that we d e s c r i b e as 'mental'."(page 87). The substance of t h i s argument i s that unconscious operations which, were they conscious, we would describe as ' i n t e l l i g e n t ' or 'mental' may themselves (ceterus paribus -- to b o l s t e r the argument) be regarded as 'mental'. I t h i n k i t i s t h i s g e n e r a l i z a t i o n which Fodor wants to p r e s s . A f a i r l y complicated wrangle could e a s i l y develop a t t h i s p o i n t , on the nature of the g e n e r a l i z a t i o n and i t s p l a u s i b i l i t y . But, even accepting the g e n e r a l i z a t i o n , there i s a s u b s t a n t i a l issue which i s being by-passed here, and which cuts deeply i n t o the a n a l y s i s given i n P s y c h o l o g i c a l E x p l a n a t i o n considered i n the way Fodor himself seems to regard i t , as an e l a b o r a t i o n of the programme of C a r t e s i a n l i n g u i s t i c s as a p p l i e d to the 35 concept of p s y c h o l o g i c a l explanation i n general. The s t a r t i n g p o i n t and sine qua non of the theory of C a r t e s i a n l i n g u i s t i c s i s the d i s t i n c t i o n between r a t i o n a l and n o n - r a t i o n a l creatures; with regard to the use of language, t h i s t r a n s l a t e s as the d i s t i n c t i o n between creatures which can express i n d e f i n i t e l y many thoughts i n freedom from e x t e r n a l s t i m u l i and a p p r o p r i a t e l y and coherently i n i n d e f i -n i t e l y many s i t u a t i o n s , and those creatures which cannot. I t i s s u r e l y to Chomsky's c r e d i t that he has not turned away from t h i s extremely important d i s t i n c t i o n , and keeps reminding both l i n g u i s t s and p s y c h o l o g i s t s who have turned away from i t that the d i s t i n c t i o n i s c r u c i a l and stands i n need of an account which i s p s y c h o l o g i c a l as opposed to p h i l o s o p h i c a l , i . e . , which looks f o r explanations f o r the d i s t i n c t i o n and not f o r the conceptual t o o l s according to which we apply the d i s t i n c t i o n . But the unconscious operations according to which we assign pause perceptions i n non-correspondence with f l u c -tuations i n the a c o u s t i c s i g n a l are by no means c l e a r l y examples of the d i s t i n c t i v e kinds of operations. Studies i n animal psychology have explored t h i s area and i t i s c l e a r that animals impose s t r u c t u r e on t h e i r perceptions i n much the same way, and so the argument that unconscious operations which, i f we were conscious of them, we would consider to be 'mental' should be considered to be mental operations i s i r r e l e v a n t . (For a study of G e s t a l t perception of monkeys, for example, 36 i n which i t i s concluded that i n f a n t monkeys, can r e l a t e p i c t u r e s to r e a l objects without p r i o r c o n d i t i o n i n g on that r e l a t i o n s h i p , see P i c t o r i a l Recognition i n the Infant Monkey, Proceedings o f the PsychPnomic S o c i e t y ' 63, by J . Hochberg.) An explanation of pause perception i n humans i s not an explanation of a d i s t i n c t i v e , c r e a t i v e operation of the human i n t e l l e c t . Nevertheless, such an explanation l i e s outside the scope of s i m p l i s t i c stimulus-response theory. In general, p o s t u l a t i n g a complex i n t e r n a l s t r u c t u r e to account f o r some competence does not thereby account f o r the p o s s i b i l i t y of behaving with a c r e a t i v e and r a t i o n a l i n t e l l e c t . For t h i s reason, we must be cautious i n attempting to d e l i n e a t e the explanatory power of an 'innateness p o s t u l a t e ' , and the p o s t u l a t e of an i n t e r n a l i z e d grammar. To be sure, i n order to account f o r any behaviour with a complex i n t e r n a l s t r u c t u r e — any ' s e r i a l l y organized behaviour' f o r example, using Lashley's terminology — we have to pos t u l a t e not only an i n t e r n a l i z a t i o n of some kind, but a l s o an innate s t r u c t u r e of some kind, i n v i r t u e of which the i n t e r n a l i z a t i o n s can be given the p a r t i c u l a r forms that they take. To the extent that b e h a v i o u r i s t s f a i l to come up with concrete proposals i n t h i s regard, they have f a i l e d to account f o r the complex behaviour of organisms. But such a c r i t i c i s m , however, i s not per se a r a t i o n a l i s t c r i t i c i s m of behaviourism, because i t i s neu t r a l with regard to r a t i o n a l and non-rational c r e a t u r e s . Unfortunately, i n s u f f i c i e n t a t t e n t i o n has been paid to the d i s t i n c t i o n between r a t i o n a l and n o n - r a t i o n a l creatures i n e l a b o r a t i o n s and demonstrations of the p o s s i b i l i t y of non-behaviourist explanations. S u r p r i s i n g l y enough, t h i s was supposed to be, on Chomsky's account, the d i s t i n c t i o n which motivated the search f o r n o n - b e h a v i o u r i s t i c e x p l a n a t i o n s i n the f i r s t i n s tance. I t i s not p l a t i t u d i n u o u s to suggest here that a good deal more work needs to be done to elaborate on the concept of the explanation i n p s y c h o l o g i c a l terms of r a t i o n a l behaviour, bearing i n mind that the explanation of some c a p a c i t y of a r a t i o n a l creature i s not thereby an explanation of a c a p a c i t y which i s unique to a r a t i o n a l creature, and c e r t a i n l y i t i s not an explanation of the c a p a c i t y of that creature to behave r a t i o n a l l y . This much seems to c o i n c i d e d i r e c t l y with Chomsky's pessimism about having made any r e a l dent i n the task r a i s e d but not solved by Descartes — to e x p l a i n the p o s s i b i l i t y of there being r a t i o n a l behaviour. But i t leaves us i n the dark as to the sense i n which i t i s a l l e g e d that a formulation of a r e c u r s i v e system for the generation of sentences provides us with an e x p l i c i t formulation of the c r e a t i v e processes of language. Once again, then, the question i s , what i s the r e l a t i o n s h i p between the presence of a device f o r the r e c u r s i v e generation of sentences i n a grammar, and the a b i l i t y 38 of a speaker to use a language c r e a t i v e l y ? As mentioned, the question i s not r a i s e d i n order to suggest that there i s no theory about the r e l a t i o n s h i p between the r e c u r s i v e generation of s e n t e n c e s t r u c t u r e s and the c r e a t i v e aspect of language u s e , f o r t h e r e i s . In fact, i t might be more accurate to say t h a t t h e r e a r e some t h e o r i e s on t h i s s u b j e c t , f o r there arc at l e a s t two points of contact. One i s the i m p l i c a t i o n which Chomsky (rather mysteriously) endorses that a creature w h i c h c a n produce and understand i n d e f i n i t e l y many sentences a p p r o p r i a t e l y i n s i t u a t i o n s e t c . cannot be a c r e a t u r e whose mental f a c u l t i e s are e x p l i c a b l e on mechanical models; i . e . , that a c r e a t u r e w h i c h can a p p r o p r i a t e l y use r e c u r s i v e l y g e n e r a t e d sentence, s t r u c t u r e cannot be 'a machine'. Thus, f o r example, i n P e r s i s t e n t Topics i n L i n g u i s t i c Theory, an a r t i c l e which b r i e f l y sketches the connection between C a r t e s i a n and modern g e n e r a t i v e l i n -g u i s t i c s , Chomsky wri t e s (Diogenes,, #51. pages 17,18)s This device (embedding of propositions) c a n be repeated i n d e f i n i t e l y so that a grammar of t h e f o r m suggested w i l l have the c a p a c i t y to generate an i n f i n i t e number of deep st r u c t u r e s . . . O f course, no e x p l i c i t system of t h i s s o r t was developed during t h i s period, although the general framework of such a system was suggested by many observations and p a r t i c u l a r analyses. I t i s worthy of note, however, that the c r e a t i v e aspect of language use was a t o p i c of considerable d i s c u s s i o n and thought i n the seventeenth century, i n the context o f the controversy over animal automatism...Descartes argues that the freedom of language use from stimulus c o n t r o l , i t s inde-pendence of what we would now c a l l " c o n d i t i o n i n g " , i t s appropriateness to s i t u a t i o n and to preceding discourse, and i t s t y p i c a l novelty a l l (emphasis added) p o i n t to the 39 existence of some s o r t of " a c t i v e p r i n c i p l e " that l i e s beyond the bounds of mechanical explanation... I t should be observed, i n c i d e n t a l l y , that there i s nothing at a l l absurd i n Descartes' argument... I t i s a l s o important to note that subsequent attempts to show that mechanical p r i n c i p l e s can account for human as w e l l as animal behaviour do not, so f a r as I can discover, attempt to r e f u t e these arguments. But j u s t what t h i s 'argument' i s remains uncle a r . In f a c t , we s h a l l only be i n a p o s i t i o n to see that there is_ an argument here once we have discovered i t s premise to be f a l s e . To a n t i c i p a t e we s h a l l be i n a p o s i t i o n a t the end of the i n v e s t i g a t i o n to see that the statement that i f a creature needs to have access to the r e c u r s i v e embedding of p r o p o s i t i o n s i n order to express i t s thoughts, then i t s use of language cannot be e x p l i c a b l e on mechanical models i s sound. But t h i s w i l l o n l y be a f t e r we have discovered that we do not need to have access to the r e c u r s i v e embedding of p r o p o s i t i o n s i n order to express our thoughts. In any case, t h i s brings us d i r e c t l y to the second theory about, or p o i n t of contact between, the r e c u r s i v e generation of sentences and the c r e a t i v e aspect of language use. In Language and Mind Chomsky s p e c i f i e s the f o l l o w i n g c o n d i t i o n of n e c e s s i t y f o r any c r e a t i v e l y useable language (L & M page 15): Following the Port-Royal theory to i t s l o g i c a l conclusions, then, the grammar of a language must contain a system of r u l e s that c h a r a c t e r i z e s deep and surface s t r u c t u r e s and the t r a n s f o r m a t i o n a l r e l a t i o n between them, and — i f i t i s to accommodate the c r e a t i v e aspect of language use — that does so over an i n f i n i t e domain of p a i r e d deep and surface s t r u c t u r e s , a p p r o p r i a t e l y r e l a t e d . 40 Undoubtedly, i t i s t h i s idea, that the presence i n a grammar of a device which r e c u r s i v e l y generates sentence s t r u c t u r e s i s a necessary c o n d i t i o n f o r a speaker to use the language s p e c i f i e d by the grammar c r e a t i v e l y which l i e s behind the a s s e r t i o n that developments i n the foundations of mathematics of the l a s t t h i r t y years ( i n the theory of r e c u r s i v e functions) provide us with 'a r e a l understanding of how a language can make i n f i n i t e use of f i n i t e means'. This point i s repeated on other occasions as w e l l . P e r s i s t e n t Topics, f o r example, v i r t u a l l y opens with the statements ( P e r s i s t e n t Topics, i n Diogenes #51 page 13): A c e n t r a l t o p i c of much current research i s what we may c a l l the 'c r e a t i v e * aspect of language use, that i s , i t s unboundedness and freedom from stimulus c o n t r o l . The speaker-hearer whose normal use of language i s " c r e a t i v e " , i n t h i s sense, must have i n t e r n a l i z e d a system of r u l e s t hat determines the semantic i n t e r p r e t a t i o n s of an unbounded s e t of sentences... And Fodor and Katz i n The Str u c t u r e of a. Semantic Theory put what appears to be the same p o i n t i n the f o l l o w i n g terms ( i n The S t r u c t u r e of Language, ed. Fodor and Katz, page 481): A f l u e n t speaker's mastery of the language e x h i b i t s i t s e l f i n h i s a b i l i t y to produce and understand the sentences of h i s language, i n c l u d i n g i n d e f i n i t e l y many that are wholly novel to him...The emphasis upon novel sentences i s important. The most c h a r a c t e r i s t i c feature of a language i s i t s a b i l i t y to make a v a i l a b l e an i n f i n i t y of sentences from which the speaker can s e l e c t appropriate and novel ones to use as the need a r i s e s . T h i s idea that only a grammar which generates sentences r e c u r s i v e l y can c o n s t i t u t e the grammar of a c r e a t i v e l y useable 41 language, or that only such a grammar can f u l l y serve the needs of a r a t i o n a l c r e a t i v e i n t e l l e c t , thus, i s not a random or i s o l a t e d remark. In i t s e x p l i c i t form i t i s repeated on a number of occasions; f u r t h e r , i t motivates and provides the bas i s f o r one of Chomsky's c r i t i c i s m s of s t r u c t u r a l i s t l i n g u i s t i c s ; and on the more t h e o r e t i c a l l e v e l , i t f i l l s , or attempts to f i l l , the gap r e f e r r e d to above concerning the r e l a t i o n s h i p between formulating a s y n t a c t i c a l l y r e c u r s i v e system of grammatical r u l e s , and e x p l a i n i n g how i t i s p o s s i b l e f o r human language to serve as the v e h i c l e f o r a c r e a t i v e , r a t i o n a l i n t e l l e c t , an i n t e l l e c t markedly and q u a l i t a t i v e l y d i f f e r e n t from that of a l l other known animals. The theory i s thus a c e n t r a l underpinning of generative t r a n s f o r m a t i o n a l l i n g u i s t i c s . Yet d e s p i t e i t s obvious importance, i t i s never supported where i t i s presented; nor has i t been c r i t i c a l l y scrutinized,. Let us examine the claim. To begin with, Chomsky gives two s o r t s of d e f i n i t i o n s of'the c r e a t i v e aspect of language use' i n h i s works, and while the statement of n e c e s s i t y (as I s h a l l continue to r e f e r to i t ) quoted i n various versions above that i f a grammar i s to accommodate the c r e a t i v e aspect of language use i t must contain a device which r e c u r s i v e l y generates sentence s t r u c t u r e s , i s obviously true on one i n t e r p r e t a t i o n of 'creative aspect of language use', i t i s not obviously true on the other. This f a c t provides us with a v i v i d i l l u s t r a t i o n of the non-synonymy of the two d e f i n i t i o n s of 42 the ' c r e a t i v e aspect of language use', and leads into the i n t e r e s t i n g question of how to determine the t r u t h of the statement of n e c e s s i t y on the second i n t e r p r e t a t i o n . For I w i l l argue, that the statement of n e c e s s i t y on the second i n t e r p r e t a t i o n would be the more i n t e r e s t i n g one with regard to c o n s t r u c t i n g an explanation of the p o s s i b i l i t y of there being r a t i o n a l creatures. Here, then, are the two d e f i n i t i o n s : i n Topics of the Theory of Syntax, and i n Current Issues i n L i n g u i s t i c Theory, the term ' c r e a t i v e aspect of language use' i s d e f i n e d i n the f o l l o w i n g way (Topics, page 1 1 ) : The most s t r i k i n g aspect of l i n g u i s t i c competence is what we may c a l l the ' c r e a t i v i t y of language", that i s , the speaker's a b i l i t y to produce new sentences, sentences that are immediately understood by other speakers although they bear no p h y s i c a l resemblance to sentences which are f a m i l i a r . A d i f f e r e n t s o r t of d e f i n i t i o n , however, i s provided i n C a r t e s i a n L i n g u i s t i c s , Aspects of the Theory of Syntax, and Language and Mind. For example, i n Language and Mind, the d e f i n i t i o n i s presented as follows (L & M page 6): This new p r i n c i p l e (whose essence i s thought) has a " c r e a t i v e aspect", which i s evidenced most c l e a r l y i n what we may r e f e r to as the " c r e a t i v e aspect of language use", the d i s t i n c t i v e l y human a b i l i t y to express new thoughts and to understand e n t i r e l y new expressions of thought, w i t h i n the framework of an " i n s t i t u t e d language", a language that i s a c u l t u r a l product subject to laws and p r i n c i p l e s p a r t i a l l y unique to i t and p a r t i a l l y r e f l e c t i o n s of general p r o p e r t i e s of mind. For easy reference, I w i l l r e f e r to the f i r s t as'the d e f i n i t i o n i n terms of novel sentence production', and to the second as 'the 43 d e f i n i t i o n i n terms of novel thought expression'. On the d e f i n i t i o n of the c r e a t i v e aspect of language use i n terms of novel sentence production, the statement of n e c e s s i t y a s s e r t s that i f a grammar of a language i s to accommodate the f a c t that i t i s w i t h i n a speaker's competence to assign s t r u c t u r a l d e s c r i p t i o n s to a t h e o r e t i c a l l y u n l i m i t e d number of s y n t a c t i c a l l y complex sentences, then the grammar must con t a i n a device f o r the r e c u r s i v e generation of sentences. And t h i s i s c l e a r l y t r u e . I f we simply consider the f a c t that there i s no t h e o r e t i c a l l i m i t on the number of embedded noun clauses that a sentence can contain, we w i l l conclude that the grammar may be c h a r a c t e r i s e d as being ' s y n t a c t i c a l l y r e c u r s i v e ' , and so a formulation of the grammar w i l l contain a system of s y n t a c t i c r e c u r s i o n s . On the second d e f i n i t i o n of the c r e a t i v e aspect of language use, the statement of n e c e s s i t y a s s e r t s that i f a grammar of a language i s to accommodate the f a c t that speakers can express i n language ever novel thoughts, then the grammar must contain a device f o r r e c u r s i v e generation of sentence s t r u c t u r e s . But t h i s i s not true i n any obvious sense; there i s no simple demonstration of i t s t r u t h corresponding to the simple demonstration given above of the t r u t h of the statement of ne c e s s i t y on the f i r s t i n t e r p r e t a t i o n of 'creative aspect of language use'. In f a c t one might rather argue to the contrary: some s i n g l e sentences can be used to express i n d e f i n i t e l y many 44 thoughts i n v i r t u e of t h e i r being able to r e f e r to i n d e f i n i t e l y many o b j e c t s . On such grounds, someone might cl a i m that a grammar which was capable of generating only the one sentence, 'The cat was on the mat', could s t i l l serve as the v e h i c l e f o r the expression of i n f i n i t e l y many thoughts. But t h i s would be an e x c e s s i v e l y r i g i d i n t e r p r e t a t i o n of the statement of n e c e s s i t y on the second d e f i n i t i o n , i n terms of thought expressions. (Besides being conceptually dubious: could such a grammar define a 'language' at a l l ? ) In p a r t i c u l a r , i t would ignore the most i n t e r e s t i n g aspect of the statement, namely the i m p l i c i t a s s e r t i o n that i n order to express a l l our thoughts we need to have a language whose grammar generates sentences r e c u r s i v e l y . This may not be demonstrably true i n any simple way, but i t i s a l s o not demonstrably f a l s e through any immediate or simple c o n s i d e r a t i o n s . Also, t h i s i n t e r p r e t a t i o n that i n order to express a l l our thoughts we need to have a language whose grammar generates sentences r e c u r s i v e l y seems to be what Chomsky had i n mind when he wrote — i n an e a r l y work, which gives us an i n s i g h t i n t o the statement of n e c e s s i t y — " I f we construct a f i n i t e - s t a t e grammar that produces only E n g l i s h sentences, we know that i t w i l l f a i l to produce an i n f i n i t e number of these sentences; i n p a r t i c u l a r i t w i l l f a i l to produce an i n f i n i t e number of true sentences, f a l s e sentences, reasonable questions that could i n t e l l i g i b l y be asked, and the l i k e " . (Three Models for the D e s c r i p t i o n of Language, i n IRE Transactions, page 116). 45 The statement of n e c e s s i t y i s , then, c l e a r l y true on the d e f i n i t i o n of ' c r e a t i v e aspect of language use' i n terms of novel sentence production, but i t i s , although i n t e r e s t i n g , not o b v i o u s l y true on the d e f i n i t i o n of ' c r e a t i v e aspect of language use' i n terms of novel thought expression. On which i n t e r p r e t a t i o n would i t be most l i k e l y to p l a y a r o l e i n any p s y c h o l o g i c a l explanation of the p o s s i b i l i t y of there being r a t i o n a l as opposed to n o n - r a t i o n a l creatures? My remarks here are only t e n t a t i v e , but I would l i k e to suggest that the statement of n e c e s s i t y would be more l i k e l y to c o n t r i b u t e toward an explanation of the p o s s i b i l i t y of there being r a t i o n a l creatures on the second i n t e r p r e t a t i o n , i n terms of novel thought expression, than on the f i r s t i n t e r p r e t a t i o n i n terms of novel sentence production. The suggestion i s o f f e r e d because of the apparent primacy of the second d e f i n i t i o n , the d e f i n i t i o n i n terms of thought expression with regard to r a t i o n a l i t y . I s h a l l argue that a creature's being able to produce and assign semantic i n t e r p r e t a t i o n s to i n f i n i t e l y many sentences i s not s u f f i c i e n t f o r i t s being r a t i o n a l ; whether i t i s necessary f o r i t s being r a t i o n a l turns out to be, or, at very l e a s t , hinges on, the statement of n e c e s s i t y on the d e f i n i t i o n of c r e a t i v e aspect of language use i n terms of novel thought expression. On the other hand, a creature's being able to express ever novel thoughts i n i t s language i s s u f f i c i e n t f o r i t s being r a t i o n a l , and i t may w e l l 4 6 be necessary f o r i t s being r a t i o n a l . Thus of the two d e f i n i t i o n s , the d e f i n i t i o n i n terms of thought expression appears to be the anchor to which the d e f i n i t i o n i n terms of sentence prod-u c t i o n i s p a r a s i t i c a l l y attached — at l e a s t i n s o f a r as the term ' c r e a t i v e aspect of language use' i s supposed to be l i n k e d to the C a r t e s i a n concept of a s p e c i a l k i nd of i n t e l l e c t . F i r s t note that the ' c r e a t i v e aspect of language use' i s supposed to be so l i n k e d i n Chomsky's account (Current Issues, page 51 as anthologized i n The S t r u c t u r e of Language, ed. Fodor & K a t z ) : ...the r e a l i z a t i o n that t h i s " c r e a t i v e " aspect of language i s i t s e s s e n t i a l c h a r a c t e r i s t i c can be traced back at l e a s t to the seventeenth century. Thus we f i n d the C a r t e s i a n view that man alone i s more than mere automatism, and that i t i s the possession of true language that i s the primary i n d i c a t o r of t h i s (see Descartes, Discourse on Method, p a r t V)... However, the d e f i n i t i o n of c r e a t i v e aspect of language use as presented i n Current Issues, i . e . , the d e f i n i t i o n i n terms of novel sentence production, i s not so l i n k e d conceptually, except v i a the other d e f i n i t i o n , i n terms of thought expression. To get a h o l d on the question, consider the f o l l o w i n g observation concerning the r e l a t i o n s h i p between l i n g u i s t i c use and l i n g u i s t i c s t r u c t u r e : A s e m a n t i c a l l y p r i m i t i v e communication-system (or language, speaking loosely) might be extremely complex s y n t a c t i c a l l y . 47 imagine, f o r example, a system of b i r d songs i n which there are, say, nine d i s t i n c t sounds, which o r n i t h o l o g i s t s represent by numerals from 1 to 9. Imagine a l s o that each i n d i v i d u a l 'speech-act' or song has no f i x e d length, and that b i r d s s i n g when there i s some danger present, when they are i n p a i r s of male and female during mating season, and when they have been alone f o r over ten minutes ('alone f o r x minutes' meaning, say that no b i r d of t h e i r species has been w i t h i n the f i e l d of v i s i o n , nor sung w i t h i n the ranee of a u d i b i l i t y of the b i r d being described as being 'alone f o r x minutes' f o r x minutes). Assume a l s o that the f o l l o w i n g s t r i n g s represent t y p i c a l songs i n the presence of danger: 1) 12323685 2) 435364 3) 914164; that the f o l l o w i n g represent t y p i c a l songs when a b i r d i s alone: 1) 2145957 2) 4265781383 3) 871; and that the f o l l o w i n g represent t y p i c a l songs when the b i r d i s mating: 1) 51247597 2) 42685365195 3) 57631. In these repre s e n t a t i o n s , the danger songs a l l add up to mu l t i p l e s of 5; the alone-songs add up to m u l t i p l e s of 3; and the mating songs add up to mul t i p l e s of 2. Assume fu r t h e r that t h i s holds true i n general, and f i n a l l y , that the only behavioural c o r r e l a t i o n s with the p a r t i c u l a r numbers chosen have to do with the i n t e n s i t y of the danger, the length of time the b i r d has been alone, and the i n t e n s i t y of the mating a c t i v i t y . Although rather u n l i k e l y , such a system of communication i s l o g i c a l l y p o s s i b l e . In accounting f o r the behaviour of b i r d s which use such a system of communication, an animal 48 p s y c h o l o g i s t would have to po s t u l a t e an i n t e r n a l i z e d grammar with a very h i g h l y s t r u c t u r e d and complex s y n t a c t i c component: the b i r d i s capable of producing and i n t e r p r e t i n g i n d e f i n i t e l y many s y n t a c t i c a l l y complex s t r u c t u r e s . But the s y n t a c t i c s t r u c t u r e s generated by the grammar can only be i n t e r p r e t e d i n a f i x e d and f i n i t e way, i n terms of three semantic dimensions which can only be weighed i n i n t e n s i t y along some s c a l e or other. Despite the f a c t that the i n t e r n a l i z e d grammar i s s t r u c t u r e d along complex mathematical l i n e s , the b i r d s using such systems of communication are under the same mental r e s t r i c t i o n s as b i r d s which operate along s i m i l a r semantic dimensions u s i n g a s y n t a c t i c a l l y simple system. We could not des c r i b e the b i r d which uses the s y n t a c t i c a l l y complex system as a creature which uses a language ' c r e a t i v e l y ' . We could not p r o p e r l y say, f o r example, that the b i r d i s capable of performing mathematical operations, even though i t must have an i n t e r n a l i z e d represen-t a t i o n of some mathematical operations. (For arguments to t h i s e f f e c t with regard t o analogous though more d i f f i c u l t questions such as whether or not an animal which has i n i n t e r n a l i z e d map of some kind can be s a i d to use the map, see J . Bennett, R a t i o n a l i t y , #10. Bennett's arguments apply a f o r t i o r i here; see a l s o i b i d , page 20 f o r a s t r o n g l y r e l a t e d consideration.) Thus, the a b i l i t y of a creature to produce and as s i g n 49 semantic i n t e r p r e t a t i o n s to i n d e f i n i t e l y many s y n t a c t i c a l l y complex objects i s not s u f f i c i e n t t o guarantee t h a t t h e creature i s r a t i o n a l , or has a ' c r e a t i v e 1 i n t e l l e c t . . B u t i s i t necessary? That i s , i s the competence t o p r o d u c e and understand i n f i n i t e l y many s e n t e n c e s a n e c e s s a r y competence f o r a creature to have i f the c r e a t u r e i s t o be r a t i o n a l ? To answer t h i s , we need to know whether or t,. or a c r e a i u r e w h i c h can express i n d e f i n i t e l y many new t h o u g h t s .in f r e e d o m f r o m e x t e r n a l s t i m u l i and a p p r o p r i a t e l y and c o h e r e n t l y must h a v e an i n t e r n a l r e p r e s e n t a t i o n of a grammar w h i c h g e n e r a t e s sentences r e c u r s i v e l y -- and t h i s e i t h e r h i n g e s , o n or i s i d e n t i c a l with, the statement o f n e c e s s i t y u s i n g t h e d e f i n i t i o n of c r e a t i v e aspect of language use i n terms of t h o u g h t expression. For consider the r e l a t i o n s h i p b e t w e e n ' r a t i o n a l c r e a t u r e ' and 'creature with an a b i l i t y t o e x p r e s s i n d e f i n i t e l y many new thoughts i n freedom from ex t e r n a l s t i m u l i and a p p r o p r i a t e l y and coherently'. There i s l i t t l e d o u b t here b u t t h a t t h e a b i l i t y to express i n d e f i n i t e l y many new t h o u g h t s i n f r e e d o m from e x t e r n a l s t i m u l i , coherently, and a p p r o p r i a t e l y , i s s u f f i c i e n t f o r a creature's b e i n g r a t i o n a l . The more p o t e n t i a l l y c o n t r o v e r s i a l questions are ( n o t relevant as a controversy here): What must a creature do i n order t o count as b e i n g a creature which can express i n d e f i n i t e l y many new t h o u g h t s , coherently, a p p r o p r i a t e l y , and i n freedom f r o m e x t e r n a l s t i m u l i ? 50 And secondly (more re l e v a n t here): must a creature be able to express i n d e f i n i t e l y many new thoughts coherently, a p p r o p r i a t e l y , and i n freedom from ex t e r n a l s t i m u l i , i n order to be pr o p e r l y described as a r a t i o n a l creature? This i s a more d i f f i c u l t matter -- nevertheless, a p o s i t i v e conclusion i s e n t i r e l y p l a u s i b l e . A number of contemporary philosphers have, to my mind, c o n v i n c i n g l y argued points c l o s e enough to t h i s one, and from which t h i s one could be drawn. In Bennett's R a t i o n a l i t y , for example, i t i s argued that a creature which has no language at a l l cannot be r a t i o n a l , and that only a creature whose language includes dated judgements can be r a t i o n a l . An argument might be constructed, I think, showing that to make a s i n g l e dated judgement one must have the competence to make an i n f i n i t y of dated judgements. In any case, i t i s c l e a r that the a b i l i t y to express i n d e f i n i t e l y many thoughts coherently and a p p r o p r i a t e l y , and i n freedom from e x t e r n a l s t i m u l i i s more c l o s e l y r e l a t e d from the conceptual p o i n t of view to r a t i o n a l i t y than the a b i l i t y to produce and as s i g n semantic i n t e r p r e t a t i o n s to i n d e f i n i t e l y many formal objects or sentences. This, however, should only be regarded as t e n t a t i v e l y suggesting that the statement of n e c e s s i t y which asserts that i f a speaker i s to be competent to produce and assign semantic i n t e r p r e t a t i o n s to i n d e f i n i t e l y many sentences the speaker must 51 have an i n t e r n a l i z e d r e p r e s e n t a t i o n of a grammar which r e c u r s i v e l y generates sentences i s not as i n t e r e s t i n g as the statement of n e c e s s i t y which a s s e r t s that i f a speaker i s to be able to express an u n l i m i t e d number of thoughts, a p p r o p r i a t e l y and coherently, and i n freedom from external s t i m u l i , then the speaker must have an i n t e r n a l i z e d r e p r e s e n t a t i o n of a grammar which r e c u r s i v e l y generates sentences, with regard to accomplishing the task r a i s e d (according to Chomsky) by Descartes: to e x p l a i n how i t i s p o s s i b l e f o r us to be r a t i o n a l . The suggestion i s only t e n t a t i v e because the concept of explanation i s d i f f i c u l t , and the statement of n e c e s s i t y under both i n t e r p r e t a t i o n s may have bea r i n g . Notice, f o r example, that the 'explanation' of c r e a t i v i t y — on whichever i n t e r p r e -t a t i o n — i s at a l e v e l removed from that of r a t i o n a l i t y now. In general, p r o v i d i n g an explanation involves p r o v i d i n g necessary and/or s u f f i c i e n t e m p i r i c a l conditions f o r the events or processes or s t a t e s - o f - a f f a i r s to be explained. But as we have seen, we can a l s o t e s t the d e f i n i t i o n of c r e a t i v i t y given a g a i n s t the concept of r a t i o n a l i t y , although t h i s t e s t i s more along conceptual than e m p i r i c a l l i n e s . Nevertheless conditions are r e l a t i n g to one another at two l e v e l s of n e c e s s i t y and/or s u f f i c i e n c y , and su'ch f o r b i d d i n g questions as: is an e m p i r i c a l l y necessary c o n d i t i o n f o r something which i s (say) a n a l y t i c a l l y s u f f i c i e n t but not necessary f o r a t h i r d 52 t h i n g to be regarded as an explanation f o r the t h i r d ? w i l l keep appearing. There i s one issu e , t h o u g h , on w h i c h the s t a t e m e n t o f n e c e s s i t y on the second i n t e r p r e t a t i o n has b e a r i n g w h i l e on the f i r s t i t has no bearing: v i z . , t h e r e ' i t i o n s h i p b e t w e e n s y n t a c t i c devices and semantic p o s s i b i l i t i e s . To s t u d y the semantic e f f e c t s of the r e c u r s i v e g e n e r a t i o n of s e n t e n c e s by checking the statement of n e c e s s i t y i n i t s e l f makes i t an absorbing and worthwhile i s s u e . The r e s t o f t h e essay, with the exception of the short remarks d i r e c t l y f o l l o w i n g ie devoted to checking the statement of n e c e s s i t y on the d e f i n i t i o n of c r e a t i v e aspect of language u s e i n term s o f e v e r n o v e l thought expression; i . e . , to a s t u d y of w h e t h e r o r n o t our languages must r e c u r s i v e l y g e n e r a t e sentences i f we are t o be able to express a l l the i n d e f i n i t e l y many t h o u g h t s we can express i n our languages. (I s h a l l argue t h a t to t h i s e x t e n t , they need not: a language — i n f a c t , a sub-grammar o f E n g l i s h — with no r e c u r s i v e generation o f deep s t r u c t u r e types for sentences can serve as t h e v e h i c l e f o r t h e e x p r e s s i o n o f a l l the thoughts expressable u s i n g t h e w h o l e grammar o f E n g l i s h . ) Before s e t t l i n g down to t h i s s t u d y , one o r two f i n a l remarks on the r e l a t i o n s h i p b e t w e e n a r a t i o n a l i s t l i n g u i s t i c s , and a study of U n i v e r s a l Grammar as c o n c e i v e d b y Chomsky. These remarks may serve to put f u r t h e r perspective on t h e s t a t e m e n t of. 5 3 n e c e s s i t y , but nothing that f o l l o w s afterwards depends on them. Chomsky adopts as a working h y p o t h e s i s t h e theory t h a t there i s a U n i v e r s a l Grammar with b o t h f o r m a l and substantive u n i v e r s a l s ; the formal u n i v e r s a l s a r e t a k e n to have a somewhat more a b s t r a c t character than t h e substantive u n i v e r s a l s ; i f a l l grammars of n a t u r a l human l a n g u a g e s o p e r a t e a t two l e v e l s of s t r u c t u r e , t r a n s f o r m a t i o n a l l y r e l a t e d , t h e n t h a t m i g h t provide an example of, or be t a k e n t o be determined b y , a formal u n i v e r s a l . The terms i n which s t r u c t u r a l d e s c r i p t i o n s are c h a r a c t e r i s e d , on the other hand, -- some category u n i v e r s a l s , f o r example — might be regarded p r o f i t a b l y as being 'substantive 1 rather than 'formal' . Chomsky, p a r t l y as an argument f o r the innateness o f f o r m a l u n i v e r s a l s , and p a r t l y to avoid confusions as t o the nature of t h e i n n a t e p r i n c i p l e s b e i n g postulated, observes on a number of o c c a s i o n s that formal u n i v e r s a l s do not appear to be s p e c i f i e d i n any way as the most l o g i c a l or a p r i o r i , the simplest forms p o s s i b l e . As an example Chomsky c i t e s the f a c t that such t r a n s f o r m a t i o n a l operations as r e f l e c t i o n , w h e r e b y a s t r i n g 'abcdefg', say, i s transformed i n t o 'gfedcba' do not ever appear, de s p i t e t h e i r formal s i m p l i c i t y , i n any n a t u r a l languages, and i t i s suggested that s t r u c t u r e dependency, which would exclude simple r e f l e c t i o n as a grammatical transformation, i s a c h a r a c t e r i s t i c of a l l transformations as s t i p u l a t e d by some formal p r i n c i p l e of the U n i v e r s a l Grammar. 54 In response to t h i s idea, Max Black writes i n Explanation In The Behavioural Sciences (an anthology e d i t e d by Borger and C i o f f i page 460): Chomsky i s able to t e l l us some of the a l t e r n a t i v e p r i n c i p l e s (or modes of organization) that are not used i n the languages we know. Let us c a l l such p r i n c i p l e s u nnatural. Are we asked to b e l i e v e that human beings cannot in f a c t use unnatural p r i n c i p l e s , at however much inconvenience, and s t i l l make themselves understood? I f so, I doubt whether we can know i t , without a t r i a l . But suppose i t turned out, as i t might, that we could speak and make ourselves understood by employing unnatural p r i n c i p l e s . Then, on Chomsky's approach we would have to po s t u l a t e some higher-order d i s p o s i t i o n f o r modifying the f i r s t - o r d e r d i s p o s i t i o n s that produce the 'natural' l i n g u i s t i c c a p a c i t i e s . Which looks s u s p i c i o u s l y l i k e a re d u c t i o ad absurdum. Chomsky r e p l i e s (Explanation In The Behavioural Sciences, page 469) Consider the f o l l o w i n g l i n e of argument (I do not, f o r the moment, argue that i t i s c o r r e c t , but only that i t i s coherent) . A theory of mind w i l l p o s t u l a t e a number of f a c u l t i e s , one, a language f a c u l t y with the p r o p e r t i e s sketched i n my paper But the knowledge of language provided by t h i s f a c u l t y i s not the t o t a l i t y of human knowledge, and i t s mechanisms do not exhaust the devices by which knowledge (better, b e l i e f ) c a n be acquired. For example, these mechanisms preclude an operation of i n t e r r o g a t i o n by l e f t - r i g h t i n v e r s i o n , but permit an operation of i n t e r r o -g ation of the h i g h l y complex s o r t that we have i n E n g l i s h . Making use of other f a c u l t i e s of the mind, a person could i n some sense 'learn' a system of communication that forms i n t e r r o g a t i v e s by l e f t - r i g h t i n v e r s i o n ; thus the experiment that Professor Black suggests would s u r e l y succeed. But i t would not disprove the assumption that the language f a c u l t y precludes such an operation. I have no doubt that the other f a c u l t i e s of mind a l s o have t h e i r i n t r i n s i c l i m i t a t i o n s and s t r u c t u r e . But nothing i s known of any importance about t h i s matter...in any event, there i s s u r e l y no incoherence i n supposing that one of the components of mind — the language f a c u l t y -- has p a r t i c u l a r p r o p e r t i e s a b i l i t i e s , and l i m i t a t i o n s . I do not, therefore, see that there i s any r e d u c t i o ad absurdum l u r k i n g dangerously i n the background. 55 Both Chomsky and Black f i n d that the thought of a creature's being able to l e a r n to use a language whose formal p r i n c i p l e s do not conform to the b i o l o g i c a l l y determined formal p r i n c i p l e s of i t s p o s s i b l e n a t u r a l languages an i n t e r e s t i n g one; however, Chomsky's r e p l y to the argument that t h i s leads to an absurd p o s i t i o n i s , I think, c o r r e c t . On the other hand, u n d e r l y i n g the query r a i s e d by Professor Black i s a d i s t i n c t i o n , or a set of d i s t i n c t i o n s with serious r a m i f i c a t i o n s f o r the theory that a U n i v e r s a l Grammar, the d i s c o v e r y of whose p r i n c i p l e s receives e m p i r i c a l confirmation, c o n s t i t u t e s an explanation of our i n t e l l e c t u a l f a c u l t y with regard to language. In Language and Mind Chomsky w r i t e s (L & M page 24): The p r i n c i p l e s that determine the form of grammar and that s e l e c t a grammar of the appropriate form on the b a s i s of c e r t a i n data c o n s t i t u t e a subject that might, f o l l o w i n g a t r a d i t i o n a l usage be termed ' u n i v e r s a l grammar'. The study of u n i v e r s a l grammar, so understood, i s a study of the nature of human i n t e l l e c t u a l c a p a c i t i e s . I t t r i e s to formulate the necessary and s u f f i c i e n t c o n d i t i o n s that a system must meet to q u a l i f y as a p o t e n t i a l human language, cond i t i o n s that are not a c c i d e n t a l l y true of the e x i s t i n g human languages, but that are r a t h e r rooted i n the human 'language capac i t y ' , and thus c o n s t i t u t e the innate o r g a n i z a t i o n that determines what counts as l i n g u i s t i c experience and what knowledge of language a r i s e s on the b a s i s of t h i s experience. Although i t i s not incoherent to suggest that there are s e v e r a l i n t e l l e c t u a l f a c u l t i e s , each with t h e i r i n t r i n s i c l i m i t a t i o n s and s t r u c t u r e , i t may w e l l be incoherent to suggest that the e m p i r i c a l l y discoverable u n i v e r s a l s of language ex p l a i n or determine 'what counts as l i n g u i s t i c experience' 56 at the same time as h o l d i n g that we can l e a r n language outside the e m p i r i c a l l y discovered language u n i v e r s a l s . I t h i n k that one of the deepest motivating f a c t o r s behind the theory of generative t r a n s f o r m a t i o n a l l i n g u i s t i c s i s the hope that e m p i r i c a l l i n g u i s t i c s w i l l e v e n t u a l l y be able to c o n t r i b u t e an explanation i n p s y c h o l o g i c a l terms of our mental f a c u l t i e s ; but f o r a proposal to be ' s c i e n t i f i c ' or e m p i r i c a l , with regard to the U n i v e r s a l Grammar, i t must be the case that i t i s a r b i t r a r y , or that there are other conceivable a l t e r n a t i v e s . Thus Chomsky w r i t e s , (Explanation i n the Behavioural Sciences, Reply, page 468): The p a r t i c u l a r assumptions of u n i v e r s a l grammar have e m p i r i c a l consequences i n terms of which they must be evaluated. I agree that these assumptions are 'unplausible i n themselves 1, but regard t h i s as t h e i r s c i e n t i f i c merit. I t i s j u s t t h i s f a c t that gives importance to the evidence that i s o f f e r e d i n support of these p r i n c i p l e s . There i s , f o r example, no a p r i o r i c o n s i d e r a t i o n that lends p l a u s i b i l i t y to a theory of s y n t a c t i c operations that precludes the formation of i n t e r r o g a t i v e s by l e f t - r i g h t i n v e r s i o n . . . . I t i s j u s t t h i s f a c t that c o n s t i t u t e s the s c i e n t i f i c i n t e r e s t of such a theory. In the case of U n i v e r s a l Semantics, whether there are any such a r b i t r a r y p r i n c i p l e s i s unclear, and, even i f there are, s u r e l y there must be others which are not a r b i t r a r y . Yet one might s t i l l want to regard the non-empirical p r i n c i p l e s of u n i v e r s a l semantics as p s y c h o l o g i c a l and explanatory rather than p h i l o s o p h i c a l . In t h i s l i g h t , the statement that i f a creature i s to be able to express i n f i n i t e l y many thoughts a p p r o p r i a t e l y and 57 coherently i n i n d e f i n i t e l y many s i t u a t i o n s i t must have a grammar which r e c u r s i v e l y g e n e r a t e s s e n t e n c e s i s a r r y c b o l o g i c a l theory which i s not obviously con f i r mat. l e or r-tucabl<? by experimental c o n s i d e r a t i o n s . At t h e same time i t i s not t r u e by d e f i n i t i o n ? i t appears, r a t h e r , t o r e q u i r e a s p e c i a l s o r t of a n a l y s i s . To see more c l e a r l y t h a t i t ^bov lc\ n o t be regarded as an e m p i r i c a l c l a i m on a oar v i t h t h e .; .< ,..© t'o&t simple r e f l e c t i o n of a s t r i n g i s not a p o s s i b l e t r a n s fox vaa t i o n a 3 operation i n a n a t u r a l language, consider t h e d i s t i n c t i o n between 'any p o s s i b l e n a t u r a l human l a n g u a g e ' , -md 'any language which can serve to e x p r e s s any t h ought w n i c h can be expressed i n any p o s s i b l e n a t u r a l human language',, The U n i v e r s a l Grammar as sketched by Chomsky a t t e m p t s t o s p e c i f y the former; but a l i n g u i s t i c s w h i c h i s u n i v e r s a l i n a more Kantian sense, i n p a r t i c u l a r , a n o n - e m p i r i c a < l i n g u i s t i c s , would be r e q u i r e d to s p e c i f y t h e n e c e s s a r y (and minimal) s t r u c t u r e s f o r the l a t t e r . In t h e s e t e r m s , the s t a t e m e n t that a device f o r the r e c u r s i v e g e n e r a t i o n o f s e n t e n c e s i s necessary f o r the f u l l expression o f t h o u g h t i s a c l a i m a b o u t the necessary equipment for any l a n g u a g e w h i c h can s e r v e t o express any thought which can be e x p r e s s e d i n any n a t u r a l human language, and determining w h e t h e r i t i s t r u e o r not cannot be thought of as a s t r i c t l y e m p i r i c a l s t u d y , or as a study which i s bounded by the p r i n c i p l e s w h i c h s p e c i f y t h e f o r m of 'any p o s s i b l e n a t u r a l human l a n g u a g e ' . A s i m i l a r d i s t i n c t i o n , i t might be noted, between e m p i r i c a l and non - e m p i r i c a l l i n g u i s t i c s i s sketched i n Strawson's a r t i c l e Grammar and Philosophy, only Strawson's d i s t i n c t i o n between ' d e s c r i p t i v e ' and ' e s s e n t i a l ' l i n g u i s t i c s focuses on problems of a s s i g n i n g l e x i c a l items to grammatical c a t e g o r i e s ; nevertheless, there i s the i n t u i t i o n there too, I think, that questions i n v o l v i n g what Chomsky thinks of as substantive u n i v e r s a l s , or elsewhere as u n i v e r s a l semantic constraints on the s t r u c t u r e of the l e x i c o n , cannot be regarded as ' d e s c r i p t i v e ' or 'em p i r i c a l ' i n the way that questions i n v o l v i n g Chomsky's k i n d of formal u n i v e r s a l s can. Note, though, that the d i s t i n c t i o n between the study o f 'any p o s s i b l e n a t u r a l human language', and 'any language which can serve t o express any thought which can be expressed by any p o s s i b l e n a t u r a l human language' does not simply r e f l e c t the d i s t i n c t i o n between studying ' u n i v e r s a l syntax and phonology on the one hand, and 'universal semantics' on the other. I have suggested that some branch of u n i v e r s a l semantics cannot be regarded as e m p i r i c a l i n the way that Chomsky's study of u n i v e r s a l syntax i s . But i t would appear that there i s a l s o a d i s t i n c t i o n between p r i n c i p l e s of syntax and phonology which determine the s t r u c t u r e of any p o s s i b l e n a t u r a l human language, and so which are, i n an important sense, not the only conceivabl p r i n c i p l e s , and hence the discovery of which are ' s c i e n t i f i c ' , or e m p i r i c a l , p r i n c i p l e s , thus which do not determine the 59 s t r u c t u r e of any language which can serve to express any thought which can be expressed i n any p o s s i b l e n a t u r a l human language — on the one hand, and, on the other, p r i n c i p l e s of syntax and (possibly) phonology which determine the s t r u c t u r e of any language which can serve to express any thought which can be expressed i n any p o s s i b l e n a t u r a l human language. These p r i n c i p l e s are not e m p i r i c a l l y confirmable or r e f u t a b l e , but are d iscovered through various s o r t s of p h i l o s o p h i c a l reasoning. An example of the f i r s t s et has already been given: the s t r u c t u r e -dependency of t r a n s f o r m a t i o n a l operations; a f a i r l y simple example of the second set might be the p r i n c i p l e that any language which can serve to express any thought which can be expressed i n any n a t u r a l language must c o n s i s t of more than a s i n g l e symbol. In f a c t , i t can be demonstrated that the minimum number of symbols f o r the surface s t r u c t u r e l e v e l of any f u l l y expressive language i s two. That i s , although i t can be demonstrated that any f u l l y expressive language must contain more than one symbol, i t cannot be demonstrated that the surface s t r u c t u r e of any f u l l y expressive language mustcontain more than two symbols, as a f u l l y expressive language could e x i s t whose surface l e v e l contained only two symbols. A code could be set up f o r w r i t t e n English,- f o r example, i n which each symbol ( i n c l u d i n g 'space') i s c o d i f i e d by a combination of zero's and one's. The requirement that there be two symbols 60 and not one a r i s e s from the f a c t t h a t 'space' must count as a s t r u c t u r a l u n i t , and i f t h e r e were only one symbol* there would be no way to 'read' t h e s t r u c t u r a l u n i t s , and t h e r e b y assign s t r u c t u r a l d e s c r i p t i o n s to t h e formal objects o f t h e surface of the code. Another example, and p e r h a p s a more i n t e r e s t i n g one o f t h l a t t e r set would be the statement t h a t some p r o p o s i c l o n a l u n i t of the deep s t r u c t u r e of the language of any r a t i o n a l c r e a t u r e must have two c o n s t i t u e n t s . The argument m i g h t be, f o r example, (again, a c c e p t i n g the view a l l u d e d t o above t h a t t h e u t t e r i n g of dated judgements i s a necessary c o n d i t i o n f o r r a t i o n a l i t y ) t h at the simplest form of a dated judgement must have two c o n s t i t u e n t s i n deep s t r u c t u r e — one f o r a f e a t u r e ' and one f o r a 'date'. Thus, non-empirical s y n t a c t i c s i s not a n u l l s u b j e c t , and there i s a d i s t i n c t i o n of both methodological and s u b s t a n t i a l i n t e r e s t between e m p i r i c a l a n d . n o n - e m p i r i c a l s y n t a c t i c s ; but i t i s not c l e a r whether t h e r e c a n be a u n i v e r s a l semantics which s p e c i f i e s t h e s t r u c t u r e o f any p o s s i b l e n a t u r a l human language without s p e c i f y i n g t h e , s t r u c t u r e of any f u l l y expressive language, and thus i t i s not c l e a r whether there can be any u n i v e r s a l semantics w h i c h i s e m p i r i c a l i n Chomsky's sense. Taking the emphasis i n the previous remarks, however, on 61 the p o i n t that there i s some non-empirical s y n t a c t i c study, once again, then, i t may be observed that d i s c o v e r i n g the t r u t h or f a l s i t y of the statement that any f u l l y expressive language must con t a i n a device f o r the r e c u r s i v e generation of sentence s t r u c t u r e s i s not an undertaking w i t h i n the bounds of e m p i r i c a l or d e s c r i p t i v e l i n g u i s t i c s ; r a ther i t under-takes to disc o v e r whether there could be a r a t i o n a l creature who uses a language which can t r a n s l a t e and express any thought expressable by humans, but which contains no r e c u r s i v e generation of sentences. To put t h i s i n p s y c h o l o g i c a l terms, i t seeks to di s c o v e r whether any creature whose r a t i o n a -l i t y and i n t e l l i g e n c e i s on a par with humans i n s o f a r as use of language i s concerned must have an innate language deter-mining s t r u c t u r e i n v i r t u e of which a system f o r r e c u r s i v e generation of sentences i s represented or canbe acquired as a n a t u r a l or ' f i r s t ' language. ( ' F i r s t ' or 'natural' language, since obviously, any creature whose i n t e l l e c t u a l i t y i s on a par with ours w i l l be capable of l e a r n i n g — to at l e a s t some degree of f l u e n c y — a language which r e c u r s i v e l y generates sentences.) At the same time we are checking, i t appears, whether or not i s i s p o s s i b l e that our b i o l o g i c a l l y determined language f a c u l t y (accepting that there may be a separate facu-l t y ) does not provide for the r e c u r s i v e generation of sentence s t r u c t u r e s , and that the r e c u r s i o n i s a feature imported from 62 other p r i n c i p l e s of i n t e l l i g e n c e . A l l t h i s , however, i s h i g h l y s p e c u l a t i v e , and I'd l i k e now to r e t u r n to a l e s s c o n t r o v e r s i a l l e v e l of problem and a n a l y s i s . Whatever the r o l e of the statement of n e c e s s i t y f o r any f u t u r e p s y c h o l o g i c a l explanation of our c r e a t i v e use of language, the statement of n e c e s s i t y i s c l e a r l y formulable and worthy of being t e s t e d f o r t r u t h . Accordingly, the r e s t of the study i s devoted to t e s t i n g the t r u t h of the statement found i n v a r i o u s forms i n the w r i t i n g s of prominent t h e o r i s t s of generative t r a n s f o r m a t i o n a l l i n g u i s t i c s t h a t i f we are to be able to express a l l our thoughts we must have access to a grammar which generates sentences r e c u r s i v e l y . C. On checking the n e c e s s i t y of r e c u r s i v e generation of:: sentences f o r the f u l l expression of thought. As has been elaborated above, i t i s not obviously true that using a system f o r the r e c u r s i v e generation of sentences i s necessary f o r the f u l l expression of thought; moreover, the means to check the c l a i m that such a system i s necessary are not immediately apparent. The c l a i m can be represented i n the f o l l o w i n g terms: A I f a language i s to serve as the v e h i c l e f o r the expression of a l l our thoughts, then i t s grammar must contain a s y n t a c t i c component which r e c u r s i v e l y generates sentences. 6 3 Roughly speaking, the meaning o f A i s c l e a r ; i t ' s the means whereby i t s t r u t h may be checked which i s not c l e a r . To begin with, however, the antecedent of the c o n d i t i o n a l i n A involves making judgements on what I s h a l l c a l l 'the expressive c a p a c i t y of a language'. To put i t very crudely, 'the expressive c a p a c i t y of a language' r e f e r s to how many thoughts can be expressed i n the language. What, then, does i t involve to make judgements on the expressive c a p a c i t y of a language? In order to make judgements on the expressive c a p a c i t y of a language, we need to i n q u i r e which thoughts the language can or cannot express. P r o c e d u r a l l y speaking, t h i s means to make judgements on the expressive c a p a c i t y of a language i s to compare i t i n expressive c a p a c i t y with reference to other languages (using 'language' i n i t s widest sense, to include n a t u r a l languages, a r t i f i c i a l languages, defined sub-portions of n a t u r a l languages, e t c . ) . I f we d i d not proceed i n t h i s fashion we would not be able to make judgements on the expressive c a p a c i t y of a language at a l l , s i n c e we would be making judgements on which thoughts the language could or could not express without having any expression of these thoughts by which to i d e n t i f y them. Since we can't i d e n t i f y or enumerate our thoughts without g i v i n g them expression, we can't judge the expressive capacity of a language without making paraphrase comparisons between languages, or between the whole language and parts of the language. 64 Thus, to meaningfully discuss the expressive c a p a c i t y of languages we need to introduce at l e a s t some terms by which to r e f e r to the comparitive status of languages with regard to t h e i r expressive c a p a c i t i e s . Here a language L n i s defined as equal i n expressive c a p a c i t y to another language L 2 i f and only i f any piece of discourse i n Lj_ can be paraphrased by a piece of d iscourse i n L 2 and v i c e v e r s a . I w i l l s i d e step the question of whether or not the n a t u r a l languages are equal to each other i n expressive c a p a c i t y . Rather I take i t that i n some meaningful sense we can regard the E n g l i s h language as ' f u l l y expressive', and that any other language, a r t i f i c i a l or n a t u r a l i s f u l l y expressive i f and only i f i t i s equal to E n g l i s h i n expressive c a p a c i t y . This i s not meant to beg any questions surrounding the 'Whorfian hypothesis' which, as I understand i t , i s not that what i s expres s i b l e i n Navajo, say, may be altogether i n e x p r e s s i b l e i n E n g l i s h , but rather that what i s e a s i l y expressed i n one language i n some cases may be expressed only with d i f f i c u l t y i n another. We are now i n a p o s i t i o n to see that A claims t h a t : A' I f a language i s to be a f u l l y expressive language, i . e . i f a language i s to be equal to E n g l i s h i n expressive capacity, i . e . , i f a language i s to be capable of generating a paraphrase f o r any sentence or s e r i e s of sentences i n E n g l i s h , then the grammar of that language must contain a device which r e c u r s i v e l y generates sentences. Having reformulated A i n t h i s way, the cla i m i s beginning to 65 look somewhat more checkable. The c l a i m w i l l be f a l s e i f a language e x i s t s which can generate a paraphrase for any sentence or s e r i e s of sentences i n E n g l i s h without c o n t a i n i n g a device which r e c u r s i v e l y generates sentences i n i t s grammar. I t w i l l be r e c a l l e d from the e x p o s i t i o n i n Chapter One, p a r t one, (see pages 7 - 9 and p. 19) that s e n t e n t i a l i n f i n i t y might come from two sources: there might be a device which r e c u r s i v e l y generates deep s t r u c t u r e types f o r sentences, and i n v i r t u e of which there would be an i n f i n i t y of sentences generated by the grammar; and secondly, there might be an i n f i n i t y of l e x i c a l items, and hence an i n f i n i t y of p o s s i b l e sentences — with or without a device f o r the r e c u r s i v e generation of deep s t r u c t u r e types f o r sentences. I f we exclude the r e c u r s i v e generation of sentences v i a an i n f i n i t y of l e x i c a l items, then the statement of n e c e s s i t y would a s s e r t : A' 1 Any f u l l y expressive language, i . e . , any language equal to E n g l i s h i n expressive capacity, i . e . , any language which can generate a paraphrase f o r any sentence or s e r i e s of sentences i n E n g l i s h , must contain a device f o r the r e c u r s i v e generation of deep s t r u c t u r e types. Mutatis mutandis, i f we exclude the r e c u r s i v e generation of deep s t r u c t u r e types then the statement of n e c e s s i t y would a s s e r t : A''' Any f u l l y expressive language, i . e . , any language equal to E n g l i s h i n expressive capacity, i . e . , any language which can generate a paraphrase f o r any sentence or s e r i e s of sentences i n E n g l i s h , must contain a l e x i c o n which contains an i n f i n i t y of items. And each of these can be checked. A' 1 implies that we need to 66 use sentences which c o n t a i n what was defined above as r e c u r s i v e l y derived c o n s t i t u e n t s to express a l l our thoughts. (A reminder by way of e x a m p l e s : r e c u r s i v e l y derived c o n s t i t u e n t s are s t r u c t u r e s l i k e 'noun clause', ' r e l a t i v e clause', and 'co-ordinated c l a u s e ' ) . So A' 1 i s f a l s e i f , , using sentences with no r e c u r s i v e l y d e r i v e d c o n s t i t u e n t s we can paraphrase any sentence of E n g l i s h . In rough terminology. A 1 ' a s s e r t s that t o express a l l our thoughts we need to use, complex sentences, and it. i s f a l s e i f we can express a l l our thoughts without using complex sentences. S i m i l a r l y , A' 1' i s f a l s e i f using a f i n i t e l e x i c o n one can paraphrase any sentence of E n g l i s h , I must emphasise at t h i s p o i n t that both these claims r e q u i r e extensive paraphrase analyses; before such an a n a l y s i s i s performed one would have to say that one of the claims might be true while the other i s f a l s e , or both are true or both are f a l s e . I t i s only the paraphrase a n a l y s i s which w i l l enable one to judge. Yet the statement of n e c e s s i t y i t s e l f can only be true i f e i t h e r A' 1 or A 1 ' ' i s t r u e . Thus we are faced with two p o s s i b l e courses of a c t i o n : we might i n v e s t i g a t e our language to see i f there are any l e x i c a l r e c u r s i o n s , and i f there are to see whether we need them to express our thoughts, or we could i n v e s t i g a t e the r e c u r s i v e mechanisms which generate an i n f i n i t y of deep s t r u c t u r e types (of which we know there are some) to see whether we need them to express a l l our thoughts. Checking e i t h e r one or the other does not amount to the f u l l check of the statement of n e c e s s i t y 67 but checking e i t h e r would c o n s t i t u t e a s o l i d beginning. The bulk of t h i s study w i l l be devoted to checking the t r u t h of A " , i . e . , to checking whether we need to have access to an i n f i n i t y of deep s t r u c t u r e types i n order to express a l l our thoughts. This choice i s not an a r b i t r a r y one. To begin with, Chomsky, P o s t a l , Fodor, Katz, and others s t a t e that a synchronic d e s c r i p t i o n of language contains only a f i n i t e l e x i c o n ('synchronic', not because they hold that a d i a c h r o n i c d e s c r i p t i o n w i l l i n c lude an i n f i n i t e l e x i c o n , but because t h e i r work i s , i n general, r e s t r i c t e d to problems i n synchronic d e s c r i p t i o n of language, and they don't seem to comment on the si.ze of the l e x i c o n i n a d i a c h r o n i c d e s c r i p t i o n ) . Thus, f o r most t h e o r i s t s , there are no l e x i c a l r e c u r s i o n s , and Chomsky (as can be seen from some of the passages i n c l u d e d above, e.g., the one beginning "the r e c u r s i v e property of the grammar...is r e s t r i c t e d . . . " page 30 above) c l e a r l y has the r e c u r s i v e generation of deep s t r u c t u r e types f o r sentences alone i n mind when he r e f e r s to the r e c u r s i v e generation of sentences. A l s o on a number of occasions Chomsky discounts the importance of s u b s t i t u t i o n of l e x i c a l items i n f i x e d grammatical frames with regard to the notion of c r e a t i v e aspect of language use, and t h i s f u r t h e r suggests that i n r e f e r r i n g to the r e c u r s i v e generation of sentences he i s r e f e r r i n g to the r e c u r s i v e generation of deep s t r u c t u r e types e x c l u s i v e l y i n the statement of n e c e s s i t y . 68 In the f o l l o w i n g chapter, then, I w i l l attempt to check as completely as p o s s i b l e through parpahrase a n a l a y s i s f o r the t r u t h of A 1', that we need to have access to an i n f i n i t y o f deep s t r u c t u r e types f o r sentences i n order to be able to express our thoughts; and then, i n the f i n a l chapter, I w i l l r e t u r n to the issue of l e x i c a l r e c u r s i o n s , and r a i s e the f u r t h e r question of ' r e c u r s i v e generation of paragraphs'. To repeat, we begin with the r e c u r s i v e generation of deep s t r u c t u r e types s i n c e these are the only r e c u r s i o n s p o s t u l a t e d and discussed by Chomsky, Fodor, Katz, and other generative t r a n s f o r m a t i o n a l t h e o r i s t s . 69 CHAPTER TWO A. P r e l i m i n a r y By d e f i n i t i o n , there are an i n f i n i t y of r e c u r s i v e l y derived sentences; t h e r e f o r e , i t i s impossible to show that a l l r e c u r s i v e l y d e r i v e d sentences can be paraphrased by s e r i e s of sentences none of which are r e c u r s i v e l y derived by conside-r i n g a l l r e c u r s i v e l y derived sentences. I d e a l l y , what we want to do, r a t h e r , i s to enumerate the types of r e c u r s i v e l y derived constituents.- and proceed with the a n a l y s i s using as many r e p r e s e n t a t i v e examples as warrant c o n s i d e r a t i o n . However, at the present date, the theory of E n g l i s h grammar has not been developed s u f f i c i e n t l y to permit one to c o n s t r u c t a l i s t of the kinds of r e c u r s i v e l y derived c o n s t i t u e n t s with any confidence. How then can we d e l i m i t the subject matter? I propose to do i t with a set of core examples which seem to be r e p r e s e n t a t i v e of r e c u r s i v e l y derived sentences, and to use them as the focus around which the grammatical g e n e r a l i -zations a t t a c h . By the end of Chapter Two, we do indeed end up with a l i s t of kinds of r e c u r s i v e l y derived c o n s t i t u e n t s but i t i s important to remember that the l i s t was not derived t h e o r e t i c a l l y , but b u i l t up sometimes i n an ad hoc way around the cases that present themselves f o r a n a l y s i s . This r e f l e c t s the more general methodological c o n s i d e r a t i o n that to demonstrate 70 the e l i m i n a b i l i t y of r e c u r s i v e l y derived sentences what I must show i s that given any sentence which contains a r e c u r s i v e l y derived c o n s t i t u e n t regardless of whether that c o n s t i t u e n t i s to be analyzed as, say, a 'noun phrase' or 'verb phrase* i n deep s t r u c t u r e , we can f i n d a paraphrase which uses no recur-s i v e l y d e r i v e d sentences. The hope i s , then, that where an a n a l y s i s i s c o n t r o v e r s i a l , the controversy w i l l not be over whether the d e r i v a t i o n involves the a p p l i c a t i o n of a r e c u r s i v e s e t of r u l e s . To use an example which s h a l l be discussed below (see page 94) i n the case of the d i s t i n c t i o n between 'noun phrase complement' and 'verb phrase complement', grammarians would agree that "Mary persuaded John to go" contains a r e c u r s i v e l y derived c o n s t i t u e n t , even i f they disagree about whether that c o n s t i t u e n t i s dominated by 'noun phrase 1 or 'verb phrase' i n deep s t r u c t u r e . In attempting to eliminate r e c u r s i v e l y derived c o n s t i t u e n t s , I w i l l b a s i c a l l y f o l l o w Jacobs and Rosenbaum's treatment of such s t r u c t u r e s i n E n g l i s h Trans formational Grammar. Roughly speaking, r e c u r s i v e l y derived c o n s t i t u e n t s are of two types: embeddings and c o - o r d i n a t i o n s . In a sentence l i k e : 2.1 The dog which ate the cat which chased the mouse which nib b l e d the cheese i s here. there are a s e r i e s of embedded clauses. On the other hand, i n a sentence l i k e : 71 2.2 One plus one i s two, and two plus two i s four, and three plus three i s s i x , and four plus four i s eight, and f i v e plus f i v e i s ten. there are a s e r i e s of co-ordinated c l a u s e s . From a s t r i c t l y t e c h n i c a l standpoint, co-ordinations must be thought of as a kind of embedding because the co-ordinated clauses must s t i l l be analysed as wholly enclosed or embedded i n s i d e the l a r g e s t s e n t e n t i a l u n i t . But i t w i l l s u i t our purposes here to simply note the 1enclosedness' of the clauses i n 2.1 i n cont r a s t with the 'independence' of the clauses i n 2.2. I t i s c l e a r t h a t there are a number of d i f f e r e n t kinds of embeddings. 2.1 contains a s e r i e s of r e l a t i v e c l a u s e s . The f o l l o w i n g sentence, though, contains some s o r t of noun clause: 2.3 E i n s t e i n b e l i e v e d that Newton b e l i e v e d that A r i s t o t l e b e l i e v e d that P l a t o b e l i e v e d that the world was made of Forms. And the f o l l o w i n g sentence contains ' i n f i n i t i v a l ' clauses (which may or may not turn out to be a species of 'noun clause') 2.4 Edna persuaded Mary t o persuade John to persuade B i l l to persuade Florence to persuade H a r r i e t to re t u r n to England. We can a l s o have embedding of p a r t i c i p i a l clauses, e.g.: 2.5 The monarch's preventing the mob's s t a r t i n g to s i n g about the previous r i s i n g of the people was opposed by the treacherous a d v i s o r . F i n a l l y , the possessive can be used to create s u c c e s s i v e l y embedded s t r u c t u r e s , e.g: 72 2.6 Your f r i e n d ' s uncle's dog's c o l l a r ' s r i n g ' s gleam i s b r i g h t . Thus we have a number of d i f f e r e n t kinds of r e c u r s i v e l y embeddable s t r u c t u r e s . Co-ordinations, a l s o , are of d i f f e r e n t s o r t s . There are simple s e n t e n t i a l c o - o r d i n a t i o n s , as i n 2.2: 2.2 One plus one i s two, and two plus two i s four, and three plus three i s s i x , and four plus four i s eight, and f i v e plus f i v e i s ten. In a d d i t i o n there are more complex s e n t e n t i a l c o-ordinations, as i n : 2.7 I f , i f Harry loves Edna, then Edna loves Harry, then Edna loves Harry, i f J u l i a loves Thomas, i f , i f J u l i a loves Thomas then Harry no longer loves J u l i a , i f , i f Harry no longer loves J u l i a then Harry loves Edna. There are a l s o co-ordinations which do not immediately appear to be s e n t e n t i a l . In p a r t i c u l a r , the f o l l o w i n g sentence apparently contains co-ordinated noun phrases: 2.8 Jones and Smith and Brown and Green were d i s c u s s i n g the l a t e s t news. And the f o l l o w i n g sentence contains coordinated adverbs and a d j e c t i v e s . 2.9 She walked s w i f t l y , determinedly, c o n f i d e n t l y , smoothly, q u i e t l y , and circumspectly to the b r i g h t , b i g , o l d , smooth, s o l i d , impressive, V i c t o r i a n mansion. I have presented, then, nine sentences each of which apparently contains a r e c u r s i v e l y derived constituent, sentences which, moreover, appear to d i f f e r from one another i n important aspects. I must emphasize, however, that t h i s l i s t i s not 7 3 t h e o r e t i c a l l y motivated, and that i t merely provides us with a po i n t of departure. i f there i s any 'theory' to j u s t i f y t h i s l i s t , then i t i s the theory of surface s t r u c t u r e , and,as we s h a l l see, the l i s t i s misleading i n some respects f o r that reason. Before s t a r t i n g i n on the a n a l y s i s , i t i s important to d i s t i n g u i s h between ' r e c u r s i v e l y derived c o n s t i t u e n t s ' on the one hand, and 'constituents which can be embedded' and 'con-s t i t u e n t s which can be co-ordinated* on the other. At the t h e o r e t i c a l l e v e l , not a l l c o n s t i t u e n t s which give r i s e to embeddings or co-ordinations need to be r e c u r s i v e l y derived, and i n f a c t not a l l embeddings are r e c u r s i v e l y derived, nor (according to Yngve) are a l l co-ordinations r e c u r s i v e l y d e r i v e d . Consider a u x i l i a r i e s , f o r example. A sentence l i k e : 2.10 John might have been being obtuse, contains a s e r i e s of (embedded or co-ordinated -- i t ' s not c l e a r which — ) a u x i l i a r i e s . However, there are l i m i t s to the embedding or c o - o r d i n a t i n g of a u x i l i a r i e s i n deep s t r u c t u r e ; i t i s c l e a r that a u x i l i a r i e s are s t r i c t l y ordered and l i m i t e d i n number as they attach to any given verb i n any given sentence or c l a u s e . Ross has a theory that modals and some a u x i l i a r i e s are main verbs of embedded clauses i n deep s t r u c t u r e (Ross, A u x i l i a r i e s as Main Verbs) but whether or 74 not Ross's theory i s a good one, we would not p o s t u l a t e r e c u r s i v e r u l e s i n the base from which to der i v e the compounding of a u x i l i a r i e s . I turn now to the substance of the paraphrase a n a l y s i s , beginning with r e l a t i v e c l a u s e s . B . R e l a t i v e Clauses. To argue that i t i s never necessary to use a r e l a t i v e clause to express a thought i n E n g l i s h , one must show t h a t whenever there i s a sentence with one or more r e l a t i v e clauses there w i l l always be a v a i l a b l e a sentence, or a s e r i e s of sentences, which do not have any r e l a t i v e clauses and yet which express the thought of the o r i g i n a l sentence or which roughly paraphrase i t . Since there i s an i n f i n i t y of sentences with r e l a t i v e clauses the best we can do i s to take samples of the d i f f e r e n t kinds of r e l a t i v e c l a u s e s . Thus, w e ' l l consider a p p o s i t i v e r e l a t i v e clauses and r e s t r i c t i v e r e l a t i v e clauses? w i t h i n the l a t t e r c l a s s w e ' l l look at clauses with a 'which' or 'who' or 'that' r e l a t i v i z e r . Further w e ' l l consider clauses i n which the r e l a t i v i z e r i s preceded by a p r e p o s i t i o n , r e l a t i v e clauses occuring i n s i d e each other, and so f o r t h , t r y i n g to include as many d i f f e r e n t types of clauses as p o s s i b l e . There are two b a s i c kinds of r e l a t i v e clauses: a p p o s i t i v e r e l a t i v e clauses and r e s t r i c t i v e r e l a t i v e c l a u s e s . The d i s t i n c t i o n between the two i n the w r i t t e n v e r s i o n i s marked by the f a c t 75 that commas surround a p p o s i t i v e c l a u s e s . No commas surround r e s t r i c t i v e r e l a t i v e c l a u s e s . (Stress d i f f e r e n c e s i n o r a l communication mark the same d i s t i n c t i o n . ) T y p i c a l examples of a p p o s i t i v e and r e s t r i c t i v e r e l a t i v e clauses, r e s p e c t i v e l y , are: 2.10a New York, which used to be l o v e l y , has r e c e n t l y become u n l i v e a b l e . 2.10b The man who knocked on the door yesterday returned again today. There i s a s i g n i f i c a n t d i f f e r e n c e i n the fu n c t i o n and use of these two kinds of r e l a t i v e c lauses. Whereas r e s t r i c t i v e r e l a t i v e clauses are used to i n d i v i d u a t e , a p p o s i t i v e r e l a t i v e clauses are not. Thus i n 2.10b the r e s t r i c t i v e clause t e l l s us which man i s being r e f e r r e d to as 'the man', whereas i n 2.10a the a p p o s i t i v e clause has no such f u n c t i o n ; i t does not s p e l l out 'which' New York i s being r e f e r r e d t o . In general, a p p o s i t i v e r e l a t i v e clauses are easy to eliminate i n paraphrase, and so I s h a l l not dwell on them at any length. 2.10a f o r example can be paraphrased by: 2.10c New York used to be l o v e l y . But i t has r e c e n t l y become u n l i v e a b l e . I t i s worth observing here that a p p o s i t i v e r e l a t i v e clauses are apparently a form of conjunction. 2.10a i s roughly synonymous with: 2.10d New York, and (though?) i t used to be l o v e l y , has r e c e n t l y become u n l i v e a b l e . 76 But there i s no synonymy between: 2.10e The man who yesterday introduced himself as John Doe i s back again today. and: 2.10f The man, and he yesterday introduced himself as John Doe, i s back again today. In other words r e s t r i c t i v e clauses cannot be construed as having a conjunctive r e l a t i o n s h i p with the noun phrases which they r e s t r i c t whereas a p p o s i t i v e s can be construed as having a conjunctive r e l a t i o n s h i p with the noun phrases to which they are i n a p p o s i t i o n . Notice that i t i s p o s s i b l e to have a conjoined sequence w i t h i n a r e s t r i c t i v e clause, as i n : 2.10g The man who came to the door yesterday and who knocked and who then entered and introduced h i m s e l f as John Doe, i s back again today. But the r e l a t i o n s h i p between the conjoined sequence and the preceding noun phrase 'the man' cannot be c o n j u n c t i v e . There i s no synonymy between 2.10g and: 2.10h The man, and he came to the door yesterday, and he knocked, and he then entered and introduced himself as John Doe, i s back again today. since the clauses i n 2.10h do not serve to i n d i v i d u a t e they do not s p e c i f y which man. Possibly, i t i s the conjunctive character of the a p p o s i t i v e clauses which makes the task of e l i m i n a t i n g them i n paraphrase e s p e c i a l l y easy. The paraphrase e l i m i n a t i o n of r e s t r i c t i v e r e l a t i v e clauses i s more c h a l l e n g i n g . As mentioned above, i n general, r e s t r i c t i v e r e l a t i v e clauses serve as i n d i v i d u a t i n g devices, so the broad question 77 i s whether we can i n d i v i d u a t e w i t h o u t using r e s t r i c t i v e r e l a t i v e c l a u s e s . For example, i n : 2.11 The woman who entered t h e p r e s s c l u b i s the a u t h o r of s e v e r a l books. 'who entered the press club' s e r v e s t o i n d i v i d u a t e t h e woman, so that the s i n g u l a r reference implied b y t h e d e f i n i t e a r t i c l e can occur. Our problem i s to f i n d a way t o in<3ivicXm*-e withou*' r e l a t i v e clauses l i k e 'who entered t h e p r e s s c l u b ' . I f we consider some of t h e r e c e n t work done on d e f i n i t e -d e s c r i p t i o n and s i n g u l a r terms, t h i s might seem l i k e a formidable task. I'm t h i n k i n g i n p a r t i c u l a r o f Zeno V e n d l e r ' s Singular Terms (Chapter 2 i n L i n g u i s t i c s i n Philosophy) , m whi-Vendler argues that the r e l a t i v e c l a u s e s t r u c t u r e 'the Noun wh w i l l introduce a s i n g u l a r term " i f and only i f t h e s e n t e n c e 'There i s an N wh....' i s e n t a i l e d b y the discourse", and fur ther that the de f i n i t e a r t i c l e with or without a r e l a t i v e clause f o l l o w i n g i t " i s a f u n c t i o n of a r e s t r i c t i v e c l a u s e attached to the noun" so that even i n the cases where t h e r e l a t i v e clause i s absent, t h e d e f i n i t e a r t i c l e p r e s u p p o s e s that some r e l a t i v e clause can go there, t h a t at some l e v e l i t is understood to be there, or that i t can be invoked through various c o n s i d e r a t i o n s . But the d i f f i c u l t y i s only apparent and not r e a l . Vendler' concern i s centred on the nature o f s i n g u l a r terms as s i g n a l l e d by the presence of the d e f i n i t e a r t i c l e . He says t h a t wherever 78 there's a d e f i n i t e a r t i c l e , then there i s a r e s t r i c t i v e r e l a t i v e clause present or presupposed. But our concern i s simply with r e l a t i v e clauses and whether or not i t i s necessary to use r e l a t i v e clauses to express our thoughts. Vendler's main concern i s not with the primacy of the r e l a t i v e clause as an expressive d e v i c e . In f a c t Vendler's treatment of r e l a t i v e clauses, r a t h e r than being i n c o n s i s t e n t , provides m a t e r i a l which d o v e t a i l s with the m a t e r i a l we w i l l examine here. Vendler argues that a sentence l i k e : 2.12 The woman i s the author of s e v e r a l books, presupposes a clause such as i s present i n 2.11: 2.11 The woman who entered the press club i s the author of s e v e r a l books. and that 2.11 presupposes: 2.13 There i s a woman who entered the press c l u b . (Although the copula doesn't imply 'real existence'.) I s h a l l take t h i s one step to the s i d e (or maybe one step further) and say that 2.13 can be paraphrased by: 2.14 Some woman entered the press c l u b . so that a sentence l i k e 2.11 can be paraphrased by a p a i r of sentences n e i t h e r of which contains a r e l a t i v e c l a u s e : 2.15 Some woman entered the press c l u b . That woman i s the author of several books. And such an a n a l y s i s i s at l e a s t roughly c o n s i s t e n t with Vendler's (A p o s s i b l e inconsistency i s over whether or not Vendler — despite the f a c t that i t i s not h i s main concern — does end 79 up implying the primacy of the r e l a t i v e clause as an expressive device. To me at l e a s t whether he does i s not c l e a r ; i f he does then I b e l i e v e that the a n a l y s i s which follows shows such an i m p l i c a t i o n to be unwarranted.) Let's look, then, at some of the problems surrounding the s t y l e of paraphrase f o r r e s t r i c t i v e r e l a t i v e clauses exemplified i n 2.15. F i r s t there i s the problem of what I s h a l l c a l l 'the l o s t p r e s u p p o s i t i o n ' . I f I say: 2.11 The woman who entered the press club i s the author of s e v e r a l books. then I am presupposing t h a t : 2.14 Some woman entered the press c l u b . But i f I say: 2.15 Some woman entered the press c l u b . That woman i s the author of s e v e r a l books. then I am no longer presupposing t h a t : 2.14 Some woman entered the press club, because I am a s s e r t i n g i t . A presupposition has been l o s t ; of course, the p r o p o s i t i o n a l content hasn't disappeared but the p r o p o s i t i o n qua presupposition has become transformed i n t o a p r o p o s i t i o n qua a s s e r t i o n . And the question i s , when a presupposition becomes e x p l i c i t or when a presupposition becomes an a s s e r t i o n , i s there something l o s t from the communication? I would l i k e to suggest that whereas there i s something clumsy about having to ass e r t where we would often (though not always presuppose, there i s nothing v i t a l l o s t from the 80 communication i f we are subject to such a c o n s t r a i n t . We can see t h i s , I think, i n the f a c t that when communication f a i l s we oft e n c l a r i f y and disambiguate by making presuppo-s i t i o n s e x p l i c i t i n j u s t t h i s way. When problems a r i s e we w i l l i n f a c t use a p a i r of sentences l i k e 2.15 to s u b s t i t u t e for a sentence l i k e 2.11. Thus, for example, i n speaking to a person who i s not f l u e n t i n the language, there i s the tendency to avoid r e l a t i v e clauses and use the paraphrase form. Next I would l i k e to b r i e f l y examine the e f f e c t s of having to use two sentences i n the place of one when the o r i g i n a l includes a negative c o n s t r u c t i o n . Consider a sentence l i k e : 2.16 I t ' s f a l s e that the woman who entered the press club i s the author of se v e r a l books. or: 2.17 John denied that the woman who entered the press club i s the author of sev e r a l books. This kind of sentence, containing a negative which covers both the main clause and the r e l a t i v e clause of the o r i g i n a l sentence (2.11) can create some problems. This kind of sentence i s q u i t e d i f f e r e n t from the kind of sentence which contains a negative c o n s t r u c t i o n wholly i n s i d e one or the other of the clauses of 2.11. By contrast-with 2.16 and 2.17, consider 2.18 The woman who entered the press club i s not the author of s e v e r a l books. 81 and: 2.19 The woman who d i d not enter the press club i s the author of s e v e r a l books. The c o n t r a s t i s given i n the f a c t that the l a t t e r p a i r can be parpahrased using the s t y l e developed f o r 2.11. The f o l l o w i n g two sentences paraphrase 2.18 and 2.19 r e s p e c t i v e l y : 2.20 Some woman entered the press c l u b . That woman i s i s not the author of several books. 2.21 Some woman d i d not enter the press c l u b . That woman i s the author of s e v e r a l books. Of course 2.21 rin g s odd to the ear, but t h i s oddness merely r e f l e c t s the p e c u l i a r i t y of i n d i v i d u a t i n g with a negative as was done i n 2.19. There i s nothing l o g i c a l l y impossible about such an i n d i v i d u a t i o n , and i n f a c t such i n d i v i d u a t i o n s do occur when i t i s understood that only one woman d i d not enter the press club. But when the negative occurs outside the r e l a t i v e clause, there i s a problem i n s o f a r as the r e l a t i o n s h i p between the negative and the presupposition i s concerned. For example, 2.16 could not always be paraphrased by 2.20 because a person could a s s e r t a p r o p o s i t i o n to be f a l s e on the grounds that i t s presupposition i s f a l s e . That i s , one could say, "It's f a l s e that the woman who entered the press club i s the author of several books", adding "there was no such woman — I deny that any woman has entered the press c l u b . " Whether such a use of 82 ' f a l s e ' or 'deny' i s l e g i t i m a t e i s c o n t r o v e r s i a l , but even f o r those f o r whom such a use of ' f a l s e ' or 'deny' i s l e g i t i m a t e , a c l e a r statement of the problem i n d i c a t e s i t s s o l u t i o n . I f a person denies a p r o p o s i t i o n on such grounds, then the grounds for the d e n i a l can a l s o be made e x p l i c i t . So that a p r o p o s i t i o n l i k e 2.16 (or 2.17, adding a sentence which makes reference to John) can be paraphrased e i t h e r by 2.20 or by something l i k e : 2.23 I deny that that woman i s the author of s e v e r a l books. No woman has entered the press c l u b . The f i r s t sentence i n 2.23 contains a r e c u r s i v e l y derived s t r u c t u r e , namely a noun clause, or, to use the terminology to be used below, a noun phrase complement. But t h i s i s not a r e l a t i v e clause, and we can leave the task of d e a l i n g with the independent e l i m i n a b i l i t y of noun phrase complements t i l l l a t e r . T h i r d , I would l i k e to examine the e f f e c t of t h i s kind of paraphrase on the l o g i c a l focus of sentences with r e l a t i v e c l auses. I f someone says "The woman who entered the press club i s the author of s e v e r a l books" i t might be i n response to the question, "What e l s e do you know of the woman who entered the press club, besides that she entered the press club?" or, to the question, "How many books i s the woman who entered the press club the author o f ? " or to se v e r a l other questions of t h i s i l k . Of course I do not mean to imply that the question must 83 have been asked; nor do I want to a s s e r t that the l o g i c of discourse requires that every utterance, or even every a s s e r t i o n , must be o f f e r e d i n answer to some such question. But when a p r o p o s i t i o n ijs o f f e r e d i n answer to some such question, e x p l i c i t l y or i m p l i c i t l y , then the p r o p o s i t i o n can be s a i d to have ' l o g i c a l focus'. To i l l u s t r a t e , i f the p r o p o s i t i o n i s o f f e r e d i n response to the question "Who i s the author of s e v e r a l books?" then the l o g i c a l focus of the statement i s on the noun phrase, "The woman who entered the press c l u b " . To d i g r e s s f o r a moment, the r e l a t i o n s h i p between the l o g i c of d i s c o u r s e and s y n t a c t i c a n a l y s i s must be strong when i t comes to the c o n s i d e r a t i o n of l o g i c a l focus s i n c e i n some languages l o g i c a l focus has s y n t a c t i c r e p r e s e n t a t i o n i n surface s t r u c t u r e . In Japanese, f o r instance, i f the l o g i c a l focus i s on the s u b j e c t of the sentence then the usual subject i n f l e x i o n 'wa' i s r e p l a c e d by the subject i n f l e x i o n 'ga'. In answer to the question "Who i s here?" one says, "Watashiga kokoni imasu" ("I" "here" "am"); i n answer to the question "Where are you?" one says, "Watashiwa kokoni imasu" ("I" "here" "am"). Thus i n at l e a s t one language the l o g i c of discourse receives syntac-t i c r e p r e s e n t a t i o n at the surface l e v e l , suggesting the p o s s i -b i l i t y that some features of the l o g i c of discourse could be represented i n a l l languages i n deep s t r u c t u r e . But, to r e t u r n to the question at hand, we want to know 84 what happens to the l o g i c a l focus of a sentence with a r e l a t i v e clause when i t i s parpahrased using the s t y l e developed above. There i s no problem when the statement (or statements i n the paraphrase version) are given i n response to an e x p l i c i t question. Thus, i n the f o l l o w i n g exchange, the l o g i c a l focus i s q u i t e c l e a r : 2.24 Mr. A: Who i s the author of s e v e r a l books? Mr. B: Some woman entered the press c l u b . That woman i s the author of s e v e r a l books„ But can the l o g i c a l focus be made when the question i s not e x p l i c i t ? I want to suggest that i t can be made as e x p l i c i t l y as i t can i n the sentence using the r e l a t i v e c l a use. Remembering that there are no surface marks i n E n g l i s h (which r e c e i v e normal orthographic r e p r e s e n t a t i o n at any rate) to mark the c o n s t i t u e n t which bears the l o g i c a l focus, there i s nothing to suggest that the presence of a r e l a t i v e clause i n any way helps to e s t a b l i s h the l o g i c a l focus. The presence of a r e l a t i v e clause does not mean that the l o g i c a l focus w i l l be on the noun phrase which includes the r e l a t i v e clause; the l o g i c a l focus could a l s o be on the whole of the verb phrase or on some pa r t of i t . Having two separate sentences i s s i m i l a r l y no bar to the l o g i c a l focus appearing i n some con s t i t u e n t of one or the other of the sentences. The problem would not be the i m p o s s i b i l i t y of having l o g i c a l focus. But there might be a problem — or so the o b j e c t i o n would run -- i n that we have two sentences i n the place of one, and 85 so we might have 'too much' l o g i c a l focus. The paraphrase discourse might suggest more questions than the o r i g i n a l d i s c o u r s e . Saying, "Some woman entered the press c l u b . That woman i s the author of s e v e r a l books," i t could be argued, seems to r e q u i r e that each of the sentences has a p a r t i c u l a r l o g i c a l focus, and thus that the f i r s t sentence i s meant to convey information i n response to one of a number of p o s s i b l e questions, and l i k e w i s e the second. Saying the one sentence, "The woman who entered the press club i s the author of s e v e r a l books" does not, on the other hand, seem to i n v o l v e the l o g i c a l focus of the presupposed p r o p o s i t i o n , "Some woman entered the press c l u b . " In general i t looks as though presuppositions do not c a r r y l o g i c a l focus, and i t could be argued then that s i n c e the paraphrase v e r s i o n c a r r i e s more l o g i c a l focus than the o r i g i n a l , the paraphrase should be r e j e c t e d as inadequate. But the argument i s not strong, as there i s no reason to say that a p r e s u p p o s i t i o n must c a r r y more l o g i c a l focus when i t i s asserted than when i t remains l a t e n t i n a r e l a t i v e clause, or that i t must c a r r y l o g i c a l focus when i t i s asserted while i t need not so long as i t remains l a t e n t i n a r e l a t i v e c l ause. I f we consider the p r o p o s i t i o n , "Some woman entered the press c l u b " out of context, there are many-questions we can imagine i t being o f f e r e d i n answer to. I t might be o f f e r e d i n answer to, "Who entered the press club?" or i n answer to "Was i t a 8 6 woman who e n t e r e d t h e p r e s s c l u b ? " o r i n answer t o " ( S o r r y I d i d n ' t h e a r you,) D i d some woman e n t e r t h e p r e s s c l u b , o r l e a v e t h e p r e s s c l u b ? " and so on. But i f t h e p r o p o s i t i o n i s b e i n g o f f e r e d as t h e t r a n s f o r m a t i o n o f a p r e s u p p o s i t i o n i n t o an a s s e r t i o n t h e n i t i s b e i n g o f f e r e d i n answer t o none o f t h e s e q u e s t i o n s . I t i s d i f f i c u l t t o s u p p l y a q u e s t i o n s u c h t h a t i t can be a n s w e r e d b y "Some woman e n t e r e d t h e p r e s s c l u b . " and s u c h t h a t t h e answer' i s t h e a s s e r t i o n , o f a p r e s u p p o s i t i o n . I n f a c t , we c a n ' t r e a l l y c a l l i t a p r e s u p p o s i t i o n any more s i n c e i t s s t a t u s as p r e s u p p o s i t i o n i s d e p e n d e n t on t h e r e l a t i v e c l a u s e w h i c h t h e s p e a k e r i s now f o r b i d d e n f r o m u s i n g . What t h e n i s i t ? I n some s e n s e i t i s t h e c o n v e y i n g o f i n f o r m a t i o n w h i c h i s n e c e s s a r y f o r t h e s p e a k e r t o c o n v e y i n o r d e r t o be a b l e t o c o n t i n u e t o t h e s u b s t a n c e o f what he w a n t e d t o communicate. I n t e r m s o f t h e l o g i c o f d i s c o u r s e . , i t i s t h e i n t r o d u c t i o n o f a f a c t i n terms o f w h i c h o t h e r f a c t s t o f o l l o w c a n become i n t e l l i g i b l e . T h i s i s not a p e c u l i a r f u n c t i o n , but one w h i c h i s o f t e n p e r f o r m e d i n d i s c o u r s e w i t h o u t t h e r e l a t i v e c l a u s e e ven when t h e r e l a t i v e c l a u s e i s a v a i l a b l e . I n some c a s e s we c o u l d n o t r e p l a c e t h e s e p a r a t e s e n t e n c e b y t h e r e l a t i v c l a u s e , b u t I c a n n o t s e e any r e a s o n s , o r c a n n o t t h i n k o f any s i t u a t i o n s i n w h i c h we c a n n o t f i l t e r t h e r e l a t i v e c l a u s e i n t o a s e p a r a t e s e n t e n c e , and t h i s i s t h e c r u c i a l p o i n t . I f a s p e a k e r s a y s , "A man k n o c k e d on t h e do o r y e s t e r d a y , a 87 i t turned out that he was your cousin, " he could n o t have s a i d , "The man who knocked on t h e d o o r y e s t e r d a y t u r n e d o u t t o be your c o u s i n " unless the h e a r e r a l r e a d y knew t h a t some man knocked on the door yesterday. B u t whenever someone s a y s "The man who knocked on the d o o r y e s t e r d a y t u r n e d o u t to be your c o u s i n " then i t i s assumed t h a t t h e l i s t e n e r a l r e a d y knows that some man knocked on the door y e s t e r d a y , -,o t h a t -using two sentences, "Some man knocked on t h e d o o r y e s t e r d a y . T h a t man turned out to be your cousin" i s d e f i c i e n t o n l y i n t h e s e n s e that i t r e i t e r a t e s shared information, and t h u s commits a s o r t of 'discourse redundancy'. The problem of m u l t i p l i c a t i o n of l o g i c a l f o c u s has t h e n revealed i t s e l f as the problem of t h e l o s t p r e s u p p o s i t i o n I n d i s g u i s e . We are a s s e r t i n g what could have been p r e s u p p o s e d ; and again I would say that such forced e x p l i c i t n e s s i s n o t a bar to communication. Let's look now at the d i f f e r e n t kinds of r e l a t i v e c l a u s e c o n s t r u c t i o n . In 2.11 the subject of t h e ma i n clause was a l s o t h e s u b j e c of the r e l a t i v e c l ause. Let's consider a sentence i n w h i c h t h e subject of the main clause i s the o b j e c t of the v e r b i n the r e l a t i v e clause, e.g: 2.25 The woman whom you not i c e d i s t h e a u t h o r of s e v e r a l books. 88 Here 'you' i s the subject of the r e l a t i v e clause and 'the woman' i s the object of the r e l a t i v e clause verb 'noticed'. This does not, however, s u b s t a n t i a l l y a f f e c t the p a r a p h r a s a b i l i t y of the sentence. Thus we can paraphrase the sentence with: 2.26 You n o t i c e d some woman. That woman i s the author of s e v e r a l books. Next, consider sentences with d i f f e r e n t a r t i c l e s or determiners. In 2.11, the subject of the main clause included the a r t i c l e 'the'. The nature of the r e l a t i v e clause changes q u i t e s u b s t a n t i a l l y when the a r t i c l e 'a' i s used. Consider f o r example: 2.27 A woman who entered the press club i s the author of se v e r a l books. This p a r t i c u l a r sentence i s , to me at l e a s t s u r p r i s i n g l y u n i n t e l l i g i b l e . I say ' s u r p r i s i n g l y ' because there i s nothing odd about using the a r t i c l e 'a' with a noun which heads a r e l a t i v e c l a u s e . Thus: 2.28 A woman who enters a press club i s l i k e l y to be a j o u r n a l i s t . i s p e r f e c t l y i n t e l l i g i b l e . The d i f f e r e n c e s between 2.27 and 2.28 i n terms of presuppositions, conditions f o r i n t e l l i g i b i l i t y e t c . are f a s c i n a t i n g , but I s h a l l not enter i n t o a d i s c u s s i o n of them. I s h a l l simply mention that using the a r t i c l e 'a' often implies some kind of g e n e r a l i t y , and that t h i s g e n e r a l i t y i s not a hindrance to p a r a p h r a s a b i l i t y . 89 Thus 2.28 can be paraphrased by: 2.29 Some women enter press c l u b s . These women are l i k e l y to be j o u r n a l i s t s . There i s a s i m i l a r s t o r y f o r r e p l a c i n g the a r t i c l e 'the' with the determiner 'every' 'each' or 'any'. I w i l l g ive only one example, u s i n g 'each'. 2.30 Each woman who enters a press club i s the author of se v e r a l books. And t h i s can be paraphrased by: 2.31 Some women enter press c l u b s . Each of these women i s the author of s e v e r a l books. Some people w i l l doubtless argue that one of the d i f f e r e n c e s between sentences with r e l a t i v e clauses i n which the head noun of the r e l a t i v e clause i s preceded by 'the' and sentences with r e l a t i v e s i n which the head noun i s preceded by 'a 1 or 'every' or 'any' or 'each' i s that the former i n v o l v e presuppo-s i t i o n s which the l a t t e r do not. I t might be s a i d , f o r example, 2.31 i s u n s a t i s f a c t o r y as a paraphrase f o r 2.30 because there might be s i t u a t i o n s i n which we would say that each woman who enters a press club i s the author of s e v e r a l books when there are no women who enter press clubs . This i s a p o s s i b i l i t y precluded by 2.31 si n c e i t a f f i r m s that some women enter press c l u b s . On the other hand, i t ' s not easy to think of convincing cases, and even assuming for the sake of the argument that a convincing case could be found, another avenue of paraphrase 90 i s open. I f there i s no pr e s u p p o s i t i o n i n 2.30 that some women enter press clubs, then 2.30 can be paraphrased' by a statement of inference, as i n : 2.3 2 I f a woman enters a press club, then she i s the author of s e v e r a l books. This sentence, of course, contains a r e c u r s i v e l y derived s t r u c t u r e as i t involves s e n t e n t i a l c o - o r d i n a t i o n ; but at l e a s t i t has no r e l a t i v e clause, and co-ordinations s h a l l be discussed below. Next, consider a sentence l i k e : 2.33 The woman whose husband entered the press club i s the author of s e v e r a l books. Here the subj e c t of the main clause i s n e i t h e r the subject nor the o b j e c t of the r e l a t i v e clause, but i s a possesive m o d i f i e r of the noun 'husband' which i s the subject of the r e l a t i v e c l a u s e . This sentence can be paraphrased without d i f f i c u l t y f o l l o w i n g the f a m i l i a r p a t t e r n : 2.34 Some woman's husband entered the press c l u b . That woman i s the author of s e v e r a l books. I f anyone i s dubious about the reference of that woman'in the second sentence then the r e l a t i o n s h i p s can be made p e r f e c t l y e x p l i c i t as i n : 2.35 Some man entered the press c l u b . That man's wife i s the author of s e v e r a l books. Continuing on the theme of the d i f f e r e n t r e l a t i v i z e r s , consider: 2.36 The place where she works i s the press c l u b . Here 'where' i s the r e l a t i v i z e r instead of 'who' or 'whom' or 'whose'. Sentences l i k e t h i s are not troublesome, and can be 91 paraphrased i n the usual way: 2.37 She works at some p l a c e . That place i s the press c l u b . This p a i r of sentences reads oddly but then so does 2.36. I f there was no reason why: 2.38 She works at the press c l u b . couldn't have been used instead of 2.36, then 2.38 would not need a paraphrase a t a l l . And i f , f o r some reason (although I can't invent one) 2.36 i s not paraphraseable by 2.38 then 2.37 presumably would c a r r y the whole load and s p e c i a l focus of 2.36. There are other s i m i l a r v a r i a n t s on the r e l a t i v i z e r ; there are sentences with r e l a t i v i z e r s 'how' 'why' and 'when*. I ' l l con-s i d e r another example using 'why'. 2.39 The reason why she entered the press club was to give a speech. Again there i s a paraphrase v e r s i o n using only a s i n g l e sentence, as w e l l as the v e r s i o n using the standard p a t t e r n using a p a i r of sentences. 2.40 She entered the press club to give a speech. 2.42 She entered the press club f o r some reason. The reason was to give a speech. Both these v e r s i o n s contain i n f i n i t i v a l c lauses; l a t e r I attempt to show that i n f i n i t i v a l clauses are el i m i n a b l e i n paraphrase. Consider next a sentence i n which the r e l a t i v i z e r i s preceded by a p r e p o s i t i o n : 2.42 The woman to whom you l i s t e n e d i s the author of se v e r a l books. 92 In the r e l a t i v e clause 'the woman' i s the object of the p r e p o s i t i o n 'to'. We can thus paraphrase as f o l l o w s : 2.43 You l i s t e n e d to some woman. That woman i s the author of s e v e r a l books. We must a l s o consider, f i n a l l y , r e l a t i v e clauses o c c u r i n g i n s i d e questions and imperatives. 2.44 Who was the woman who entered the press club? With the question there i s s t i l l the pr e s u p p o s i t i o n t h a t some woman entered the press club, so we can paraphrase w i t h : 2.45 Some woman entered the press c l u b . Who was t h a t woman? And the imperative works i n the same way: 2.46 Name the woman who entered the press c l u b . 2.47 Some woman entered the press c l u b . Name that woman. While the examples given here are by no means exhaustive even of the types of r e l a t i v e clauses c o n s t r u c t i o n , I b e l i e v e that enough of the d i f f e r e n t types have been e x h i b i t e d to warrant the co n c l u s i o n that r e l a t i v e clauses as they occur i n i n d i v i d u a l sentences are eliminable i n paraphrase. But what of r e l a t i v e clauses which are m u l t i p l y embedded i n one another? A f t e r a l l , what I am t r y i n g to demonstrate i s that the r e c u r s i v e devices which generate m u l t i p l e embeddings and s u c h - l i k e constru-c t i o n s are not necessary f o r language to ca r r y on i t s expressive f u n c t i o n s . We want to look then at a sentence which contains r e l a t i v e clauses embedded i n s i d e other r e l a t i v e c l a u s e s . In t h i s l i g h t , consider the f o l l o w i n g sentence: 93 2.48 The woman who entered the press club which was named a f t e r the man who founded the paper which s e l l s more copies than any other paper i n Vancouver i s the author of s e v e r a l books. Using the technique of r e l a t i v e clause e l i m i n a t i o n developed f o r r e l a t i v e clauses as they occur i n independent clauses, we can unpack a sentence with m u l t i p l y embedded r e l a t i v e clauses s t a r t i n g from the inmost c l a u s e . 2.49 Some paper s e l l s more copies than any other paper i n Vancouver. Some man founded that paper. Some press c l u b was named a f t e r that man. Some woman entered that press club. That woman i s the author of s e v e r a l books. Transforming from a c t i v e to passive and passive to a c t i v e wherever p o s s i b l e , i t might a l s o be observed, helps to keep the references smooth, thus consider: 2.1 The dog which ate the cat which chased the mouse which n i b b l e d the cheese i s here. Using the a c t i v e v o i c e , t h i s could be paraphrased by: 2.51 Some mouse n i b b l e d the cheese. Some cat chased that mouse. Some dog ate that c a t . That dog i s here And u s i n g the pas s i v e v o i c e f o r more f l u i d reference, i t would be paraphrased by: 2.52 The cheese was nib b l e d by some mouse. That mouse was chased by some cat. That cat was eaten by some dog. That dog i s here. However I do not wish to emphasise t h i s p o i n t about the transformation f o r the sake of f l u i d i t y of a c t i v e to passive and passive to a c t i v e . I t i s not always p o s s i b l e i n the f i r s t p lace and i t i s not necessary i n the second. I merely want to c l a i m 94 that r e l a t i v e clauses, both appo el i m i n a b l e whether they occur i n conjoined or embedded. We do no order to express our thoughts . s i t i v e and r e s t r i c t i v e , are independent clauses or m u l t i p l y t need r e l a t i v e clauses i n C. Noun and Verb Phrase Complements. Having attempted to demonstrate that a language might be f u l l y expressive although i t contains no r e l a t i v e c lauses, I turn now to an examination of two other types of dependent clause c o n s t r u c t i o n s , noun and verb phrase complements. B a s i c a l l y speaking, we are i n t e r e s t e d i n the f o l l o w i n g kinds of sentences: 3.1 John expected that Jean w i l l come back. 3.2 John expected Jean to come back. 3.3 John persuaded Jean to come back. 3.4 John expected Jean to see B i l l . 3.5 John persuaded Jean to see B i l l . 3.6 John's meeting Jean was c o i n c i d e n t a l . Each of these sentences contains a dependent clause. The surface suggests that there are three kinds of clauses l i s t e d here: those which begin with 'that' ( l i k e 3.1), those that have an i n f i n i t i v a l c o n s t r u c t i o n , ( l i k e 3.2, 3.3, 3.4, and 3.5), and those that have a gerund preceded by a possesive, ( l i k e 3.6). Pre-transformational grammar c l a s s i f i e d these sentences together as containing 'noun clauses', and i t was s a i d , at l e a s t i n 95 some pr e - t r a n s f o r m a t i o n a l t e x t s , that there were three kinds of noun c l a u s e s r u s i n g modern names 'that + sentence' noun clauses; i n f i n i t i v a l , or ' f o r . . . t o . . . ' noun clauses; and gerundial, or 'possesive + .. . . i n g ' noun cl a u s e s . In terms of surface s t r u c t u r e , there i s some j u s t i f i c a t i o n f o r t h i n k i n g of a l l three kinds of clauses as being noun cla u s e s . I f we examine 3 . 1 , 3.2, and 3.3, we w i l l see that 3 . 1 i s roughly synonymous with 3.2, and that 3.2 i s i d e n t i c a l to 3.3 except f o r the d i f f e r e n t verb. There was no reason to doubt, nor i s there now any reason to doubt that 3 . 1 contains a noun c l a u s e . Since 3.2 contains an i n f i n i t i v a l clause that can express the same thought as the 'that + sentence' clause. and s i n c e 3.2 and 3.3 both contain s i m i l a r i n f i n i t i v a l clauses on the surface, the conclusion that both 3.2 and 3.3 contain i n f i n i t i v a l noun clauses was a n a t u r a l one. I t was concluded that i n general i n f i n i t i v a l clauses were noun clauses. Transformational evidence, however, suggests that whereas some i n f i n i t i v a l clauses are noun clauses, others are not. Taking 3.2 and 3.3 as examples, we can see that 3.2 takes a transformation which 3.3 r e j e c t s : 3.7 What John expected was f o r Jean to come back. 3.8* What John persuaded was for Jean to come back. Now t h i s transformation i s c a l l e d the ' c l e f t sentence' transformation, and i n other ways has been e s t a b l i s h e d as a t e s t f o r the presence of a noun phrase (see J and R, E n g l i s h Transformational Grammar, page 39). This suggests that some 96 i n f i n i t i v a l clauses might not be dominated by 'noun phrase' i n deep s t r u c t u r e , and thus would not be 'noun clauses' or 'noun phrase complements'. There i s , as w e l l , other evidence that clauses f o l l o w i n g 'persuade' do not have the same deep s t r u c t u r e a n a l y s i s as clauses f o l l o w i n g 'expect'. Consider, f o r example, 3.4 i n comparison with 3.5. 3.4 John expected Jean to see B i l l . 3.5 John persuaded Jean to see B i l l . I f we p a s s i v i z e the clauses, we have the two sentences: 3.9 John expected B i l l to be seen by Jean. 3.10 John persuaded B i l l to be seen by Jean. Although 3.4 and 3.9 might be used i n d i f f e r e n t s i t u a t i o n s , they are roughly synonymous; but 3.5 and 3.10 are not at a l l synonymous. In 3.5 'Jean' i s the object of the verb 'persuade', whereas i n 3.10 ' B i l l ' i s the object of the verb'persuade'. In both 3.4 and 3.9 the whole clause i s the object of the verb 'expect'. And these d i f f e r e n c e s i n grammatical r e l a t i o n s must be accounted f o r i n the deep s t r u c t u r e a n a l y s i s . One way to account f o r these d i f f e r e n c e s i s to post u l a t e that some i n f i n i t i v a l clauses are immediately dominated not by 'noun phrase' but by 'verb phrase' i n deep s t r u c t u r e . The deep s t r u c t u r e f o r : 3.2 John expected Jean to come back. 97 might be: Sentence Noun phrase Verb phrase Verbal Noun phrase I sentence John. expected Jean to come back whereas the deep s t r u c t u r e f o r : 3.3 John persuaded Jean to come back, would be: Sentence Noun phrase Verb phrase Verbal Noun phrase Noun / Jean Sentence John persuaded And the deep s t r u c t u r e f o r : 3.4 John expected Jean to see B i l l would be something l i k e : Sent e tnr^e^^^ Noun phrase Verb phrase . Verbal Noun phrase I s e n t e n c e i - -John expected Jean to see B i l l whereas the deep s t r u c t u r e f o r : 3.5 John persuaded Jean to see B i l l . would be something l i k e : Jean to come back 98 Sentence / \ Noun phrase Verb phrase Verbal Noun pharase Sentence I Noun I John persuaded Jean Jean to see B i l l This would mean that noun clauses, or noun phrase complements are generated by a rewrite r u l e l i k e : 3.11 Noun phrase — ——> Sentence whereas verb phrase clauses, or verb phrase complements are generated by a r e w r i t e r u l e l i k e : 3.12 Verb phrase > Verbal + Noun phrase + Sentence To summarize what i s relevant f o r our purposes, t h i s treatment analyzes 'that + sentence' clauses and 'possesive +...ing clauses i n general as houn phrase complements, and i n f i n i t i v a l , or ' f o r . . . t o . . . ' clauses as noun phrase complements when the c l e f t transformation works and when, i f the verb i s t r a n s i t i v e , p a s s i v i z i n g the clause preserves synonymy, but as verb phrase complements when the c l e f t sentence transformation does not work,and when, i f the verb i s t r a n s i t i v e , p a s s i v i z i n g the clause does not preserve synonymy. This a n a l y s i s i s s t i l l a contro-v e r s i a l one, but, as the controversy i s not over whether or not the clauses i n sentences 3.1 through to 3.6 are r e c u r s i v e l y derived s t r u c t u r e s , since i t i s c l e a r that they are, the work which follows does not hinge on the adequacy of the transforma-t i o n a l a n a l y s i s . I t i s the task of t h i s study to show that r e c u r s i v e l y derived s t r u c t u r e s are eliminable i n paraphrase, 99 however they are analyzed. On the other hand, the transforma-t i o n a l a n a l y s i s i s an a i d i n the sense that the b e t t e r we can c l a s s i f y the sentences, the b e t t e r we can examine the d i f f e -rent types . Before c o n t i n u i n g d i r e c t l y to the question of the e l i m i n a b i l i t y of these clauses i n paraphrase, I would l i k e to make note o f an i n t e r e s t i n g and important d i s t i n c t i o n , a ' d i s t i n c t i o n which i s u s u a l l y t r e a t e d as a d i s t i n c t i o n i n verb types. (Later, however, i n the d i s c u s s i o n of 'how' and 'why' complements, we come against some evidence that t h i s i s not only a d i s t i n c t i o n i n verb types.) The d i s t i n c t i o n i s between f a c t i v e and non-factive verbs. When f a c t i v e verbs are used with or without a negative, then, i f the complementing clause i s expressed i n p r o p o s i t i o n a l form, the t r u t h of the p r o p o s i t i o n i s presupposed. For example, i n a sentence l i k e : 3.13 John knew that Mary had gone. i f John knew that Mary had gone, then the p r o p o s i t i o n that Mary had gone, or the clause expressed i n p r o p o s i t i o n a l form, 'Mary had gone' i s presupposed. To give another example, i n : 3.14 John knew Mary to have gone. i f John knew M a r y t o have gone, then the complementing clause expressed i n p r o p o s i t i o n a l form, 'Mary had gone' i s presupposed to be t r u e . On the other hand, i t does not follow from, or i s not presupposed by: L O O 3.15 John b e l i e v e d Mary to have gone, that Mary had gone. Again: 3.16 John expects that Mary w i l l go. does not imply or presuppose that Mary w i l l go, whereas: 3.17 John knows that Mary w i l l go. does imply or presuppose that Mary w i l l go. The same d i s t i n c t i o n appears when the negative i s used with the verb. Thus: 3.17a John d i d n ' t know that Mary had gone, presupposes that Mary went, while: 3.17b John d i d n ' t b e l i e v e that Mary had gone, does not c a r r y t h i s p r e s u p p o s i t i o n that Mary had gone. A l s o : 3.17c John didn' t expect that Mary would go. c a r r i e s no p r e s u p p o s i t i o n that M a r y went, i s going, or w i l l go. Further, 'Yes/No' questions ( i . e . , questions which can be answered by "Yes" or "No") i n d i c a t e the same d i s t i n c t i o n with these verbs. Thus: 3.17d Did John know that Mary had gone? presupposes that Mary had gone, while: 3.17e Did John b e l i e v e that Mary had gone? does not presuppose that Mary had gone. 'Believe' and 'expect* are thus non-factive verbs, whereas 'know' i s a f a c t i v e verb, the c o n d i t i o n of the presupposition of a complementing clause when expressed as a p r o p o s i t i o n being both s u f f i c i e n t and necessary f o r a verb's being f a c t i v e . 1 0 1 Now the f a c t i v e / n o n - f a c t i v e d i s t i n c t i o n i n verbs l i k e • believe' 'know', 'expect', and ' r e a l i z e ' has been discussed; f o r a grammatical treatment for example, see Fact by P. and C. K i p a r s k i . But the f a c t i v e / n o n - f a c t i v e aspects of verbs l i k e 'persuade', 'convince', 'force', 'want', 'need', 'cause', 'condescend', and 'prevent' have not to my knowledge been tr e a t e d or analyzed. Yet the d i s t i n c t i o n , i t appears, a p p l i e s over these verbs as w e l l . The f a c t i v e / n o n - f a c t i v e d i s t i n c t i o n has been a p p l i e d to verbs which, l i k e 'know' 'expect' 'believe' and ' r e a l i z e ' , c h a r a c t e r i s t i c a l l y are followed by 'that + sentence' noun phrase complements; however, the d i s t i n c t i o n would a l s o appear to apply to verbs which c h a r a c t e r i s t i c a l l y are followed by i n f i n i t i v a l complements, i . e . to verbs which c h a r a c t e r i s t i c a l l y are followed by ' f o r . . . t o . . . ' noun or verb phrase complements. For example, i n a s s e r t i n g : 3.18 John forced Mary to go. one i s presupposing that Mary went, that Mary i s going, or that Mary w i l l go. S i m i l a r l y , i n a s s e r t i n g : 3.19 John caused Mary to go. one i s presupposing that Mary went (or i s going?) On the other hand, i n a s s e r t i n g : 3.20 John wanted Mary to go. one i s not presupposing that Mary went, i s going, or w i l l go. And s i m i l a r l y , i n a s s e r t i n g : 102 3.21 John persuaded Mary to go. one i s not presupposing that Mary went, i s going, or w i l l go, although there i s an i m p l i c a t i o n that Mary intended or i s intending to go. The i m p l i c a t i o n that she d i d go, i s going, or w i l l go f a i l s , of course because although a person may intend to do something, they may be prevented from doing i t . This b r i n g s us to 'prevent' which i s i n another c l a s s s t i l l . In a s s e r t i n g : 3.22 John prevented Mary from going. one i s presupposing that Mary d i d not go. We might say, then, that 'force' and 'cause' are f a c t i v e , that 'want' and 'persuade' are n o n - f a c t i v e , and that 'prevent' i s n e g a t i v e l y f a c t i v e , i n that i t presupposes that the complementing clause when expressed as a p r o p o s i t i o n i s f a l s e . Now there are many points here deserving of a more accurate a n a l y s i s . For example, there has been throughout t h i s d i s c u s s i o n an unwritten i n s t r u c t i o n to 'express the complemen-t i n g clause i n p r o p o s i t i o n a l form'; but t h i s i n s t r u c t i o n needs c l a r i f i c a t i o n . In 3.22, f o r instance, how does one derive the p r o p o s i t i o n 'Mary d i d not go' from the surface s t r i n g 'Mary from going'? Secondly there i s the not unrelated problem of tense. Some verbs seem to for c e the expression of a p r o p o s i t i o n i n the past tense, whereas others do not. This i s p u z z l i n g and requires an a n a l y s i s . T h i r d l y , there i s the r e l a t i o n s h i p between 'non-factive' and 'negatively f a c t i v e ' which could be 103 f u r t h e r explored and d e l i n e a t e d . Apparently 'negatively f a c t i v e ' i s a species of ' f a c t i v e ' with a p r o p o s i t i o n l i k e 'A prevented B from doing X' r e c e i v i n g an a n a l y s i s something along the l i n e s of 'A caused B to not do X'; but the exact s p e c i f i c a t i o n s of the mechanism 'negatively f a c t i v e ' i s u n c l e a r . And f i n a l l y there i s the question of the extent to which the ' f a c t i v e ' verbs which, l i k e 'force', take verb phrase complements pass the important t e s t of the negative and the question. As i t i s not always the case that: 3.22a John d i d n ' t f o r c e Mary to go. presupposes that Mary went, one might argue that f o r c e i s not f a c t i v e . But s i n c e the negative t e s t i s not i n f a l l i b l e even for a paradigm f a c t i v e verb l i k e 'know', only an extended i n v e s t i g a t i o n could judge on the appropriateness of applying the d i s t i n c t i o n to verbs which take verb phrase complements. Thus: 3.22b John could not have known that Mary went, does not n e c e s s a r i l y presuppose that Mary went, e s p e c i a l l y i f there i s a heavy s t r e s s on 'known'. Here, however I w i l l attempt to do no more than make note of the d i s t i n c t i o n , observing that i t a p p l i e s or may apply over a broad range of verbs, i n c l u d i n g verbs which c h a r a c t e r i s t i c a l l y take i n f i n i t i v a l clauses; i n demonstrating the e l i m i n a b i l i t y of c e r t a i n kinds of clauses l a t e r on the d i s t i n c t i o n or at l e a s t some v e r s i o n of i t w i l l reappear. 104 Noun Phrase Complements: So f a r , i n d i s t i n g u i s h i n g between noun and verb phrase complements mention has been made of three kinds of noun phrase complements: 'that + sentence' ' f o r . . . t o . . . ' and 'possessive + . ..ing' complements. I t was implied that a l l 'that + sentence'clauses were noun phrase complements, as w e l l as a l l 'possessive + ...ing' complements being noun phrase complements. However, those sentences i n which there i s a verb which i s c h a r a c t e r i s t i c a l l y followed by a noun phrase and by an i n f i n i t i v a l verb phrase complement, but which i s followed by a noun phrase and a 'that + sentence' clause v i o l a t e t h i s r u l e . For example, i n : 4.1 John persuaded Jack that the world i s f l a t , we have a verb, 'persuade' which c h a r a c t e r i s t i c a l l y takes a noun phrase and an i n f i n i t i v a l verb phrase complement, but i t i s followed i n t h i s case by the noun phrase 'Jack:' and by the 'that + sentence' clause 'that the world i s f l a t * . Is t h i s clause a noun or a verb phrase complement? I f we consider the grammatical r e l a t i o n s surrounding t h i s clause, we come up against the problem of ' J a c k ' s ' r e l a t i o n to 'the world i s f l a t ' . 'Jack' appears to be the subject and 'that the world i s f l a t ' , the object, of a missing verb. Considering the synonymy of 4.1 with: 4.2 John persuaded Jack to b e l i e v e that the world i s f l a t , i t would be reasonable to postulate a deep s t r u c t u r e for 4.1 105 which follows the verb phrase complement p a t t e r n : 4.3 Sentence Noun John persuaded J a c k Jack to b e l i e v e that the world i s f l a t I make note of t h i s c l a s s of exceptions without f u r t h e r comment, f e a r i n g to enter any f u r t h e r i n t o the t h i c k e t of d e s c r i p t i v e t r a n s f o r m a t i o n a l l i n g u i s t i c s l e s t we never leave i t . More important at t h i s p o i n t i s the question of whether noun phrase complements are only of the three types mentioned above. They are not. In a d d i t i o n to these three types — 'that sentence', ' f o r . . . t o . . . ' , and 'possessive + ...ing' c l a u s e s — there are those noun phrase clauses which, c l o s e l y resembling 'that + sentence' clauses, are introduced by 'whether', ' i f ' , 'why', 'how', and 'what'. Having mentioned eight types of noun phrase complements, I w i l l move on now to an a n a l y s i s of t h e i r e l i m i n a b i l i t y i n paraphrase. 'That + sentence' noun phrase complements. I t ' s c l e a r that we sometimes use sentences with noun phrase complements when we needn't do so. Consider f o r instance: 4.1 I t ' s f a l s e that Harry i s b a l d . 106 Here we have t h e noun p h r a s e complement " t h a t H a r r y i s b a l d " ; however, t h e t h o u g h t t h a t i t ' s f a l s e t h a t H a r r y i s b a l d can be e x p r e s s e d , and o f t e n w o u l d be e x p r e s s e d w i t h o u t t h e noun p h r a s e complement, as i n : 4.2 H a r r y i s n o t b a l d . W i t h o u t g o i n g i n t o t h e r e l a t i o n s h i p o f 4.1 t o 4.2 and t h e a d e q u a c y o f 4.2 as a p a r a p h r a s e f o r 4.1, i t c a n be o b s e r v e d t h a t , however good o r b a d i t i s as a p a r a p h r a s e , 4.2's a v a i l a b i l i t y i s d e p e n d e n t on t h e p r e s e n c e o f t h e v e r b a l ' t r u e ' o r t h e v e r b a l ' f a l s e ' . Thus: 4.3 I t ' s t r u e t h a t H a r r y i s b a l d , i s , a t l e a s t s o m e t i m e s , p a r a p h r a s e a b l e b y : 4.4 H a r r y i s b a l d . b u t : 4.5 I t ' s a good t h i n g t h a t H a r r y i s b a l d . i s n e v e r p a r a p h r a s e a b l e b y e i t h e r 4.2 o f 4.4. Thus, n e i t h e r p r o v i d e s us w i t h a g e n e r a l s t y l e o f p a r a p h r a s e w h i c h w i l l w o rk o v e r a b r o a d r a n g e o f c a s e s . L e t ' s c o n s i d e r t h e more r e p r e -s e n t a t i v e sentence.: 4.6 Mary b e l i e v e s t h a t H a r r y i s b a l d . I n t h e m a t e r i a l i m m e d i a t e l y f o l l o w i n g I w i l l e n t e r t a i n a number o f p a r a p h r a s e a p p r o a c h e s t o t h i s k i n d o f s e n t e n c e , f i n a l l y f o c u s i n g on t h e one w h i c h seems t o me t o o f f e r t h e most p r e c i s i o n and t o have a p p l i c a t i o n o v e r the b r o a d e s t r a n g e o f c a s e s . 107 Let's begin with the f o l l o w i n g observation: One of the most economical and pervasive devices of language i s the use of s u b s t a n t i v a l expressions to abbre-v i a t e , summarize, and connect. Having made a s e r i e s of d e s c r i p t i v e statements, I can comprehensively connect with these the remainder cf my discourse by the use of such expressions as ' t h i s s i t u a t i o n ' or 'this s t a t e of a f f a i r s ' ; j u s t as, having produced what I regard as a set of reasons for a c e r t a i n c o n c l u s i o n I allow myself to draw breath by saying 'Since these things are so, then...', i n s t e a d of p r e f a c i n g the e n t i r e s t o r y by the conjunction. A s i t u a t i o n or s t a t e of a f f a i r s i s , roughly, a set This i s a casual remark of Strawson's appearing i n the context of a d i s c u s s i o n of the r e l a t i o n s h i p between f a c t s , s i t u a t i o n s and states of a f f a i r s (see L o g i c o - L i n q u i s t i c Papers, page 197) . I t i s i n t e r e s t i n g i n that i t i m p l i c i t l y r e f e r s to paraphrase on a s y n t a c t i c v a r i a b l e which i s the subject of t h i s chapter, and a l s o contains the b a s i c p r i n c i p l e of the s t y l e s of noun clause paraphrase which I w i l l now consider. Further, i n On Saying That ( i n Words and Objections, ed. Harman and Davidson) Davidson addresses h i m s e l f d i r e c t l y to t h i s subject and o f f e r s a paraphrase f o r sentences l i k e 4.6 along the f o l l o w i n g l i n e s : 4.7 Harry i s b a l d . Mary b e l i e v e s t h i s . Davidson, however, does not discuss the adequacy of t h i s s t y l e of paraphrase as f u l l y as we s h a l l want to i n t h i s study. Let us b e g i n : t h e n by asking: i s 4.7 an adequate paraphrase of 4.6? One problem here i s the status of the statement "Harry i s b a l d " . That Mary b e l i e v e s that Harry i s b a l d does not e n t a i l or presuppose that Harry i s b a l d . The verb 'believe' i s non-f a c t i v e , and so separating the two clauses i n t o independent 108 sentences leads us i n t o expressing a statement which the o r i g i n a l does not e n t a i l or presuppose. The s i t u a t i o n i s some-what better, though,using a f a c t i v e verb l i k e 'know'. 4.8 Mary knows that Harry i s b a l d . Now when we paraphrase with: 4.9 Harry i s b a l d . Mary knows t h i s . we are somewhat c l o s e r to an adequate paraphrase. How, then, can we handle the sentences with non-factive verbs? What we need, e s s e n t i a l l y , i s some kind of device f o r u t t e r i n g a sentence without a s s e r t i n g i t . The use of questions i s i n t e r e s t i n g i n t h i s regard. Thus, to paraphrase: 4.12 Mary does not b e l i e v e that Harry i s b a l d , we could use: 4.13 Is Harry bald? Mary does not b e l i e v e i t . The question used i n t h i s way can be h e l p f u l although we wouldn't want to place e x c l u s i v e r e l i a n c e on i t . We sometimes c a l l questions of t h i s type ' r h e t o r i c a l ' to d i s t i n g u i s h them from the questions we ask when we don't know the answers, or when we want others to f u r n i s h the answers. But we would want to d i s t i n g u i s h two kinds of r h e t o r i c a l questions: questions which are d r a m a t i c a l l y r h e t o r i c a l and used f o r e f f e c t s such as irony and emphasis, and on the other hand, questions which we might dub ' s y n t a c t i c a l l y r h e t o r i c a l ' and are used or can be used to introduce p r o p o s i t i o n s i n t o discourse without g i v i n g them 109 a s s e r t o r i a l f o r c e . Also, s i n c e we sometimes ask questions which themselves co n t a i n noun clauses, we would sometimes end up with s t r i n g s of d i f f e r e n t kinds of questions and a convention would have to be introduced i n t o language to d i s t i n g u i s h the s y n t a c t i c a l l y r h e t o r i c a l questions from the other kinds of questions. Thus to paraphrase the question: 4.14 D i d Mary b e l i e v e that Harry i s bald? we 1d have: 4.15 Is Harry bald? Did Mary b e l i e v e i t ? And some convention would be h e l p f u l , perhaps necessary, to i n d i c a t e that the speaker i s not asking "Is Harry b a l d ? " i n order to r e c e i v e an answer. We might a l s o want to e n t e r t a i n use o f suppose', as i n 'suppose Harry i s b a l d ' ; t h i s introduces some problems, however, i n s o f a r as 'suppose Harry i s b a l d ' i s a transformation of 'suppose that Harry i s b a l d ' ; the approach nevertheless could be j u s t i f i e d , I think, but i t s j u s t i f i c a t i o n would be somewhat complicated at t h i s p o i n t . In any case, there i s another device i n the language f o r s t r i p p i n g a p r o p o s i t i o n of i t s a s s e r t o r i a l f o r c e , and perhaps a b e t t e r one, and that i s the use of quotation marks. Thus, we can paraphrase: 4.16 Mary b e l i e v e s that Harry i s b a l d . with: 4.17 Mary b e l i e v e s the f o l l o w i n g p r o p o s i t i o n . "Harry i s ba 110 Further, u s i n g the quotation marks around the p r o p o s i t i o n we can a l s o give the p r o p o s i t i o n a proper name and then use the proper name as the subject of the verb i n the main clause of the o r i g i n a l : 4.18 "Harry i s b a l d . " C a l l the preceding p r o p o s i t i o n 'A'. Mary b e l i e v e s A. Of the s t y l e s of noun clause paraphrase we've so f a r considered, t h i s l a s t s t y l e with the proper name gives the most p r e c i s i o n and i s the e a s i e s t to apply s y s t e m a t i c a l l y and a c c u r a t e l y over a broad range of cases. Let's look, f o r example, at how a s e r i e s of m u l t i p l y embedded 'that + sentence' noun clauses would be handled using t h i s s t y l e of paraphrase: 4.19 John b e l i e v e s that Mary denied that Edna affirm e d that P h i l l i p knows that Harry i s b a l d . 4.20 C a l l the f o l l o w i n g p r o p o s i t i o n 'A'. "Harry i s b a l d . " C a l l the f o l l o w i n g p r o p o s i t i o n , 'B'. " P h i l l i p knows A C a l l the f o l l o w i n g p r o p o s i t i o n , ' C . "Edna affirm e d B C a l l the f o l l o w i n g p r o p o s i t i o n , 'D'. "Mary denied C." John b e l i e v e s D. Now i t i s not necessary to use the proper names? 4.19 could a l s o be paraphrased without g i v i n g the pr o p o s i t i o n s proper names, e.g., i n the f o l l o w i n g way: 4.21 "Harry i s b a l d . " " P h i l l i p knows the preceding proposition." "Edna affirm e d the preceding proposition." "Mary denied the preceding p r o p o s i t i o n . " John b e l i e v e s the preceding p r o p o s i t i o n . However the proper name s t y l e i s somewhat more p r e c i s e and perhaps e a s i e r to follow as w e l l and so I w i l l tend to focus on i t i n the f o l l o w i n g account. I l l I claim, then, that the proper name s t y l e of paraphrase has broad and systematic a p p l i c a b i l i t y over a large range — indeed, over a l l -- noun phrase complements. Let's look at some of the problems which a r i s e i n t r y i n g to make i t work. Perhaps the most general c r i t i c i s m of t h i s kind of para-phrase which might be presented i s the f o l l o w i n g : We can know that a p r o p o s i t i o n i s true, i t might be s a i d , but we cannot 'know a p r o p o s i t i o n ' . I f the phrase 'know a p r o p o s i t i o n ' i s to have any meaning then i t can only mean 'know a p r o p o s i t i o n to be t r u e . ' Thus i f I say "Mary knew the p r o p o s i t i o n " i t can only mean, "Mary knew the p r o p o s i t i o n to be t r u e . " S i m i l a r l y , the argument would go, i f I say "Mary b e l i e v e d the p r o p o s i t i o n " i t can only mean "Mary b e l i e v e d the p r o p o s i t i o n to be t r u e " . I f so then the paraphrase v e r s i o n i s short f o r , or is t r a n s f o r m a t i o n a l l y dependent on, a sentence which contains an embedded c l a u s e . There are three avenues which can be followed i n response to t h i s o b j e c t i o n to the proper name s t y l e of paraphrase. The f i r s t response, and perhaps the weakest, i s to deny that to 'know a p r o p o s i t i o n ' or to 'know A' where 'A* i s the name of the p r o p o s i t i o n , i s short f o r , or t r a n s f o r m a t i o n a l l y dependent upon, 'to know a p r o p o s i t i o n to be true' or to 'know A to be tr u e ' . This response i s best supported by c o n s i d e r a t i o n of constructions i n which the verb takes a pronoun as obje c t . Consider the fo l l o w i n g sentences: 112 4.22 Mary knew i t at the time. 4.23 Mary b e l i e v e d i t at the time. In both 4.22 and 4.23 the pronoun ' i t ' r e f e r s to some p r o p o s i t i o n , and i s the d i r e c t o b j e c t of the verbs 'know' and'believe'. I f we can have pronouns as d i r e c t objects of these verbs, then there would be no reason why we cannot s u b s t i t u t e proper names for these pronouns. The question then turns on whether or not a sentence l i k e , "Mary knew i t at the time," i s short f o r , or i s t r a n s f o r m a t i o n a l l y dependent upon a sentence l i k e : 4.24 Mary knew i t to be true at the time. So an argument here would shape up i n t h i s way: i n one camp there would be those who'd argue that 4.22 contains a d e l e t i o n , and i n the other camp, those who'd argue that 4.24 contains a redundancy. Since t h i s argument would be a very tangled one to pursue, i t i s fortunate that, although one may be unconvinced that 4.22 contains a d e l e t i o n , and one may suspect rather t h a t 4.24 contains a redundancy, one need not r e l y on t h i s s u s p i c i o n . This response i s the weakest not on i t s merits, then, but because i t leads i n t o a tangled argument which, for the time being, we might do j u s t as w e l l to leave unconsidered. Let's look at the other responses. The second response focuses on the status of the clause ' i t to be t r u e ' . Let us assume, f o r the sake of argument, that 4.22, "Mary knew i t at the time" i s short for or t r a n s f o r m a t i o n a l l y dependent upon 4.24, "Mary knew i t tobe true at the time." 113 Let us a l s o assume f o r the sake of argument, that: 4.25 Mary knew A. i s short f o r or t r a n s f o r m a t i o n a l l y dependent upon: 4.26 Mary knew A to be tru e . We are assuming then, f o r the sake of argument, that the proper name s t y l e of paraphrase i s s t i l l dependent on s e n t e n t i a l embedding, s i n c e ' i t to be true', or 'A to be true' i s an embedded cl a u s e . The second response focuses on the status o f these clauses and asks whether or not these are r e c u r s i v e l y derived s t r u c t u r e s , or are n e c e s s a r i l y r e c u r s i v e l y derived s t r u c t u r e s . Notice that there i s a c r u c i a l d i f f e r e n c e between a noun clause l i k e 'that Harry agrees' and a noun clause l i k e ' i t to be true' or 'that i t i s true' or 'A to be t r u e ' . 'That Harry agrees' i s a noun clause which contains a verb which can i t s e l f be followed by another noun clause, e.g. 'That Harry agrees that Mary knows that John came back'. On the other hand ' i t to be true' or 'that i t i s true' or 'A to be true' does not contain a verb which permits f u r t h e r embedding of noun cla u s e s . The copula does permit f u r t h e r embedding of r e l a t i v e clauses, but these, as we have seen, are eli m i n a b l e i n paraphrase: 4.27 Mary knew A to be a p r o p o s i t i o n which had some appeal. 4.28 Some prop o s i t i o n s have some appeal. Mary knew A to be that kind of p r o p o s i t i o n . Furthermore, i t can be seen that i n d i s e n t a n g l i n g m u l t i p l y 114 embedded noun clauses, and t h i s point i s the c r u c i a l one, even admitting the dependency of 'knew A ' on 'knew A to be true' we would never need more than one embedded 'to be true' clause per sentence. For example to disen t a n g l e : 4.29 Mary b e l i e v e s that John knows that Edna knows that Tom i s a l i a r . we would have: 4.3 0 "Tom i s a l i a r . " C a l l the preceding p r o p o s i t i o n 'A'. "Edna knows A to be t r u e . " C a l l the preceding p r o p o s i t i o n , 'B'. "John knows B to be t r u e . " C a l l the preceding p r o p o s i t i o n , 'C'. Mary b e l i e v e s C to be true. And i n 4 . 3 0 there i s never more than one embedded 'to be tr u e ' clause per sentence. Now i f an embedded clause i s not to occur more than once per sentence we can construct base r u l e s which generate such clauses and which b l o c k them from occuring more than once. Base r u l e s which generate noun clauses such that they may occur only once are not r e c u r s i v e r u l e s , and the 'to be true' clauses so generated are not r e c u r s i v e l y derived s t r u c t u r e s . To see i n concrete terms how such r u l e s could be constructed, consider the f o l l o w i n g s et of base r u l e s : 4 . 3 1 S > Noun phrase + Verb phrase Verb phrase > Verb + ST_ SJ_ > Noun phrase^ +" Verb phrase-j_. The s u b s c r i p t s i n these r u l e s prevent the rul e s from being a p p l i e d r e c u r s i v e l y . Only one clause could appear i n any sentence 115 generated by these r u l e s ; and the r u l e s are not r e c u r s i v e s i n c e no category dominates i t s e l f , 'S* being a d i s t i n c t category from 'S^1 and so on. Thus, the second response to the c r i t i c i s m that phrases l i k e 'know a p r o p o s i t i o n ' are short f o r , or t r a n s f o r m a t i o n a l l y dependent upon phrases l i k e 'know a p r o p o s i t i o n to be true" i s the demonstration that m u l t i p l y embedded 'that + S' noun phrase complements can be eliminated i n paraphrase using a maximum of one embedded clause per sentence. This p a r t i c u l a r strategem i s an important one, and s h a l l be used l a t e r on a number of occasions. In f u t u r e I w i l l c a l l a set of base r u l e s l i k e 4.31 a set of base r u l e s which 'permit non-recursive s e n t e n t i a l embedding'. The t h i r d response focuses on the verb. Again we may grant, for the sake of argument that 'Mary knew p r o p o s i t i o n A' or 'Mary knew a p r o p o s i t i o n ' or 'Mary knew A', where 'A' i s the name of a p r o p o s i t i o n , i s short f o r , or t r a n s f o r m a t i o n a l l y dependent upon 'Mary knew p r o p o s i t i o n A to be true' or 'Mary knew a p r o p o s i t i o n to be true' or 'Mary knew A to be true', where A i s the name of a p r o p o s i t i o n , r e s p e c t i v e l y . But now, l e t us examine the r e l a t i o n s h i p s between the f o l l o w i n g l o c u t i o n s : 4.3 2 Mary had an i n s i g h t . 4.33 Mary grasped a concept. 4.34 Mary asked a question. 116 4.35 Mary presented an argument. 4.36 Mary argued a p r o p o s i t i o n . 4.37 Mary asserted a p r o p o s i t i o n . 4.38 Mary u t t e r e d a p r o p o s i t i o n . 4.39 Mary b e l i e v e d a p r o p o s i t i o n . 4.40 Mary knew a p r o p o s i t i o n . I t h i n k i t w i l l be agreed that i n the f i r s t four cases, there i s no question of dependency on some deleted clause. 'To have an i n s i g h t ' f o r example, i s not dependent on 'to have an i n s i g h t to be true' or some such c o n s t r u c t i o n ; 'to grasp a concept' s i m i l a r l y i s not dependent on 'to grasp a concept to (be true) (have a p p l i c a t i o n ) ' or any such c o n s t r u c t i o n , and so on f o r 'to ask a question' and 'to present an argument'. Now of the next three, I thin k i t w i l l a l s o be agreed that 4.38, i n i n f i n i t i v a l form, 'to u t t e r a p r o p o s i t i o n ' i s not dependent on'to u t t e r , a p r o p o s i t i o n to be true (to be heard) 1 or any such c o n s t r u c t i o n , and s i m i l a r l y , although t h i s i s j u s t a b i t l e s s c l e a r , 'to a s s e r t a p r o p o s i t i o n ' does not seem to be dependent upon 'to a s s e r t a p r o p o s i t i o n to be true', while the status of 'to argue a p r o p o s i t i o n ' i s perhaps close to that of 'to b e l i e v e a p r o p o s i t i o n ' and 'to know a p r o p o s i t i o n ' . However, grant i n g that 'know' and 'believe' are d i f f e r e n t from 'ask' and 'grasp' and 'utter', what i s to prevent us from c o n s t r u c t i n g new verbs, say 'known ' and 'believei'based on the 117 c o n s t r u c t i o n s p o s s i b l e with 'ask' and 'grasp'? i f someone were to object on the grounds that there are some deep ( p o s s i b l y obscure) l i n g u i s t i c - p h i l o s o p h i c a l processes at work which enable us to speak of 'asking questions' without having to depend on clauses when we can't speak of 'knowing a p r o p o s i t i o n ' without depending on an embedded clause, then that would be pushing an a l r e a d y tenuous argument somewhat too f a r . Consider what i t i s to 'have an i n s i g h t ' . To have an i n s i g h t i n many i n t e r e s t i n g ways resembles what i t i s to know a p r o p o s i t i o n , or to know a p r o p o s i t i o n to be t r u e . For someone to know a p r o p o s i t i o n , the p r o p o s i t i o n must be true; i n order to have an i n s i g h t , the i n s i g h t must be v a l i d or must apply. Thus, "Jones had the i n s i g h t that Hamlet was unconsciously jealous of Claudius" implies that Hamlet was unconsciously jealous of Claudius, j u s t as " E i n s t e i n knew that Newton's mechanics was inadequate" implies that Newton's mechanics was inadequate. Or, f o r example, i n the same way that we cannot say "I knew that de Gaulle was crowned King of France, but l a t e r I found out that I was wrong" but must rather say something l i k e , "I thought I knew (or I believed) that de Gaulle was crowned King of France, and l a t e r I found out that I was wrong", we cannot say, "I had an i n s i g h t i n t o Hamlet's r e l a t i o n s h i p with C o r d e l i a , but l a t e r I found out that C o r d e l i a i s a character from Lear" but must rather say something l i k e , "I thought I had an i n s i g h t i n t o Hamlet's r e l a t i o n s h i p with C o r d e l i a but l a t e r 118 I found out that C o r d e l i a i s a character from Lear" or something l i k e t h a t . Grasping a concept a l s o resembles knowing a p r o p o s i t i o n i n t h i s way. J u s t as we could not say "Tom knew that 2 + 2 = 5 , and l a t e r I taught him that 2 + 2 = 4 " ; but would have to say something l i k e , "Tom thought he knew (Tom believed) that 2 + 2 = 5 , and l a t e r I taught him that 2 + 2 = 4," we could not say "Tom grasped the lesson about the sum of 2 + 2, and l a t e r I discovered that he hadn't grasped i t " b u t would have to say something l i k e "Tom thought he grasped the lesson (or I thought he grasped the lesson) but l a t e r i t turned out that he hadn't"-But n e i t h e r 'to grasp a concept' nor 'to have an i n s i g h t ' are dependent on 'to grasp a concept to be true' or 'to have an i n s i g h t to be true' or any such c o n s t r u c t i o n , r e s p e c t i v e l y ; and there i s no reason why we could not have, assuming 'to know a p r o p o s i t i o n ' i s dependent on 'to know a p r o p o s i t i o n to be true', a verb, know^ f o r which 'to know-j_ a p r o p o s i t i o n ' has the meaning of 'to know a p r o p o s i t i o n to be true' without i t s being dependent on the l a t t e r c o n s t r u c t i o n . I b e l i e v e that t h i s response a l s o serves to put the issue i n t o a proper perspective, s i n c e we are not e x c l u s i v e l y i n t e r e s t e d i n the dependencies which e x i s t i n a language, but more broadly, i n those dependencies which must e x i s t i n any language f o r the language to be f u l l y expressive. And s u r e l y a a language could e x i s t i n which 'to know^ ^ a p r o p o s i t i o n ' has 1 1 9 the meaning of 'to know a p r o p o s i t i o n to be true' and i n which the former c o n s t r u c t i o n i s not short f o r or t r a n s f o r m a t i o n a l l y dependent upon the l a t t e r , j u s t as i n E n g l i s h , 'to grasp a concept' (or 'understand a pr o p o s i t i o n " ) or 'have an i n s i g h t ' i s not short f o r or t r a n s f o r m a t i o n a l l y dependent upon a c o n s t r u c t i o n which i t s e l f contains an embedded clause. Let's r e t u r n to have another look at what the proper name s t y l e of paraphrase i n v o l v e s . For examples 4.8 Mary knows that Harry i s b a l d . i s paraphrased under the proper name s t y l e of paraphrase by: 4.41a C a l l the f o l l o w i n g p r o p o s i t i o n 'A'. "Harry i s b a l d . Mary knows A. or: 4.41b C a l l the f o l l o w i n g p r o p o s i t i o n 'A1. "Harry i s b a l d . " Mary knows A to be t r u e . or: 4.41c C a l l the f o l l o w i n g p r o p o s i t i o n 'A'. "Harry i s b a l d . " Mary knowis A. We have three options. For someone who b e l i e v e s 'to know a p r o p o s i t i o n ' i s not short f o r or t r a n s f o r m a t i o n a l l y dependent on 'to know a p r o p o s i t i o n to be true', paraphrase a w i l l s u f f i c e . For someone who b e l i e v e s "to know a p r o p o s i t i o n * i s dependent upon 'to know a p r o p o s i t i o n to be true', there i s para phrase b i n which we use a maximum of one embedded clause per sentence, and £ i n which a l l s e n t e n t i a l embedding i s eliminated, and we use a new verb which can take 'proposition' as a d i r e c t object without dependence on a clause such as 'proposition to 120 be t r u e ' . Whatever one b e l i e v e s about the status of 'know a p r o p o s i t i o n ' , 'that + S' clauses can be eliminated using some v a r i a n t of the proper name s t y l e of paraphrase. I am claiming, then, that the proper name s t y l e of paraphrase works i n p r i n c i p l e f o r 'that + sentence* noun phrase complements. However, i f we consider verbs other than the ones we have so far been c o n s i d e r i n g , we w i l l discover that t h i s s t y l e of paraphrase has been incompletely c h a r a c t e r i z e d . Consider, f o r example, the verb 'hope', which a l s o takes 'that + sentence' noun phrase complements: 4.42 Mary hopes that Harry w i l l be b a l d . I f we f o l l o w the model we have been p r e v i o u s l y using, we w i l l paraphrase with: 4.43 C a l l the f o l l o w i n g p r o p o s i t i o n 'A*. "Harry w i l l be b a l d " . Mary hopes A. Now the phrase "Mary hopes A" sounds p e c u l i a r . The sentence i s not helped much i f we i n s e r t the word 'proposition', y i e l d i n g , "Mary hopes p r o p o s i t i o n A". And going back one step f u r t h e r , r e f l e c t i n g t h a t i n pa r t the use of proper names may be j u s t i f i e d i n terms of t h e i r being s u b s t i t u t i o n s on pronouns, we w i l l a l s o consider the c o n s t r u c t i o n , "Mary hopes i t " , which, u n s u r p r i s i n g l y , sounds no l e s s p e c u l i a r . I f we want "to use 'hope' without having a clause as d i r e c t object, then perhaps we would say "Mary hopes so". But ' i t ' and 'so* do not play p a r a l l e l r o l e s i n transforma-t i o n a l processes. 121 Consider f u r t h e r the c o n t r a s t between the verbs 'hope' and 'understand' when we t r y to attach the noun phrase 'that propo-s i t i o n ' as d i r e c t o b j e c t . In one case we have the i n t e l l i g i b l e sentence "Mary understands that p r o p o s i t i o n " , and i n the other, the u n i n t e l l i g i b l e sentence, "Mary hopes that p r o p o s i t i o n We can speak of 'understanding a p r o p o s i t i o n ' , but we cannot speak of 'hoping a p r o p o s i t i o n ' . Moreover, many verbs which take 'that + sentence' noun phrase complements s i m i l a r l y r e j e c t the term ' p r o p o s i t i o n ' as object; that i s , the verb 'hope' i s not merely an i n t e r e s t i n g exception. To f u r t h e r define the problem, even i f a verb can be followed by the term 'pr o p o s i t i o n ' without v i o l a t i n g the r u l e s of grammar, the term 'pr o p o s i t i o n ' w i l l not always be the appropriate term to f o l l o w the verb i n a given paraphrase s i t u a t i o n . For example, the verb 'mention' does not grammatically r e j e the term ' p r o p o s i t i o n ' as o b j e c t . We can speak of 'mentioning a p r o p o s i t i o n ' . But i f Jack mentions that John i s back, can we say that Jack i s mentioning a p r o p o s i t i o n , or that Jack i s mentioning such-and-such a p r o p o s i t i o n to be true? Or i f Jack has n o t i c e d that the l i g h t has turned green, has he n o t i c e d a p r o p o s i t i o n , or that a p r o p o s i t i o n is* true? So the problem i s not merely i n the grammatical r e l a t i o n -ship between the verb and the term 'proposition', but a l s o i n th< 122 appropriateness of the term 'pr o p o s i t i o n ' as a c h a r a c t e r i z e r of the content of the'that + sentence' complement i n question. How can we deal with t h i s problem? In the cases so f a r r a i s e d there are terms other than 'p r o p o s i t i o n ' which a p p r o p r i a t e l y c h a r a c t e r i z e the contents of the c l a u s e s . Thus i f Mary hopes that Harry w i l l r e t u r n then she i s hoping that some event w i l l occur; i f Jack mentions that John i s back, then he i s mentioning that a c e r t a i n s t a t e -o f - a f f a i r s i s the case; i f Jack has n o t i c e d that the l i g h t has turned green, then he has n o t i c e d that a c e r t a i n event has occurred. I suggest, then, that part of a paraphrase a n a l y s i s of 'that + sentence' clauses involves the s e l e c t i o n of a term from a l i s t of terms i n c l u d i n g 'proposition', ' s i t u a t i o n ' , 'event', ' s t a t e - o f - a f f a i r s ' so that the term s e l e c t e d a p p r o p r i a t e l y c h a r a c t e r i z e s the content of the 'that + sentence' complement i n question. And i n what follows I w i l l argue f o r the adequacy i n general of the format exemplified i n the f o l l o w i n g paraphrase for d e a l i n g with 'that + sentence' noun phrase complements. 4.44 i s the sentence being treated, and 4.45 i s the paraphrase v e r s i o n . 4.44 Jack n o t i c e d that the l i g h t turned green. 4.45 "The l i g h t turned green." The preceding p r o p o s i t i o n r e f e r s to an event- C a l l that event 'A'. j a c k n o t i c e d that event A occured. In 4,45 there i s s t i l l a 'that + sentence' noun phrase complement, 123 namely, 'that event A occurred.' This noun phrase complement is p a r a l l e l to the i n f i n i t i v a l complement discussed e a r l i e r , ' p r o p o s i t i o n A to be true', i n the r o l e i t plays i n the paraphrase a n a l y s i s . Both clauses may be introduced by r u l e s which permit a maximum of one clause per sentence, hence by non-recursive base r u l e s . Using such clauses a maximum of once per sentence m u l t i p l e embeddings o f ' t h a t + sentence 1 clauses can be e l i m i n a t e d . Given the adequacy of 4.45 as a paraphrase f o r 4.44, or in general, t h a t the s t y l e of paraphrase exemplified i n 4.45 i s adequate, we w i l l be able to undertake the e l i m i n a t i o n of m u l t i p l y embedded 'that + sentence' noun phrase complements. In each case where we have m u l t i p l y embedded 'that + sentence' complements we w i l l be able to paraphrase u s i n g a s e r i e s of sentence sequences of the form exemplified i n 4,45. The problems here are not so much with the usefulness of the paraphrase s t y l e to disentangle m u l t i p l y embedded 'that + sentence* clauses, but with the adequacy of the paraphrase s t y l e on i n d i v i d u a l c l a u s e s . Thus, i f we can adequately c h a r a c t e r i z e the contents of the 'that + sentence' clauses then i t doesn't matter whether the clauses themselves contain clauses or not. For example, f o r a sentence l i k e "John i n s i s t e d that Mary returned" our paraphrase v / i l l include "John ins i s ted. that event A occurred," and f o r a sentence l i k e "John i n s i s t e d that Mary discovered that Edna returned" our paraphrase w i l l include, again, "John 124 i n s i s t e d t h a t e v e n t A o c c u r r e d " -- o n l y i n t h e s e c o n d c a s e , e v e n t A w i l l r e f e r t o Mary's d i s c o v e r y and n o t to Edna's r e t u r n . To g i v e a more d e t a i l e d t r e a t m e n t , and o f a s e n t e n c e w h i c h i n c l u d e s many k i n d s o f v e r b s , c o n s i d e r : 4.46 J a c k remembered t h a t Mary d i s c o v e r e d t h a t J o h n knew t h a t Edna b e l i e v e d t h a t Tim l e a r n t t h a t P a t r e t u r n e d . T h i s c a n be p a r a p h r a s e d b y t h e f o l l o w i n g s e r i e s o f s e n t e n c e s , u s i n g ' t h a t + s e n t e n c e " c l a u s e s a maximum o f once p e r s e n t e n c e , and b e g i n n i n g w i t h t h e i n n e r m o s t embedded c l a u s e : 4.47 " P a t r e t u r n e d . " The p r e c e d i n g p r o p o s i t i o n r e f e r s t o an e v e n t . C a l l t h a t e v e n t 'A'. "Tim l e a r n t t h a t e v e n t A o c c u r r e d . " C a l l t h e p r e c e d i n g p r o p o s i t i o n ' B ' . "Edna b e l i e v e d p r o p o s i t i o n B t o be t r u e . " C a l l t h e p r e c e d i n g p r o p o s i t i o n ' C . "John knew p r o p o s i t i o n C t o be t r u e . " The p r e c e d i n g p r o p o s i t i o n r e f e r s t o a s t a t e - o f - a f f a i r s . C a l l t h a t s t a t e - o f - a f f a i r s 'D'. "Mary d i s c o v e r e d t h a t s t a t e - o f - a f f a i r s D o b t a i n e d . " The p r e c e d i n g p r o p o s i t i o n r e f e r s t o an e v e n t . C a l l t h a t e v e n t 'E ' . J a c k remembered t h a t e v e n t E o c c u r r e d . I w i l l now c o n s i d e r t h e a d e q u a c y o f t h e p a r a p h r a s e s t y l e e x e m p l i f i e d i n 4.45. T h e r e a r e f o u r b a s i c p r o b l e m s h e r e : (1) I s t h e r e a s u f f i c i e n t l y r i c h l i s t o f c h a r a c t e r i z e r s so t h a t w h a t e v e r we may s a y v i a a ' t h a t + s e n t e n c e ' c l a u s e we can p a r a p h r a s e u s i n g t h e proper-name s t y l e o f p a r a p h r a s e , d ependent as i t i s upon s u p p l y i n g a c h a r a c t e r i z e r ? (2) Can we s u c c e s s f u l l y r e f e r t o and i n d i v i d u a t e p a r t i c u l a r e v e n t s , s i t u a t i o n s , s t a t e s - o f - a f f a i r s , p r o c e s s e s e t c . v i a s u c h s t a t e m e n t s as "The p r e c e d i n g p r o p o s i t i o n r e f e r s t o an e v e n t . 125 C a l l that event 'A'."? (3) Can we give proper names to any and a l l i n d i v i d u a l events, p r o p o s i t i o n s , s i t u a t i o n s , s t a t e s - o f - a f f a i r s , processes, etc.? The proper name s t y l e of paraphrase requires that we be able to do so. (4) The proper name s t y l e of paraphrase would force speakers to c h a r a c t e r i z e where our language p r e s e n t l y does not. How damaging i s t h i s c o n s t r a i n t f o r the adequacy of the para-phrase s t y l e ? I w i l l consider each of these problems i n t u r n . (1) I want to argue that E n g l i s h has a l i s t of chara c t e r -i z e s s u f f i c i e n t l y r i c h so that whatever i s captured i n a 'that + sentence' clause can be a p p r o p r i a t e l y c h a r a c t e r i z e d by a term such as 'pro p o s i t i o n ' , ' s i t u a t i o n ' , 'event', ' s t a t e - o f - a f f a i r s ' or 'process'. In a d d i t i o n to t h i s e m p i r i c a l point, I want to argue the t h e o r e t i c a l p o i n t that the problem of c h a r a c t e r i z e r choice — even i f i t were insuperable f o r some language — would not be s u f f i c i e n t to undermine the cl a i m that we do not need r e c u r s i v e l y derived s t r u c t u r e s to express our thoughts. A language might be poor i n c h a r a c t e r i z e r s , but the enrichment of i t s l i s t could s t i l l be undertaken (without i n v o l v i n g r e c u r s i v e mechanisms.) To "the extent that the demonstration of the em p i r i c a l point about the s u f f i c i e n c y of the E n g l i s h l i s t of c h a r a c t e r i z e r s i s s u c c e s s f u l , my e n t e r p r i s e does not hinge on the success of the demonstration of the t h e o r e t i c a l p o i n t . 126 One way, i f not the only way, to demonstrate the e m p i r i c a l p o i n t i s to do i t from a sample of cases. The f o l l o w i n g i s a l i s t of verbs which c h a r a c t e r i s t i c a l l y take 'that + sentence' complements, and the l i s t i s , I believe, r e p r e s e n t a t i v e of the types of verbs which take 'that + sentence' complements: 'know', 'believe', ' a f f i r m ' , 'deny', 'allege', 'argue', 'learn', 'see' (as i n "John saw that he was wrong"), 'seem' (as i n " i t seems that John was wrong"), 'propose', ' r e a l i z e ' , 'say', 'hope', 'notice', 'mention', 'remember', 'dream' (as i n "I dreamt that I won the I r i s h " ) , 'fear', 'hint', 'discover', and 'warn'. In general, to know that (where the dots are to be f i l l e d with any s e n t e n t i a l complement) i s to know that some p r o p o s i t i o n i s true, or to know some p r o p o s i t i o n to be true, or to know that some event has occurred or i s o c c u r r i n g or w i l l occur, or to know that some s i t u a t i o n or s t a t e - o f - a f f a i r s was, i s , or w i l l be the case. And the same i s true of the other verbs i n the l i s t . I t would be tedious to go through them a l l , but consider, say, that to discover t h a t . . . . . i s to discover that some s i t u a t i o n obtains (or obtained, or w i l l obtain) or that some event has occurred (or i s o c c u r r i n g or w i l l occur), and that s i m i l a r l y , to remember t h a t . . . . i s to remember that some event occurred, or some s i t u a t i o n or s t a t e -o f - a f f a i r s obtained, and so f o r t h . There are, to be sure, 127 e x c e p t i o n a l verbs, but these too have appropriate c h a r a c t e r i z e r s . Thus, f o r example, to propose that the p r i c e be dropped by 5* per l b . i s not to propose that some s i t u a t i o n or s t a t e - o f -a f f a i r s w i l l obtain, nor i s i t to propose that some event w i l l occur, but i t i s to propose a proposal, and far such a s p e c i a l verb, the term 'proposal' would serve as the appropriate c h a r a c t e r i z e r . Coming back to the non-exceptional verbs, even i n the most d i f f i c u l t cases, where the s e n t e n t i a l complements r e f e r to human agents (e.g."John knew that Mary discovered the c u l p r i t , " ) we can a p p r o p r i a t e l y c h a r a c t e r i z e the events captured by the clau s e s . Even i f c h a r a c t e r i z i n g an act of disc o v e r y as an event, say, i s not an exhaustive c h a r a c t e r i z a t i o n , i t i s s u f f i c i e n t f o r our purposes. N a t u r a l l y , d i s t i n c t i o n s are i n order between the d i f f e r e n t kinds of events, processes, s t a t e s - o f - a f f a i r s , e t c . but from no p h i l o s o p h i c a l standpoint that I am aware of would i t make sense to say that the things which can be r e f e r r e d to v i a s e n t e n t i a l complements cannot be ch a r a c t e r i z e d as 'events', ' s i t u a t i o n s ' , ' s t a t e s - o f - a f f a i r s ' , 'processes', e t c . And even i f one were to disagree, say, that a discovery may be a p p r o p r i a t e l y c h a r a c t e r i z e d as an event, presumably one would be able to o f f e r some other more appropriate c h a r a c t e r i z a t i o n (e.g. rather than "John knew that event A occurred," "John knew that act A occurred."). 128 We have seen enough examples, I think, to warrant the conc l u s i o n that E n g l i s h has a l i s t of c h a r a c t e r i z e r s s u f f i c i e n t l y r i c h to handle the problem of c h a r a c t e r i z e r choice. But over and above t h i s e m p i r i c a l p o i n t i s the t h e o r e t i c a l one that the l i s t of c h a r a c t e r i z e r s i s not f i x e d , and may, I claim, be enriched. In what follows I w i l l explore one p o s s i b i l i t y f o r the enrichment of the l i s t of c h a r a c t e r i z e r s . Consider the f o l l o w i n g sentences, foc u s i n g on the r e l a t i o n s h i p s between the verbs and the objects of the verbs. 4 .48 Mary thought a thought. 4 .49 Mary b u i l t a b u i l d i n g . 4 .50 Mary invented an i n v e n t i o n . 4 .51 Mary ass e r t e d an a s s e r t i o n . 4 .52 Mary proposed a p r o p o s i t i o n . 4 .53 Mary cooked a cookie. 4 .54 Mary discussed a d i s c u s s i o n . 4 .55 Mary s a t i r i z e d a s a t i r e . What c l a s s i f i e s these sentences as i n some sense belonging together i s the f a c t that i n each case the verb and the object have a common stem. However, w i t h i n t h i s c l a s s of sentences, there i s a d i v i s i o n defined by the f o l l o w i n g c o n t r a s t s : i t i s not true that whatever one may cook i s a cookie, but i t i s true that whatever one may invent i s an invention; i t i s not 129 true that whatever one may s a t i r i z e i s a s a t i r e — one may f o r example, s a t i r i z e p o l i t i c i a n s , and p o l i t i c i a n s are not s a t i r e s ; but i t i s true that whatever one may a s s e r t i s an a s s e r t i o n . In some sense, the noun 'thought' serves to c h a r a c t e r i z e whatever can stand as an object of the verb 'thought'. S i m i l a r l y , the noun ' a s s e r t i o n ' c h a r a c t e r i z e s whatever can stand as object of the verb 'assert'. On the other hand, the noun 'cookie' does not c h a r a c t e r i z e whatever can stand as object of the verb 'cook', nor does the noun 'discussion' c h a r a c t e r i z e whatever can stand as object of the verb 'discuss* — one can discuss the moon, f o r example, as w e l l as a d i s c u s s i o n . The moon i s not a d i s c u s s i o n . Now i n the manner of the terms 'as s e r t i o n ' , 'proposition', and 'invention', although not i n the manner of the terms 'b u i l d i n g ' , 'cookie', 'discussion', and ' s a t i r e ' , could we not use the stems of the verbs which take noun clauses to make words which would c h a r a c t e r i z e the objects of t h e i r r e s p e c t i v e verbs? Consider, f o r example, the verb ' t e s t i f y ' , and the foll o w i n g sentence with ' t e s t i f y ' : 4.56 The witness t e s t i f i e d that the accused was i n F l o r i d a at the time of the crime. Without using any new c h a r a c t e r i z e r s or new appropriate objects, we could paraphrase 4.56 along the l i n e s p r e v i o u s l y developed with: 130 4.57 "The accused was i n F l o r i d a at the time of the crime." The preceding p r o p o s i t i o n r e f e r s to a s t a t e - o f - a f f a i r s . C a l l that s t a t e - o f - a f f a i r s 'A'. The witness t e s t i f i e d that s t a t e - o f - a f f a i r s A obtained. Or with: 4.58 "The accused was i n F l o r i d a at the time of the crime." C a l l the preceding p r o p o s i t i o n 'A'. The witness t e s t i f i e d that p r o p o s i t i o n A was a f a c t . But now we can a l s o consider the p o s s i b i l i t y of using new nouns l i k e ' t e s t i f i c a t i o n ' , and a paraphrase v e r s i o n l i k e : 4.59 "The accused was i n F l o r i d a at the time of the crime." The preceding p r o p o s i t i o n r e f e r s to a t e s t i f i c a t i o n . C a l l t h i s t e s t i f i c a t i o n 'A'. The witness t e s t i f i e d t e s t i f i c a t i o n A. Here ' t e s t i f i c a t i o n ' i s a new word using the stem of the verb ' t e s t i f y ' , and i s a term which c h a r a c t e r i z e s whatever can stand as object of the verb ' t e s t i f y ' . (Whether ' t e s t i f i c a t i o n ' i s an unecessary a d d i t i o n to the vocabulary of E n g l i s h , con-s i d e r i n g that E n g l i s h already has the words ' t e s t i m o n i a l ' and 'testimony' i s not c l e a r ; but probably there are a number of reasons why i f we used 'testimony' or 'testimonial' instead of ' t e s t i f i c a t i o n ' we would at l e a s t be modifying the use of the words'testimony' and 'testimonial'.) Using such nouns, the l i s t of c h a r a c t e r i z e r s would be g r e a t l y enriched. For each verb which takes a 'that + sentence' noun phrase complement we could have a c h a r a c t e r i z e r appropriate to that p a r t i c u l a r verb. Now undoubtedly there are problems associated with such a r a d i c a l expansion of the c h a r a c t e r i z e r l i s 131 F o r example, t h e q u e s t i o n m i g h t be r a i s e d w h e t h e r t h e s e new terms a r e p r o p e r l y d e s c r i b e d as c h a r a c t e r i z e r s , as t h e more g e n e r a l t e r m s ' f a c t ' , ' s i t u a t i o n ' , ' e v e nt e t c . , a r e o r w h e t h e r t h e y a r e r e a l l y j u s t g r a m m a t i c a l dummies i n d i c a t i n g t h e p r e s e n c e o f a p p r o p r i a t e o b j e c t s . However, I w i l l n o t a t t e m p t t o do j u s t i c e t o s u c h q u e s t i o n s . I o n l y want t o r e i n f o r c e t h e p o i n t t h a t t h e p r o b l e m o f c h a r a c t e r i z e r c h o i c e i s a f u n c t i o n o f t h e l a n g u a g e as i t e x i s t s , and t h a t l a n g u a g e s m i g h t be c o n s t r u c t e d w i t h f a r r i c h e r l i s t s o f c h a r a c t e r i z e r s and te r m s d e s i g n a t i n g - a p p r o p r i a t e o b j e c t s . ( 2 ) I want t o c l a i m t h a t we c a n s u c c e s s f u l l y r e f e r t o and i n d i v i d u a t e e v e n t s , s i t u a t i o n s , s t a t e s - o f - a f f a i r s , and s o on v i a s u c h s t a t e m e n t s a s , "The p r e c e d i n g p r o p o s i t i o n r e f e r s t o an e v e n t ( s i t u a t i o n ) ( s t a t e - o f - a f f a i r s ) . C a l l t h a t e v e n t ( s i t u a t i o n ) ( s t a t e - o f - a f f a i r s ) ' A ' . " L e t ' s t a k e , f o r example, t h e r e f e r e n c e made i n 4.47: 4.47 " P a t r e t u r n e d . " The p r e c e d i n g p r o p o s i t i o n r e f e r s t o an e v e n t . C a l l t h a t e v e n t 'A'. T h e r e c a n b e l i t t l e q u e s t i o n h e r e o f w h i c h e v e n t i s b e i n g r e f e r r e d t o . However, i f t h e r e f e r e n c e were u n c l e a r t o someone, and t h e q u e s t i o n was r a i s e d , "Which e v e n t ? " o r "What e v e n t ? " , we c o u l d s a y t h a t t h e e v e n t was, " P a t ' s r e t u r n . " F o r f u r t h e r c l a r i f i c a t i o n , t h e n , t h e f o l l o w i n g f o r m a t i s a v a i l a b l e : 4.47a " P a t ' s r e t u r n . " The p r e c e d i n g p h r a s e r e f e r s t o an e v e n t . C a l l t h a t e v e n t 'A'. 132 Now l e t ' s consider a more d i f f i c u l t case (again from 4.47): 4.47 "Mary discovered that s t a t e - o f - a f f a i r s D obtained." The preceding p r o p o s i t i o n r e f e r s to an event. C a l l that event 'E'. There are two verbs here, 'discovered' and 'obtained', but the event wh ch we want to r e f e r to i s Mary's discovery. Thus, i f there i s a d i f f i c u l t y with 4.47 because of the presence of two verbs i n the f i r s t sentence, f u r t h e r c l a r i f i c a t i o n i s p o s s i b l e v i a n o m i n a l i z a t i o n . Here a f u r t h e r c l a r i f i c a t i o n would read something l i k e : 4.47b "Mary discovered that s t a t e - o f - a f f a i r s D obtained." The preceding p r o p o s i t i o n r e f e r s to a discovery by Mary. C a l l Mary's discovery, 'event E'. Notice that s i n c e the nominalizations 'a discovery by Mary' and 'Mary's disc o v e r y ' cut out the object of the verb, sentences l i k e , "The preceding p r o p o s i t i o n r e f e r s to a d i s c o v e r y by Mary" and " C a l l Mary's discovery, 'event E'" can be generated using base r u l e s which permit a maximum of one embedded clause per sentence. This i s an important p o i n t as i t shows that c l a r i f i c a t i o n s l i k e 4.47b can be made without i n v o l v i n g r e c u r s i v e base r u l e s . I w i l l consider one f u r t h e r example, again from 4.47: 4.47 "John knew p r o p o s i t i o n C to be t r u e . " The preceding p r o p o s i t i o n r e f e r s to a s t a t e - o f - a f f a i r s . C a l l that s t a t e - o f - a f f a i r s 'D'. I argue here that we frequently i n d i v i d u a t e s t a t e s - o f - a f f i a r s v i a p r o p o s i t i o n s i n j u s t t h i s way, and that we do not need to further c l a r i f y by introducing a nominalization l i k e 'John's 133 knowing'. Someone might, f o r example, say"John knows that Mary has been informed of the w r i t and that's the s t a t e - o f - a f f a i r s at the present time." Further c l a r i f i c a t i o n of which s t a t e - o f -a f f a i r s i s unecessary. I claim, then, that i t i s p o s s i b l e to i n d i v i d u a t e events, s i t u a t i o n s , e t c . , v i a such statements as "The preceding propo-s i t i o n (phrase) r e f e r s to an event ( s i t u a t i o n ) ( s t a t e - o f - a f f a i r s ) .' (3) Given that we can s u c c e s s f u l l y r e f e r to and i n d i v i d u a t e events, s i t u a t i o n s , s t a t e s - o f - a f f a i r s e t c . v i a such statements as "The preceding p r o p o s i t i o n r e f e r s to an event ( s i t u a t i o n ) ( s t a t e - o f - a f f a i r s ) " the attempt to name these events, s i t u a t i o n s , s t a t e s - o f - a f f a i r s etc., w i l l not be problematic. This follows, I claim, because any i n d i v i d u a t e d event, s i t u a t i o n , s t a t e - o f -af f a i r s , process or whatever can be named. We do i n f a c t give proper names not only to m a t e r i a l objects, animals, people, places, i n s t i t u t i o n s , etc., but a l s o to events, s i t u a t i o n s , and s t a t e s - o f - a f f i a r s . Not only do we name r e a l animals, (Fido) and unreal animals (Pegasus), but we al s o name events which have occurred (The Boston Tea Party, The D e f e n e s t r a t i o n of Prague), and events which have not yet occurred and perhaps never w i l l (The Second Coming). In add i t i o n , we name s i t u a t i o n s of various kinds, (Zeugzwang, i n chess). Without d i f f i c u l t y we can name events l i k e Pat's return or Mary's discovery, whether or not they have or w i l l take p l a c e . 134 Anything we can i n d i v i d u a t e (I'm t a l k i n g about the dog which..." or "I'm t a l k i n g about the event which...") we can r e f e r to pronomially ("that dog", "that event"). And anything we can r e f e r to pronomially can serve as the bearer of a name ("Call t h a t dog,'Fido'", " C a l l t h a t event, 'A'"). Now i t i s true that questions l i k e "Is event A i d e n t i c a l with event B?" are not resolved by the g i v i n g of the names, but to r a i s e questions of i d e n t i t y here would be i r r e l e v a n t s i n c e these questions are present regardless of whether the events have been given names, ("Is t h i s event the same as that event?"), I conclude that we can give proper names to the events, p r o p o s i t i o n s , s i t u a t i o n s , s t a t e s - o f - a f f a i r s , etc., which are r e f e r r e d to i n 'that + sentence' clauses. (4) F i n a l l y , I want to suggest that having to character-i z e i s not a severe r e s t r i c t i o n . I t ' s t r u e that having to c h a r a c t e r i z e forces the speaker to provide some information which he or she would not otherwise have done. A person who says "I remembered that John was i n Spokane" i s not c h a r a c t e r i z i n g the object of the memory as an event, a s i t u a t i o n , or whatever, whereas the paraphrase format does force the c h a r a c t e r i z a t i o n ("I remembered that s i t u a t i o n A obtained".) But the r e l a t i o n s h i p between the object being c h a r a c t e r i z e d and the c h a r a c t e r i z e r i s a conceptual one; f o r c i n g the speaker to c h a r a c t e r i z e i s only f o r c i n g the speaker to make e x p l i c i t what i s already i m p l i c i t . For that reason, the 135 imposition cannot be regarded as serious enough to place the paraphrase s t y l e i n jeopardy. In general, I take i t as a datum (as does Strawson i n the passage quoted above), that we do c h a r a c t e r i z e the contents of noun phrase complements when we are forced to, i . e . , that there are s i t u a t i o n s when we f i n d i t convenient, and, sometimes, unavoidable. In f a c t , i f a Quinean were to object that sentences do not r e f e r to p r o p o s i t i o n s , and that the r e i f i n g of such e n t i t i e s i s bad philosophy, I would challenge him or her to t r y to e x p l a i n the content of some very complex r e c u r s i v e l y d e r i v e d sentence, f u l l of noun phrase complements, and ask whether such a sentence, and the s t r u c t u r a l d e s c r i p t i o n which we can a s s o c i a t e with i t must be taken to 'express a thought' which we, i n v i r t u e of our performance l i m i t s w i l l never be able to grasp. On such an occasion, I take i t as a datum to be accounted f o r , and not something which needs to be j u s t i f i e d , t h a t we are capable of grasping the thoughts, and that we do proceed to produce various c h a r a c t e r i z e r s to help us i n the grasping of the 'message'. An argument that having to c h a r a c t e r i z e i s a serious imposition might be constructed, however, claiming that i f we are forced to provide c h a r a c t e r i z e r s "there w i l l be a systematic gap i n the expressive c a p a c i t y of the language. How, f o r example, would we paraphrase a sentence l i k e : 13 6 4-60 John r e a l i z e d that the melting of snow was a process. (Although the problematic n o m i n a l i z a t i o n i n t h i s example, 'the melting of snow' i s a kind of r e c u r s i v e l y derived s t r u c t u r e which I attempt to el i m i n a t e l a t e r on, I w i l l consider t h i s problem here i n order to deal with the deepest problems of the proper name s t y l e of paraphrase together; and the paraphrase e l i m i n a t i o n of'the melting of snow' w i l l i nvolve the proper name s t y l e of paraphrase.) Part of the paraphrase of a sentence l i k e 4.60 us i n g the proper name s t y l e of paraphrase, w i l l i nclude something l i k e : 4.61 John r e a l i z e d that process A was a process. And the argument would be that r e a l i z i n g the melting of snow to be a process d i f f e r s from r e a l i z i n g some process to be a process — i n f a c t , i f the users of a language were o b l i g e d to use c h a r a c t e r i z e r s i t would be d i f f i c u l t f o r 4.61 to have any meaning. To t h i s c r i t i c i s m I would argue that a sentence l i k e 4.61 can indeed be given an appropriate reading. Using Quine's 'transparent' and 'opaque' d i s t i n c t i o n (see Word & Object #30), i t i s c l e a r l y arguable that on the opaque co n s t r u a l of 4.61 any paraphrase of 4.60 containing 4.61 would be inadequate. For an opaque co n s t r u a l of 4.61 'process A was a process' i s t a u t o l o g i c a l , and i t i s unclear as to how anyone can have r e a l i z a t i o n s about t a u t o l o g i e s . But then the opaque co n s t r u a l 137 i s not the r e l e v a n t c o n s t r u a l since the problem with the tautology on the opaque reading of 4.61 i s not present i n 4.60 (And i f one were to suspect that i t i s , then the paraphrase i n c l u d i n g 4.61 would no longer be inadequate to 4.60 s i n c e the opaque reading of 4.61 W O u l d r e f l e c t the problematic status of the o r i g i n a l , 4.60.) The s u b s t a n t i a l question i s , can 4.61 be given a t r a n s -parent c o n s t r u a l ? I f i t can, there i s no reason why we would have to r e j e c t 4.61 as an inadequate c o n s t r u c t i o n . Although I can f i n d no reason to challenge the p o s s i b i l i t y of construing 4.61 t r a n s p a r e n t l y , i t might s t i l l be worthwhile to mention paraphrases f o r 4.60 other than 4.61- These other paraphrases may perhaps most u s e f u l l y be regarded as j u s t i f i c a t i o n s f o r the p o s s i b i l i t y of g i v i n g 4.61 a transparent reading, i . e . , as p r o v i d i n g i n a somewhat concrete way j u s t what the transparent reading or readings of 4.61 i s or might be. I t appears that a sentence l i k e 4.60 can only be meaningful i f we understand i t i n one of two ways: John i s r e a l i z i n g that some i s a process, and we have some c h a r a c t e r i z e r other than 'process' a v a i l a b l e to f i l l i n the blank; or John i s r e a l i z i n g that 'the melting of snow1 r e f e r s to a process. This might be e a s i e r to see from" another example. What does i t mean to say that John r e a l i z e d that L e s l i e was a woman, unless i t means e i t h e r that John r e a l i z e d that the name ' L e s l i e ' , which he had been using i n some context, r e f e r r e d , i n that context, 138 to a woman, or that John r e a l i z e d that some person, whom he knew to be named ' L e s l i e ' , was a woman? i n the second case, we do have another c h a r a c t e r i z e r for L e s l i e a v a i l a b l e , namely 'person'. I f a person has no other way of c h a r a c t e r i z i n g the melting of snow than with 'process', then perhaps to r e a l i z e that the melting of snow i s a process i s to r e a l i z e that the melting of snow r e f e r s to a process. I f one argues that one can r e a l i z e the melting of snow to be a process without r e a l i z i n g t h a t 'the melting of snow' r e f e r s to a process, then perhaps another c h a r a c t e r i z e r i s l u r k i n g , e.g. 'sequence of events', 'happening' or even 'thing'. In r e a l i z i n g the melting of snow to be a process i n that case, one i s r e a l i z i n g that some sequence of events, happening (or thing) i s a process. The problem i s more complex i n s o f a r as 'the melting of snow' i s a s e n t e n t i a l n o m i n a l i z a t i o n whereas ' L e s l i e ' i s a proper name, and so the p a r a l l e l might f a i l . But i n any case, and t h i s i s the response I would l i k e to s t r e s s , i t i s c l e a r that the paraphrase language would not be forced to use some p o s s i b l y inadequate phrase l i k e "John r e a l i z e d that process A was a process", but would have at l e a s t two ways of expressing that thought. I conclude that having to c h a r a c t e r i z e i s not an over-whelmingly serious d e f i c i e n c y of the proper name s t y l e of para-phrase. 1 3 9 Looking at the matter from a broader perspective, I b e l i e v e that i t i s i n s t r u c t i v e that such a problem should appear at a l l . We are t r y i n g to eliminate noun clauses, and i n doing so are coming up against an apparent requirement that we c h a r a c t e r i z e the content of these clauses. Why i s t h i s so? Perhaps the reason i s t h i s ; a noun clause nominalizes, but i t does not c h a r a c t e r i z e th*? content o f what i t nominalizes (at l e a s t not normally; some might: the K i p a r s k i s , e.g., cla i m that f a c t i v e noun clauses have ' f a c t ' heading them i n deep s t r u c t u r e ) . On the other hand, i f we do not use a nominalized sentence, i . e . , a noun clause, i t i s d i f f i c u l t to nominalize without a l s o c h a r a c t e r i z i n g . Thus, f o r example i f we say "John hopes that the guests w i l l a r r i v e , " the p r o p o s i t i o n "the guests w i l l a r r i v e " h a s been nominalized, but no character-i z a t i o n of the content of t h i s n o m i n a l i z a t i o n i s provided. However, when i t comes to naming, p r o v i d i n g a c h a r a c t e r i z e r i s usu a l . Thus, the usual form f o r naming i s exemplified i n such naming acts as "I name t h i s s h i p the Queen Mary," or "Let us r e f e r to t h i s argument as the Cosmological Argument." In the f i r s t case, the object being named was c h a r a c t e r i z e d as a ship, and i n the second, the object being named was c h a r a c t e r i z e d as an argument. One might elaborate on t h i s f u r t h e r and suggest that i t is not only usual to provide a c h a r a c t e r i z e r when naming, but that the a v a i l a b i l i t y of some c h a r a c t e r i z e r i s a necessary co n d i t i o n f o r the success of naming. 140 Consider that i n general the more abs t r a c t the object being named, the more pronounced i s the tendency to provide the c h a r a c t e r i z e r when g i v i n g the name. i t i s ea s i e r to name a dog without invoking the c h a r a c t e r i z e r 'dog' than i t i s to name an argument without invoking the c h a r a c t e r i z e r 'argument', and i t i s e a s i e r to name an argument without invoking the c h a r a c t e r i z e r 'argument' than i t i s to name a fear without invoking the c h a r a c t e r i z e r 'fear'. But t h i s does not suggest that g i v i n g proper names can be done without the a v a i l a b i l i t y of c h a r a c t e r i z e r s . I f anything, i t suggests the contrary, that g i v i n g a proper name i s dependent on having a c h a r a c t e r i z e r a v a i l a b l e . The reason that i t i s e a s i e s t to give a proper name without a c t u a l l y g i v i n g the c h a r a c t e r i z e r with a concrete o b j e c t i s t h a t i t i s so easy, i f r e q u i r e d , to supply the c h a r a c t e r i z e r . We have an explanation, then f o r our coming up against the c h a r a c t e r i z e r problem. The g i v i n g of proper names depends on the a v a i l a b i l i t y of c h a r a c t e r i z e r s . (For a view s i m i l a r to t h i s one about the dependency of proper names on c h a r a c t e r i z e r s , see Peter Geach, Reference and Generality, Chapter 2 #34.) Before concluding, I would l i k e to consider one fu r t h e r (miscellaneous) o b j e c t i o n to the s t y l e of paraphrase f o r noun clauses introduced above. I t might be hypothesized that a sentence with an i n d e x i c a l expression (such as 'the previous proposition') r e f e r r i n g to something p r e v i o u s l y introduced 141 l i n g u i s t i c a l l y contains the l i n g u i s t i c expression i t s e l f i n the deep s t r u c t u r e headed by the category covering the i n d e x i c a l expression, I can suggest no reasons f o r enter-t a i n i n g the hypothesis, but i t i s coherent and thus perhaps worth c o n s i d e r i n g on those grounds alone. The hypothesis would hold that a sentence l i k e : 'The previous p r o p o s i t i o n i s about an event' has i n i t s deep s t r u c t u r e the l i n g u i s t i c expression through which the i n d e x i c a l expression 'the previous p r o p o s i t i o n ' r e f e r s to some p r o p o s i t i o n . Thus the deep s t r u c t u r e of "The previous p r o p o s i t i o n i s about an event" when 'the previous p r o p o s i t i o n ' r e f e r s to the p r o p o s i t i o n that i t rai n e d yesterday would be something l i k e : Sentence Np \ S / \ Np Vp I / \ N V adv I | \ the p r o p o s i t - i t rained yesterday i s about an event xon And the hypothesis would hold that t h i s i s transformed i n t o a surface s t r u c t u r e with two sentences l i k e : Adv. Det index N I I I \ Det N rained yesterday the previous 1 i s J v , p r o p o s i t i o n about an event 1 4 2 (Here 'S' i s 'Sentence', ' Np' i s 'Noun phrase', 'Vp' i s Verb phrase', 'Det' i s Determiner' 'Vc' i s Verb complement', 'Adv' i s 'Adverb', 'P' i s 'Prepo s i t i o n ' ; the d e t a i l s of the c o n s t i t u e n t a n a l y s i s are not r e l e v a n t here; what i s r e l e v a n t i s the combination of the deep s t r u c t u r e s of the two sentences i n t o a s i n g l e deep st r u c t u r e . ) The suggestion would have a number of p e c u l i a r features --p r i n c i p a l l y , the breakdown of u n i f i e d r e c u r s i v e l y d e r ived s t r u c t u r e s i n t o simpler n o n - r e c u r s i v e l y derived s t r u c t u r e s through t r a n s f o r m a t i o n s . This i s an unusual suggestion, th e r e f o r e , but i f i t were to hold, the a n a l y s i s presented here would be s e r i o u s l y damaged, since the paraphrase s t y l e would s t i l l be dependent on r e c u r s i o n i n the deep s t r u c t u r e . However, consider the f o l l o w i n g two sentences; 4.62 The f o l l o w i n g sentence ends with a p e r i o d . The previous sentence ends with a p e r i o d . The two sentences are r e a d i l y comprehensible, and both are impeccably grammatical, nor are there any paradoxes of references such as the l i a r paradox whose ghosts they irrevocab r a i s e . The references of the expressions 'the f o l l o w i n g sentence' and 'the previous sentence' are q u i t e c l e a r , and both sentences are unparadoxically t r u e . Yet i f the hypothesis introduced above were to hold, the deep s t r u c t u r e of the f i r s t of these sentences would be ( i n i t i a l l y ) : 143 the sentence the previous sentence ends with a p e r i o d ends with a p e r i o d The sentence t r i a n g u l a t e d here would be analysed as a sentence with an i n d e x i c a l noun phrase r e f e r r i n g to something introduced l i n g u i s t i c a l l y , and so the deep s t r u c t u r e of the f i r s t sentence would be f u r t h e r s p e c i f i e d as: S the sent- the f o l l o w i n g . , ^ ence sentence ends ends with a peri o d ends with a p e r i o d with a p e r i o d However, the t r i a n g u l a t e d sentence is now i d e n t i c a l to the whole sentence whose deep s t r u c t u r e we are attempting to provide Under the hypothesis that expressions such as 'the f o l l o w i n g sentence' or 'the preceding sentence' are analyzed as i n c l u d i n g 144 the sentences being r e f e r r e d to, the deep s t r u c t u r e of the two sentences of 4.63 would each have to be i n f i n i t e l y l a r g e . Any s t r u c t u r a l d e s c r i p t i o n , however, must be f i n i t e , and so the hypothesis i s impossible. To conclude, I want to clai m that the proper name s t y l e of paraphrase provides a workeable s t y l e f o r 'that + sentence' noun phrase complements. The paraphrase involves c h a r a c t e r i z i n g the contents of the embedded p r o p o s i t i o n , g i v i n g a proper name to the content so c h a r a c t e r i z e d , and then l e t t i n g the proper name repl a c e the embedded clause as object of the main verb, or l e t t i n g a s i n g l e n o n - r e c u r s i v e l y derived clause containing the proper name replace the embedded clause as object of the main verb. In the case of m u l t i p l y embedded 'that + sentence' noun phrase complements, one would paraphrase beginning with the inmost embedding. We can express any thought we want using a maximum of one 'that + sentence' noun phrase complement per sentence. The proper name s t y l e of paraphrase, with s u i t a b l e v a r i a t i o n s , forms the ba s i s of much of the paraphrase a n a l y s i s which f o l l o w s . 'What' cla u s e s : Some verbs can be followed by what I s h a l l c a l l 'what' clauses. A t y p i c a l example of such a verb i s 'hold'; and a t y p i c a l example of a sentence with a 'what' clause i s : 5.1 John now holds what he pr e v i o u s l y denied. 14 5 We are not here d i s c u s s i n g cases, however, where we have a noun phrase complement whose f i r s t word i s 'what' and i n which 'that' can be i n s e r t e d p r i o r to 'what' and f o l l o w i n g the main verb. As an example of the l a t t e r c l a s s of sentences, consider: 5.2 John thinks what the stationmaster d i d was r i g h t . I f we i n s e r t the word 'that' between 'thinks' and "what' we have the grammatical sentence: 5.3 John thinks that what the stationmaster d i d was r i g h t By c o n t r a s t , 'that' cannot be s i m i l a r l y i n s e r t e d i n 5.1; i f i t i s i n s e r t e d , the r e s u l t i n g sentence i s ungrammatical 5.4* John now holds that what he p r e v i o u s l y denied. In 5.1 there i s only one s e n t e n t i a l noun phrase, namely 'what he p r e v i o u s l y denied'; i n 5.2 there are two s e n t e n t i a l noun phrases, namely 'what the stationmaster d i d was r i g h t ' , and 'what the stationmaster d i d ' . Here we are co n s i d e r i n g 'what5 clauses when they are the objects of verbs; i n 5.2 the l a r g e clause is a 'that + sentence' clause with 'that' deleted, and the 'what' clause, 'what the stationmaster did* i s the subject of the verb 'was'. Later we w i l l deal with 'what' clauses when they are the subjects of verbs. lWhat' clauses d i f f e r from 'that + sentence' noun phrase complements i n that the object of the verb i n the complement is the noun 'what', and thus the name 'what' clauses, rather 146 than, say, 'what + sentence' c l a u s e s . The 'what' i s part of the sentence, and not a separate element. i n 'that + sentence" complements, the verb i n the embedded sentence may be t r a n s i t i v e or i n t r a n s i t i v e ; i n 'what' clauses the verb must be t r a n s i t i v e since i t takes 'what' as an object. Furthermore, and t h i s i s an important feature of these clauses, since the 'what' i s always the ob j e c t of the verb, i t i s impossible to have a noun phrase complement embedded i n s i d e a 'what' clause. The 'what' blocks t h i s s i n c e i t i s already given as the object of the verb of the embedded clause. Thus we cannot have a sentence l i k e : 5.5* John now holds what he p r e v i o u s l y denied that the c h i r o p r a c t o r i s a doctor. S i m i l a r l y , although we can have: 5.6 John b e l i e v e s what Mary b e l i e v e s that Tim b e l i e v e s , we cannot have: 5.7* John b e l i e v e s what Mary b e l i e v e s that Tim b e l i e v e s that c h i r o p r a c t o r s are doctors. In 5.6, 'what' i s the object of the t h i r d verb (counting from the l e f t ) and thus blocks the a d d i t i o n of the noun phrase complement 'that c h i r o p r a c t o r s are doctors' to the t h i r d 'believe' i n 5.7. Also, we cannot have: 5.8* John b e l i e v e s what Mary b e l i e v e s what Tim b e l i e v e s . As before, the presence of the f i r s t 'what', i f i t i s taken to be the object of the f i r s t verb 'believes', blocks the a d d i t i o n of a 'what' clause as object of the f i r s t ' b e l i e v e s ' . 147 Thus a 'what' clause must be the l a s t or inmost embedded clause, and s i n c e 'what' clauses do not permit the a d d i t i o n of f u r t h e r clauses, i t i s not necessary to regard 'what' clauses as r e c u r s i v e l y derived s t r u c t u r e s . 'What' clauses can be introduced by base r u l e s which permit non-recursive s e n t e n t i a l embedding. However i t i s p o s s i b l e to e l i m i n a t e these clauses i n paraphrase, and s i n c e i t i s p r e f e r a b l e to e l i m i n a t e sent-e n t i a l embedding wherever p o s s i b l e , i t w i l l be worthwhile to demonstrate the e l i m i n a b i l i t y of 'what' c l a u s e s . Notice f i r s t that the sentence c o n t a i n i n g the 'what' clause presupposes the t r u t h of a p r o p o s i t i o n underlying the clause. For example, i n 5.1, the main sentence 'John now holds what he p r e v i o u s l y denied' presupposes that John p r e v i o u s l y denied something. (To put i t more acc u r a t e l y , i f someone were to s i n c e r e l y say that John now holds what he p r e v i o u s l y denied, then they would be presupposing that John p r e v i o u s l y denied something.) This gives us a headstart i n making a paraphrase: 5.1 John now holds what he p r e v i o u s l y denied. 5.8 John p r e v i o u s l y denied something. C a l l that something 'A'. John now holds A-In t h i s paraphrase, the term 'something' serves as the charac-t e r i z e r f o r the proper name. If we want to paraphrase using a c h a r a c t e r i z e r which i s i n some way s p e c i f i c a l l y appropriate to the verb (though not n e c e s s a r i l y uniquely appropriate), we could paraphrase 5.1 with: 148 5.9 John p r e v i o u s l y denied some p r o p o s i t i o n . John now holds that p r o p o s i t i o n . or: 5.10 John p r e v i o u s l y denied some p r o p o s i t i o n . C a l l that p r o p o s i t i o n 'A'. John now holds p r o p o s i t i o n A. This s t y l e of paraphrase bears a marked resemblance to the s t y l e used f o r r e l a t i v e clause e l i m i n a t i o n , and t h i s f a c t provides us with a clue as to how to regard the 'what' cla u s e s . From the p o i n t of view of d e s c r i p t i v e l i n g u i s t i c s , the fo l l o w i n g observations must be taken as purely — perhaps w i l d l y — s p e c u l a t i v e , but from the point of view of t h i s study, i t i s worthwhile to see how 'what' clauses can be regarded as transformations of r e l a t i v e c l a u s e s . Consider the r e l a t i o n s h i p between: 5.1 John now holds what he p r e v i o u s l y denied. and: 5.11 John now holds the p r o p o s i t i o n which he p r e v i o u s l y denied. To take another example, consider the r e l a t i o n s h i p between the fol l o w i n g p a i r of sentences: 5.12 John wants what you have. 5.13 John wants that which you have. The n a t u r a l way to paraphrase both 5.1 and 5.11 i s given i n 5.10 and the n a t u r a l way to paraphrase both 5.12 and 5.13 would be: 5.14 You have something. John wants that. The suggestion that 'what' clauses can be regarded as transformed r e l a t i v e clauses, however, i s merely that — a suggestion. The 149 paraphrase a n a l y s i s of 'what' clauses i s not dependent on i t . ( Z e l l i g H a r r i s made a s i m i l a r p o i n t about r e l a t i v e clauses as un d e r l y i n g 'what' clauses, apparently, i n l e c t u r e s ; f o r t h i s information I am indebted to Thomas Patton.) There i s a d i s t i n c t c l a s s of sentences whose surface s t r u c t u r e resembles that of 'what' clauses. These are what I w i l l c a l l 'what' i n d i r e c t questions. A t y p i c a l example i s : 5.15 John asked what the problem was. The deep s t r u c t u r e of such sentences i s d i f f e r e n t from that of 'what' c l a u s e s . For one thing, they cannot be regarded as transformations of r e l a t i v e c l a u s e s . But more important than t h e i r d i f f e r e n c e s i n t h i s context i s t h e i r s i m i l a r i t y i n that they need not be considered as r e c u r s i v e l y d e r i v e d s t r u c t u r e s . Thus: 5.15b *John asked what Mary asked what Tim asked, i s u n i n t e l l i g i b l e . 'What' i n d i r e c t questions, l i k e other 'what' clauses cannot be used r e c u r s i v e l y . The demonstration that follows, then, i s unnecessary, but i t i s i n t e r e s t i n g to observe that 'what' i n d i r e c t questions, l i k e other 'what' clauses are eliminable i n paraphrase. Thus, for 5.15 we could have: 5.15c "What was the problem?" John asked the preceding question. or: 5.15d "What was the problem?" C a l l the preceding question John asked question A. 150 This u n f o r t u n a t e l y sounds as though John spoke the very words "What was the problem?", and t h i s i s not n e c e s s a r i l y implied by the o r i g i n a l 5.15. We would have to s t i p u l a t e some d i s t i n g u i s h i n g convention between r e p o r t i n g quotes and 'de-modalizing' quotes, i . e . , quotes which s t r i p a p r o p o s i t i o n of i t s usual f o r c e . This would s u r e l y be an important convention to have i n any case. To summarize, I have argued that 'what' clauses cannot be used r e c u r s i v e l y , and that they are i n any case e l i m i n a b l e i n paraphrase. Secondly, a c l a s s of sentences s i m i l a r i n surface s t r u c t u r e to 'what' clauses, namely 'what' i n d i r e c t questions, although they d i f f e r i n deep s t r u c t u r e from 'what' clauses proper, are l i k e 'what' clauses proper i n that they cannot be used r e c u r s i v e l y . I argue as w e l l that they are i n any case eliminable i n paraphrase. 'Whether' and ' i f noun phrase complements. Verbs l i k e 'wonder', 'inquire' 'know', 'discover', ' f i g u r out', and 'ask' can take complements introduced by 'whether' and ' i f ; I s h a l l c a l l these 'whether' and ' i f complements. T y p i c a l examples of such complements are: 6.1 Mary wondered whether John would r e t u r n . 6.2 Mary wondered i f John would r e t u r n . Here, as before, we are not concerned with those sentences i n which 'that' i s i n s e r t a b l e . We are not concerned, f o r example 151 with: 6.3 Mary supposed i f John would r e t u r n so would h i s br o t h e r . In 6.3 we can i n s e r t 'that* so as to y i e l d the grammatical sentence: 6.4 Mary supposed that i f John would r e t u r n so would his b r o t h e r . In n e i t h e r 6 . 1 nor 6 . 2 can 'that' be so i n s e r t e d ; i f we d i d we would have the f o l l o w i n g ungrammatical sentences: 6.5 *Mary wondered that whether John would r e t u r n . 6.6 *Mary wondered that i f John would r e t u r n . In 6.3 there i s a 'that + sentence' noun phrase complement, namely, ' (that) i f John would r e t u r n so would h i s brother' which i t s e l f i s composed of two co-ordinated clauses, ' i f John would r e t u r n ' and 'so would h i s brother'. Such co-ord i n a t i o n s we w i l l deal with l a t e r . Now i t i s easy to see that 'whether' and ' i f ' clauses, l i k e r e l a t i v e clauses and l i k e 'that + sentence' noun phrase comple-ments, but u n l i k e 'what' clauses, can be used to construct i n d e f i n i t e l y many sentences. A s i n g l e example here w i l l s u f f i c e : 6.7 Mary wondered whether John wondered whether Tim wondered whether Edna would r e t u r n . In 6.7 the f i v e dots i n d i c a t e the p o s s i b i l i t y of i n d e f i n i t e l y many nested complements. Perhaps the best (but not the only) way to t r e a t 'whether' and ' i f clauses i s to use them, permitting a maximum of one such 152 clause per sentence, along the l i n e s of the ( b ) - s t y l e of para-phrase developed on page 119 f o r 'that + sentence' noun phrase complements. Under that paraphrase s t y l e , the sentence "Mary knows that Harry i s b a l d " was paraphrased by the sequence of sentences " C a l l the f o l l o w i n g p r o p o s i t i o n 'A'.. 'Harry i s b a l d . ' Mary knows A to be t r u e . " Here too we would have to choose appropriate c h a r a c t e r i z e r s . For example, i f Mary wonders whether John l e f t , then she i s wondering whether a c e r t a i n event has occurred so 'event' would be the appropriate c h a r a c t e r i z e r . Thus f o r : 6.7a Mary wondered whether John l e f t . we could paraphrase with: 6.7b "John l e f t . " The preceding p r o p o s i t i o n r e f e r s to an event. C a l l that event 'A'. Mary wondered whether event A occurred. As was the case with 'that + S' noun phrase complements, the advantage of r e p l a c i n g one 'whether' noun phrase complement with another appears when we are d e a l i n g with clauses which are m u l t i p l y embedded. Thus, i f Jack i s wondering whether Mary wondered whether John l e f t , then Jack i s s t i l l wondering whether a c e r t a i n event took place; so that: 6.7c Jack wondered whether Mary wondered whether John l e f t , can be paraphrased by the f o l l o w i n g s e r i e s of sentences i n which 153 'whether' clauses are permitted to occur a maximum of once per sentence: 6.7d "John l e f t . " The preceding p r o p o s i t i o n r e f e r s to an event. C a l l that event 'A'. "Mary wondered whether event A occurred." The preceding p r o p o s i t i o n r e f e r s to an event. C a l l t h a t event 'B'. Jack wondered whether event B occurred. Thus we have one option f o r paraphrase using a maximum of one 'whether' clause per sentence. Other options are a l s o a v a i l a b l e ; however, I take i t that the f i r s t option i s adequate and that what follows i s only of secondary i n t e r e s t . To paraphrase such clauses, we might a l s o , f o r example, make e x p l i c i t reference to the truth-value of the p r o p o s i t i o n . I f we take i t that the f o l l o w i n g two sentences are synonymous then we can paraphrase the f i r s t u s i ng the s t y l e of paraphrase already developed to deal with the second: 6.1 Mary wondered whether John would r e t u r n . 6.13 Mary wondered what the truth-value of the p r o p o s i t i o n 'John would re t u r n ' was. 6.13 can be paraphrased i n the f o l l o w i n g way, using the s t y l e developed f o r 'what' clauses which'contain other p r o p o s i t i o n a l elements: 6.14 "John would r e t u r n . " C a l l the preceding p r o p o s i t i o n 'A'. Mary wondered what the truth-value of p r o p o s i t i o n A was . However t h i s paraphrase i s dependent on the synonymy of 6.1 and 6.13, and does not improve on the paraphrase already developed 154 si n c e i t too i s dependent on s e n t e n t i a l embedding used a maximum of once per sentence. In a l l these examples we have been d e a l i n g with 'whether' cl a u s e s . ' I f ' clauses work i n j u s t the same way as 'whether' clauses; i n f e c t , the general r u l e i s that sentences with ' i f ' clauses are synonymous with sentences with 'whether* clauses so long as the two sentences are i d e n t i c a l except that where one has the word 'whether', the other has the word " i f . Thus 6.1 and 6.2 are synonymous, and 6.2 can be paraphrased by 6.9 6.1 Mary wondered whether John returned. 6.2 Mary wondered i f John returned. 6.9 "John returned." The preceding p r o p o s i t i o n r e f e r s to an event. C a l l that event 'A'. Mary wondered whether event A occurred. (or: Mary wondered i f event A occurred.) To summarize, 'whether' and ' i f * clauses can be used recur-s i v e l y . Recursive uses of these clauses can be eliminated i n para-phrase. One s t y l e involves c h a r a c t e r i z i n g the content of the clause and using a maximum of one 'whether'or ' i f clause per sentence. 'Why and 'how' noun phrase complements. A c l a s s of complements very s i m i l a r to 'whether' and ' i f complements i s the c l a s s of 'why' and 'how' complements. The cl a s s of verbs which take 'whether' and ' i f complements appears to be i d e n t i c a l to the cl a s s of verbs which take 'why' and 'how' 155 complements. Thus we have 7.1 Mary wondered John had returned. knew asked f i g u r e d out in q u i r e d discovered whether why i f how In 7.1 any of the verbs w i l l go with any of the complements. But 'why' and 'how' complements d i f f e r from 'whether' and " i f complements i n one important respect: these verbs used with 'why' and 'how' complements are f a c t i v e , whereas the same verbs followed by 'whether' and ' i f complements are not. To ask why John returned, or to wonder why John returned, i s to presuppose that John returned; s i m i l a r l y , to f i g u r e out how John returned, or to wonder how John returned i s to pre-suppose that John returned. But to wonder i f John returned i s not to presuppose that John returned; and again, to wonder whether John returned i s not to presuppose that John returned. This may suggest that the concept of ' f a c t i v e ' involves two f a c t o r s : the verb, and the kind of cl a u s e . (For a s i m i l a r account of t h i s p o i n t see Fact, by P. and C. Ki p a r s k i . ) The verb 'know' followed by a 'that + sentence' clause i s f a c t i v e , but followed by a 'whether' or ' i f clause i s not f a c t i v e ; the verb 'guess' followed by a 'that + sentence' clause i s not f a c t i v e , b u t followed by a 'how' or 'why' clause may be f a c t i v (Thus i f one says "Mary guessed how John returned" one may be presupposing that John returned; i n f a c t , one u s u a l l y i s 156 presupposing that John returned.) This poi n t does not play a d i r e c t r o l e in the argument here, but i t does suggest that the concept of ' f a c t i v e ' cannot be regarded simply as a c l a s s -i f i c a t i o n of verbs, but must a l s o take i n t o account the r e l a t i o n s h i p between the verb and the type of clause which follows i t . Here I s h a l l c a l l both verbs and clauses ' f a c t i v e ' or 'non-f a c t i v e ' . I f , i n making an utterance which contains a verb and a noun phrase complement, one i s presupposing the t r u t h of the p r o p o s i t i o n expressed i n the complement, then I s h a l l c a l l e i t h e r the verb or the clause ' f a c t i v e ' . This i s not a wholly s a t i s f a c t o r y approach i n general terms, because i t does not provide a c l a s s i f i c a t i o n which w i l l enable one to know j u s t by the verb and the kind of complement whether the utterance w i l l i n v o l v e a f a c t i v e p r e s u p p o s i t i o n — but f o r the purpose of t h i s study i t w i l l s u f f i c e , as the a n a l y s i s here proceeds v i a p a r t i c u l a r examples. To r e t u r n from t h i s d i g r e s s i o n , the f a c t that 'how' and 'why' complements are f a c t i v e makes paraphrase e l i m i n a t i o n of them somewhat e a s i e r . Beginning with 'how' complements, to wonder how a task was accomplished i n general i s to wonder by what means or i n what manner the task was accomplished. Thus, consider: 7.2 Mary wondered how John had returned. We can paraphrase 7.2 with: 1 5 7 7.3a John had returned in some manner. Mary wondered what that manner was. or : 7.3b John had returned by some means. Mary wondered what those means were, 'Why* complements can be eliminated i n a s i m i l a r fashion; to know why John returned i s to know the reason f o r John's return, or the cause of John's r e t u r n . Take f o r example: 7.4 Mary knew why John returned. We could paraphrase 7.4 with e i t h e r : 7.5 Something caused John's r e t u r n . Mary knew what that cause was. or: 7.6 John returned f o r some reason. Mary knew what that reason was. In 7.5 we have the phrase 'John's retur n ' , which i s most l i k e l y a s e n t e n t i a l n o m i n a l i z a t i o n (but see Chomsky, Some Remarks on Nominalization which e n t e r t a i n s the idea that 'return' be l i s t e d i n the l e x i c o n as n e i t h e r verb nor noun.) I f i t i s a nominali-zation, then i t would rec e i v e an a n a l y s i s as a r e c u r s i v e l y der-ived s t r u c t u r e , and would eventually have to be eliminated i n paraphrase. Later s t r u c t u r e s such as these w i l l be tr e a t e d as nominalizations and discussed with regard to paraphrase e l i m i n -a t i o n ; here i t w i l l s u f f i c e to f u r t h e r break down the f i r s t sentence i n 7.5, so that we have the f o l l o w i n g paraphrase f o r 7.4: 7.7 John returned. Something caused t h a t event. Mary knew that cause. The choice between 7.7 and 7.6 of course i s not an a r b i t r a r y 158 one since the two paraphrases are not synonymous; I take i t that the d i s t i n c t i o n between 7.6 and 7.7 r e f l e c t s an ambiguity i n the meaning of the o r i g i n a l , 7.4. Now 'why' and 'how' complements can be used to form m u l t i p l y embedded c o n s t r u c t i o n s . For example: 7.8 Mary wondered why Tim wondered why Edna wondered why John returned. Here we could paraphrase with: 7.9 John returned f o r some reason. C a l l that reason 'A'. Edna wondered what reason A was. Edna wondered what reason A was f o r some reason. C a l l t h a t reason 'B'. Tim wondered what reason B was. Tim wondered what reason B was f o r some reason. C a l l that reason ' C . Mary wondered what reason C was. In making the paraphrase, i t was assumed that i n each clause i n the o r i g i n a l the wondering was about a reason rather than a cause; the paraphrase s t y l e using 'cause' instead of 'reason', or using combinations of 'cause' and 'reason' would be s i m i l a r . To summarize, 'how' and 'why' complements follow from the same verbs which take ' i f and 'whether' clauses, but utterances with 'how' and 'why' clauses c h a r a c t e r i s t i c a l l y involve f a c t i v e presuppositions whereas utterances with 'whether' and ' i f clauses do not. In a general way both s i n g l y and m u l t i p l y embedded 'why' and 'how' clauses are eliminable i n paraphrase by making e x p l i c i t references to 'manner', 'means', 'cause' or ' reason' . 159 Noun phrase complements as subjects of verbs. Having t r e a t e d noun phrase complements when they appear as objects of main verbs or verbs which are main verbs of noun phrase complements, the treatment of noun phrase complements when they are subjects i n r e l a t i o n to the main verbs i s not a d i f f i c u l t matter. Again I w i l l begin by co n s i d e r i n g 'that + sentence' c l a u s e s . 8.1 That Mary went i s a c e r t a i n t y . 'That + sentence' complements appearing as the subject of the verb present few probelms. Thus f o r 8.1, we could paraphrase: 8.2 "Mary went." C a l l the preceding p r o p o s i t i o n 'A'. P r o p o s i t i o n A i s a c e r t a i n t y . 'What' clauses can u s u a l l y be paraphrased by g i v i n g a proper name to whatever object the 'what' stands f o r , as i n the f o l l o -wing example: 8.3 What Mary d i d was r i g h t . 8.4 Mary performed some a c t i o n . C a l l that a c t i o n 'A'. A c t i o n A was r i g h t . Of course the 'what' can r e f e r to an object as w e l l as to an a c t i o n ; but t h i s does not cause f u r t h e r complications: 8.5 What Mary saw was h o r r i f y i n g . 8.6 Mary saw something. C a l l that t h i n g 'A'. A was h o r r i f y i n g . (The temptation i s to l e t the l a s t sentence of 8.6 read 'A was h o r r i f y i n g to Mary'; but the o r i g i n a l does not a c t u a l l y imply that the object was h o r r i f y i n g to Mary, as i t only a s s e r t s that 160 some object or other was h o r r i f y i n g . i f i n some context the i m p l i c a t i o n i s that the object was h o r r i f y i n g to Mary then the l a s t sentence of 8.6 could of course include 'to Mary'.) 'Whether' and ' i f complements can be eliminated along the l i n e s used to e l i m i n a t e them when they were objects of verbs. For example: 8.7 Whether Mary went i s i n doubt. 8.8 "Mary went." C a l l the preceding p r o p o s i t i o n 'A'. The t r u t h value of A i s i n doubt. Again, 'how' and 'why' complements can be paraphrased by making reference to means, manner, cause, or reason. Thus, f o r example: 8.10 How Mary went i s unknown. 8.11 Mary went (in some manner)(by some means). This (manner)(means) i s unknown. And: 8.12 Why Mary went i s unknown. 8.13a Mary went f o r some reason. This reason i s unknown, b Mary went. Something caused t h i s a c t i o n . That cause i s unknown. F i n a l l y , i t i s worth observing that m u l t i p l e i t e r a t i o n s of s e n t e n t i a l aubjects are most u n c h a r a c t e r i s t i c , i f not ungra-mmatical i n E n g l i s h . A t y p i c a l example would be: 8.14a That that that Mary saw John i s a c e r t a i n t y i s a tautology i s f a l s e . Nevertheless, some might claim that 8.14a follows the r u l e s of E n g l i s h grammar, and must be regarded as grammatical. For that reason i t i s appropriate to note that such m u l t i p l e i t e r a t i o n s 161 present no s p e c i a l problems. 8.14a, f o r example, can be paraphrased by: 8.14b "Mary saw John." C a l l the previous p r o p o s i t i o n 'A'. "P r o p o s i t i o n A i s a c e r t a i n t y . " C a l l the previous p r o p o s i t i o n 'B'. "P r o p o s i t i o n B i s a tautology." C a l l the previous p r o p o s i t i o n * C . P r o p o s i t i o n C i s f a l s e . I conclude, then, that noun phrase complements as subjects of verbs are e l i m i n a b l e i n paraphrase. Noun phrase complements f o l l o w i n g nominalized verbs. Nominalizations of various kinds give r i s e to s t r u c t u r e s which can be i t e r a t e d i n d e f i n i t e l y . Consider, f o r example: 9.1 John's b e l i e f that Mary had returned was ignored. In 9.1, the noun phrase complement 'that Mary had returned" i s the object of a verb which has been nominalized. The verb i s 'believed' and the no m i n a l i z a t i o n i s ' b e l i e f * . Now p r o p o s i t i o n s l i k e 9.1 presuppose p r o p o s i t i o n s l i k e : 9.2 John b e l i e v e d that Mary had returned. We can, then, paraphrase 9.1 by the f o l l o w i n g : 9.3 John b e l i e v e d that Mary had returned. This b e l i e f was ignored. S i m i l a r l y , 9.4 can be paraphrased by 9.5: 9.4 John's r e p u d i a t i o n of what he p r e v i o u s l y s a i d was t i m e l y . 9.5 John repudiated what he p r e v i o u s l y s a i d . This r e p u d i a t i o n was tim e l y . 162 I have already discussed the e l i m i n a t i o n of the 'that + sentence' and 'what' clauses such as are present i n 9.3 and 9.5; so f o r 9.1 we would have 9.6: 9.6 "Mary had returned." John b e l i e v e d the preceding p r o p o s i t i o n . This b e l i e f was ignored. And f o r 9.4 we could have e i t h e r 9.5 (as 'what' clauses can be introduced without r e c u r s i v e r u l e s , ) or: 9.7 John p r e v i o u s l y s a i d something. He repudiated t h a t . This r e p u d i a t i o n was t i m e l y . Continuing on to 'whether' and ' i f complements, consider: 9.8 John's i n q u i r y whether (if)Mary had returned was postponed. This can be paraphrased by: 9.9 John i n q u i r e d whether ( i f ) Mary had returned. This i n q u i r y was postponed. 9.9 contains a 'whether' ( ' i f ) clause which i n turn could be eliminated as follows (using the second s t y l e of paraphrase for 'whether' c l a u s e s ) : 9.10 "Mary returned." John in q u i r e d what the t r u t h value o f the preceding p r o p o s i t i o n was. This i n q u i r y was postponed. F i n a l l y : 9.11 John's d i s c o v e r i n g why Mary returned was timely would be paraphrased by: 9.12 John discovered why Mary returned. This discovery was t i m e l y . and f u r t h e r by: 9.13a Mary returned f o r some reason. John discovered that reason. That discovery was timely. 63 or by: 9.13b Mary returned. Something caused that r e t u r n i n g . John discovered that cause. That discovery was t i m e l y . Sentences with many of these s t r u c t u r e s compounded are hard to come by, but they can be constructed. Consider f o r example: 9.14 John's b e l i e f that Mary's b e l i e f that Tim's b e l i e f that Edna had returned was untrue, was untrue was untrue,. I n t u i t i o n s may d i f f e r as to whether 9.14 should be considered as a grammatical E n g l i s h sentence; on the grounds that i t i s u n i n t e l l i g i b l e unless i t i s disentangled i n a mathematical s o r t of way, some may argue that the r u l e s of E n g l i s h should be constructed so as to block such sentences from formation. Others however might d i f f e r , and i t i s the r e f o r e appropriate to note that the sentence can be paraphrased without any r e c u r s i v e l y derived s e n t e n t i a l embedding: 9.15 "Edna had returned." Tim b e l i e v e d the preceding p r o p o s i t i o n to be t r u e . "Tim's b e l i e f was untrue." Mary b e l i e v e d the preceding p r o p o s i t i o n to be t r u e . "Mary's b e l i e f was untrue." John b e l i e v e d the preceding p r o p o s i t i o n to be t r u e . John's b e l i e f was untrue. 'Whether1 , ' i f , 'how1 , and 'why' complements when m u l t i p l y embedded f o l l o w s i m i l a r patterns i n paraphrase. Perhaps a s i n g l e example w i l l s u f f i c e : 9.16 John's i n q u i r y whether Mary's i n q u i r y whether Tim's i n q u i r y was postponed was postponed was postponed. Again, some might argue that sentences l i k e 9.16 are not r e a d i l y u n i n t e l l i g i b l e ; i n any case, we can paraphrase 9.16 with: 164 9.17 Tim made an i n q u i r y . Mary i n q u i r e d whether Tim's i n q u i r y was postponed. John in q u i r e d whether Mary's i n q u i r y was postponed. (I t h i n k i t w i l l be agreed, i n f a c t , that 9.14 and 9.16 are only i n t e l l i g i b l e i f we disentangle them somewhat along the l i n e s of 9.15 and 9.17, r e s p e c t i v e l y . ) The m u l t i p l y embedded 'whether' clauses of 9.16 have been disentangled i n 9.17 using a maximum of one 'whether' clause per sentence. 'For...to...' noun phrase complements. I w i l l now consider ' f o r . . . t o . . . ' noun phrase complements. T y p i c a l examples of these complements are: 1Q.1 Mary expected the doctor to examine John. 10.2 P u l l i n g the t r i g g e r caused the gun to f i r e . 10.3 For Metternich to leave A u s t r i a was a s t o n i s h i n g . In 10.1 the clause i s 'the doctor to examine John'; i n 10.2 i t i s 'the gun to f i r e ' ; and i n 10.3 i t i s 'For Metternich to leave A u s t r i a ' . In only the t h i r d case does the 'fo r ' a c t u a l l y appear i n the surface s t r u c t u r e of the sentence. These complements are c a l l e d ' f o r . . . t o . .. ' complements because of the evidence that the ' f o r ' appears i n the deep s t r u c t u r e and i s sometimes deleted i n the transformation i n t o the surface s t r u c t u r e . For a t r a n s f o r m a t i o n a l account of these complements, see Jacobs & Rosenbaum, Chapter 20. We can begin with the observation that i n many cases the ' f o r . . . t o . . . ' clause can be replaced with a 'that + sentence' 165 c l a u s e . Thus, fo r example, we can say: 10.4 Mary expected that the doctor w i l l examine John. 10.5 That Metternich l e f t A u s t r i a was a s t o n i s h i n g , instead of 10.1 and 10.3 r e s p e c t i v e l y . However i t i s not always p o s s i b l e to paraphrase ' f o r . . . t o ' . . ' complements with 'that + sentence' complements, as the f o l l o w i n g attempt for 10.2 i l l u s t r a t e s : 10.6 *That the t r i g g e r was p u l l e d caused the gun to f i r e . For that reason, I w i l l not attempt to analyze ' f o r . . . t o . . . ' noun phrase complements i n terms of 'that + sentence' noun phrase complements, (even though a p a r t i a l a n a l y s i s of them along those l i n e s might be p o s s i b l e to construct.) Let's begin the a n a l y s i s by c o n s i d e r i n g 10.2; the verb i s f a c t i v e ; thus we have a headstart i n the paraphrase, as the p r e s u p p o s i t i o n can be separated out and can be asserted as a separate sentence, without quotes. 10.7 The gun f i r e d . P u l l i n g the t r i g g e r caused t h i s t o happen. In the second sentence of 10.7, however, there i s s t i l l the phrase ' p u l l i n g the t r i g g e r ' which i s , presumably, a t r a n s -formation of a sentence i n t o a noun phrase. Thus, (to a n t i c i p a t e a l i t t l e ) we would f u r t h e r break 10.7 down so as to y i e l d something l i k e : 10.8 The gun f i r e d . C a l l that event 'E'. Someone p u l l e d the t r i g g e r . C a l l that a c t i o n 'A!. A c t i o n A caused event E. 1.66 ( 1Poss...ing' complements i n general w i l l be discussed below, see page 172 ) • Next, l e t ' s consider the example with the non-factive verb, 10. 10.1 Mary expected the doctor to examine John. To a s s e r t that Mary expected the doctor to examine John i s not to presuppose that the doctor d i d examine John, i s examining John, or w i l l examine John. The paraphrase, then, w i l l i n v o l v e the use of quotes: 10.9 "The doctor w i l l examine John." The preceding p r o p o s i t i o n r e f e r s to a p o s s i b l e event. Mary expected that event to occur. Now there are two problems here. One i s the tense of the verb i n the p r o p o s i t i o n . And the second i s the 'to occur" clause i n the l a s t sentence of the paraphrase. Foll o w i n g the usual approach to such clauses with i n t r a n -s i t i v e verbs, we can allow t h e i r c o n s t r u c t i o n , so long as the ru l e s permit only one such clause per sentence. There i s another a l t e r n a t i v e , as suggested e a r l i e r , and that i s to use a m o d i f i c a t i o n of the verb. Using such an approach, the para-phrase would read: 10.10 "The doctor w i l l examine John." The preceding p r o p o s i t i o n r e f e r s to a p o s s i b l e event. Mary expected^that event. 10.10 follows the c_ s t y l e of paraphrase developed on page 119. What then about the problem of tense? Now no time i s e x p l i c i t l y r e f e r r e d to i n 10.1, but we cannot understand 10.1 167 unless we understand that i n r e l a t i o n to Mary's act of expecting the examination i s a future event. i n general ' f o r . . . t o ' clauses contain no e x p l i c i t references to time, i . e . , the verbs are not tensed, but the events, s i t u a t i o n s , acts and so on which these verbs r e f e r to are not s p e c i a l kinds of 'timeless' events. In order to understand sentences with ' f o r . . . t o ' c l a u s e s , we must be able to r e l a t e the events and so on r e f e r r e d to i n the ' f o r . . . t o ' clauses to the events and so on r e f e r r e d to i n the main clauses with regard to time. In 10.9 or 10.10 the tense of the p r o p o s i t i o n i n quotes supplies t h i s r e l a t i o n s h i p , and thus makes e x p l i c i t what i s i m p l i c i t i n 10.1. Since we must regard some tense as being i m p l i c i t i n the verb of the ' f o r . . . . t o ' clause, and since the choice of tenses receives r e s t r i c t i o n s from the verb of the main clause, grammarians might want to analyze ' f o r . . . . t o ' clauses as being ambiguous and c o n t a i n i n g s p e c i f i c a t i o n s as to tense i n deep s t r u c t u r e . We need not explore t h i s p o s s i b i l i t y here, but note that the ' f o r . . . t o ' clauses may more p r o f i t a b l y be regarded as being ambiguous rather than n o n - s p e c i f i c as to tense. There are only a number of tense i n t e r p r e t a t i o n s p o s s i b l e for the verbs of the ' f o r . . . t o . . . ' clauses, and unless some tense i s supplied, the sentence cannot be i n t e r p r e t e d . Another s o l u t i o n to the problem of tense i s worth e n t e r t a i n i n g ; consider the c r e a t i o n of a modal a u x i l i a r y which 1.68 would enable one to make an a s s e r t i o n without having t o s p e c i f y one of the standard tenses such as present, past, future past, or past p e r f e c t — a modal which would be used e x c l u s i v e l y wherever we now use the ' f o r . . . t o " c o n s t r u c t i o n s . Borrowing on the sound and meaning, as w e l l , of the modal a u x i l i a r y 'may' I w i l l l e t 'moy' represent t h i s new modal a u x i l i a r y . The para-phrase f o r 10.1 would read, using t h i s modal: 10.11 "The doctor moy examine John" The preceding p r o p o s i t i o n r e f e r s to an event. Mary expected that event to occur. Now i n a matter l i k e the c r e a t i o n of a new modal a u x i l i a r y to do a p a r t i c u l a r job i t i s not easy to know what would count as a j u s t i f i c a t i o n f o r such an invention, and what would count as a good c r i t i c i s m of the invention, but i f there were to be any general o b j e c t i o n to t h i s p a r t i c u l a r modal, perhaps i t would be something l i k e the f o l l o w i n g : The new modal i s used to capture the ambiguous time of the o r i g i n a l i n f i n i t i v a l construc-t i o n , and yet the o r i g i n a l i n f i n i t i v a l c o n s t r u c t i o n w i l l not be s u f f i c i e n t to l e t the clause stand alone as a sentence. What i s there to suggest that the modal w i l l help the clause t o stand alone as a sentence? In answer to t h i s I would argue that the modal i s not intended as a device which automatically guarantees the v i a b i l i t y of the clause as a sentence. But such a sentence could be meaningful. Consider the fol l o w i n g sentence: 10.12 The doctor has examined John, or the doctor i s examining John, or the doctor w i l l examine John. 169 The meaning of t h i s sentence would not be i d e n t i c a l to the meaning of: 10.13 The doctor moy examine John. but i t would be s i m i l a r ; the modal would make the a s s e r t i o n , but leave the tense ambiguous over the whole range of tenses. A sentence l i k e : 10.14 The doctor had examined John, or the doctor has examined John, or the doctor i s examining John, or the doctor w i l l examine John, or the doctor w i l l have examined John. includes more tenses, but i t reads p e c u l i a r l y ; i t ' s hard, f o r instance, to t h i n k of 'the doctor had examined John" and 'the doctor has examined John' as a l t e r n a t i v e s . But t h i s p e c u l i -a r i t y i s not r e l e v a n t f o r us. The modal i s not intended as a r e p r e s e n t a t i o n of a sentence l i k e 10.14 which asserts a d i s -j u n c t i o n of a s t r i n g of p r o p o s i t i o n s , each of which contains another tense, but i s rather an assertion with an ambiguity of tense. The question might then be r a i s e d , why use a modal a u x i l i a r y ? Why not use the i n f i n i t i v a l form i t s e l f ? I f one did, the para-phrase would read: 10.15 "The doctor to examine John" The preceding p r o p o s i t i o n r e f e r s to an event. Mary expected that event to occur. Indeed, there i s no reason, other than to mark the m o d i f i c a t i o n of E n g l i s h grammar. 'The doctor to examine John* i s not a grammatical sentence i n E n g l i s h as i t i s p r e s e n t l y c o n s t i t u t e d , 170 and the departure must be marked. i f we underlined 'to* to mark the departure, the same purpose would be served. The paraphrase would then read: 10.16 "The doctor to examine John." The preceding propo-s i t i o n r e f e r s to an event. Mary expected that event to occur. 'To' or 'moy' would not n e c e s s a r i l y be c l a s s i f i e d as modal a u x i l i a r i e s , but could be considered to be s p e c i a l a u x i l i a r i e s of tense. F i n a l l y , I would l i k e to r e i t e r a t e that t h i s departure i s not o b l i g a t o r y ; 10.9, whose f i r s t sentence reads, "The doctor w i l l examine John", i s an adequate paraphrase of the o r i g i n a l sentence with the ' f o r . . . t o . . . ' embedding, 1 0 . L In general, ' f or . . . to . . . ' clauses, l i k e 'that •+ sentence clauses, (but u n l i k e 'poss....ing' clauses) can be e a s i l y used to create m u l t i p l e embeddings. These m u l t i p l y embedded ' f o r , , , t o . . . ' clauses can be eliminated i n paraphrase along the usual l i n e s . For example: 10.17 Mary expected Tim to expect Edna to expect the doctor to examine John. Using the c_ s t y l e of paraphrase with the '"moy8 auxiliary,, a paraphrase would read: 10.18 "The doctor moy examine John." The preceding-p r o p o s i t i o n r e f e r s to an (event) (act of examination) "Edna moy expecti that (event) (act of examination) The preceding p r o p o s i t i o n r e f e r s to an (event) (act of expectation) . Tim moy expect-^that (event) (act of expectation) "The preceding p r o p o s i t i o n r e f e r s to and (event) (act of expectation). Mary expectied that (event) (act of expectation.) 171 The c h a r a c t e r i z e r choice i s u n c e r t a i n here, not because none i s a v a i l a b l e , but because i t i s unclear which i s best. Perhaps, "The preceding p r o p o s i t i o n r e f e r s to a kind of event. Mary expected such an event to occur" or the l i k e should a l s o be considered. Sentences with m u l t i p l y embedded ' f o r . . . t o ' clauses can a l s o be handled using E n g l i s h grammar with no m o d i f i c a t i o n s . Thus 10.17 could be paraphrased by the f o l l o w i n g s e r i e s of sentences: 10.19 "The doctor w i l l examine John." The preceding p r o p o s i t i o n r e f e r s to an act of examination. C a l l that act 'A'. "Edna w i l l expect the doctor to perform act A." The preceding p r o p o s i t i o n r e f e r s to an act of expectation. C a l l that act *B' . "Tim w i l l expect Edna to perform act B." The preceding p r o p o s i t i o n r e f e r s t o an a c t of expectation. C a l l that, act 'C' . Mary expected Tim to perform act C. The ' f o r . . . t o ' clauses i n 10.19 are permitted a maximum of once per sentence. 10.19 might be c r i t i c i s e d on t h e grounds that the problem of c h a r a c t e r i z e r - choice was not adequately solved. Can we say that i f John expected Mary to do something, then John performed an act of expectation? In general, can we say that to expect i s to perform an act? I f not then we could use a s l i g h t l y d i f f e r e n t kind of paraphrase f o r 10.17, one patterned more on the model of 10.18, thus: 10.20 "The doctor w i l l examine John." The preceding p r o p o s i t i o n r e f e r s to an act of examination. "Edna w i l l expect that act of examination to occur." The preceding p r o p o s i t i o n r e f e r s to an act of 172 expectation. "Tim w i l l expect that act of expectation to occur." The preceding p r o p o s i t i o n r e f e r s to an act of expectation. Mary expected t h a t act of expectation to occur. We could a l s o use proper names to c l a r i f y some of the pronomial references of 10.18 and 10.20. To summarize, ' f o r . . . t o ' clauses do not always serve unique expressive f u n c t i o n s , being f r e q u e n t l y paraphraseable by 'that + S' clauses; but some ' f o r . . . t o ' clauses cannot be paraphrased using other c l a u s e s . The tense of the verb i n a ' f o r . . . t o ' clause i s l e f t u n s p e c i f i e d , but some tense i s always implied. In paraphrase we can use the implied tense or e l s e a new modal a u x i l i a r y or new a u x i l i a r y of tense. M u l t i p l e embeddings of ' f o r . , to' clauses are easy to construct, and can be eli m i n a t e d i n paraphrase u s i n g the same techniques as f o r the e l i m i n a t i o n of i n d i v i d u a l ' f o r . . . t o . . ' clauses, beginning with the inmost embeddings. 'Poss....ing' noun phrase complements. I turn now to 'poss....ing' noun phrase complements. A t y p i c a l example of such a complement i s : 11.1 Mary opposed John's making an a p p l i c a t i o n . The complement here i s 'John's making an a p p l i c a t i o n ' ; the name 'poss...ing' i s taken from the f a c t that the subject i s i n the possessive -- 'John's' -- and the verb -- i n t h i s case'making'— i s i n the gerundive. L 7 3 'Poss...ing' complements, l i k e ' f o r . . . t o . . . ' complements, do not s p e c i f y as to time. Thus i n 11.1 i t i s unclear whether John's a p p l i c a t i o n i s before, simultaneous with, or l a t e r than Mary's o p p o s i t i o n . For t h i s reason, one s t y l e of paraphrase w i l l use the new a u x i l i a r y , 'moy' or 'to': 11.2 "John to_ make an a p p l i c a t i o n . " Mary opposed t h i s a c t i o n . In general, 'poss...ing' complements do not lend them-selves to m u l t i p l e embeddings; however the f a c t that such m u l t i p l e embeddings are u n c h a r a c t e r i s t i c does not mean that the ru l e s of E n g l i s h prevent such m u l t i p l e embeddings from o c c u r r i n g . Thus we can have sentences l i k e : 11.3 Mary opposed Tim's preventing John's making an a p p l i c a t i o n . These m u l t i p l e embeddings of 'poss...ing' complements can be eliminated i n paraphrase f o l l o w i n g the usual procedure. 11.4 "John moy make an a p p l i c a t i o n . " "Tim moy prevent John's a c t i o n . " Mary opposed Tim's a c t i o n . Using a new modal or a u x i l i a r y of tense i s not the only way to paraphrase 'poss...ing' noun phrase complements. J u s t as with ' f o r . . . t o ' noun phrase complements, the events r e f e r r e d to by the 'poss...ing' clauses are not 'timeless'. Although they may only be p o s s i b l e as opposed to a c t u a l events, they are s t i l l events which are only p o s s i b l e to occur at designated times i n r e l a t i o n to the events r e f e r r e d to i n the main clauses of such 174 sentences. A sentence l i k e 11.1 i s ambiguous to the extent that i t does not make c l e a r the time r e l a t i o n s h i p between Mary's act of opposing and John's act of applying. In comm-u n i c a t i o n s i t u a t i o n s , context would disambiguate. In a paraphrase v e r s i o n which d i d not use the new modal or a u x i l i a r y of tense, the implied tenses could be used and the sentences would not be ambiguous i n the f i r s t p l a c e . Thus f o r 11.1 we could use one of the f o l l o w i n g sentences: a. "John a p p l i e d . " b. "John i s a p p l y i n g . " c. "John w i l l apply." followed by: 11.5 The preceding p r o p o s i t i o n r e f e r s to an a c t . c a l l t hat act 'A'. Mary opposed act A. The convention could be that the tense of the verb i n a., b., oi c. marks the time r e l a t i o n between the event r e f e r r e d to i n the main clause and the event of the 'poss...ing' clause. Verb phrase complements. I have r e f e r r e d above, (pages 94 - 99 ) to the reasons f o r d i s t i n g u i s h i n g between noun and verb phrase complements; at t h i s p o i n t I w i l l attempt to demonstrate the e l i m i n a b i l i t y of the l a t t e r i n paraphrase. The f i r s t p o i n t to note i s that verb phrase complements, u n l i k e t h e i r l i n g u i s t i c cousins, ' f o r . . t o . . ' noun phrase complements, are not c h a r a c t e r i s t i c a l l y paraphraseable using 'that + sentence' noun phrase complements. Thus, "John expect-1 7 5 Mary to be back e a r l y " , a sentence with a ' f o r . . . t o . . . ' noun phrase complement, i s roughly synonymous with 'John expects that Mary w i l l be back e a r l y ' , a sentence with a 'that + sentence' noun phrase complement. But the f o l l o w i n g sentence: 12.1 John persuaded Mary to be back e a r l y . which contains the verb phrase complement 'Mary to be back e a r l y ' i s not roughly synonymous with: 12.2 John persuaded Mary that she would be back e a r l y . On the other hand, verb phrase complements resemble ' f o r . , t o . . ' noun phrase complements i n s o f a r as both complements have verbs i n the i n f i n i t i v a l form and therefore are u n s p e c i f i c as to time. A c c o r d i n g l y , one paraphrase a n a l y s i s of verb phrase complements uses an u n s p e c i f i c tense as i n d i c a t e d by the new a u x i l i a r y , ' to_' . Before going i n t o p a r t i c u l a r cases, i t w i l l be u s e f u l to s p e l l out the p e c u l i a r i t i e s of the verb phrase complement with regard to expressive f u n c t i o n . Of a l l embedded-sentence s t r u c t u r e s the verb phrase complement i s perhaps the most complex. We have seen reasons why, i n a sentence with a verb phrase complement l i k e : 12.1 John persuaded Mary to come back e a r l y , we must regard the subject of the verb i n the complement as the object of the main verb. Thus, "Mary1 i s the object of the main verb 'persuade', and a l s o the subject of the complement's verb, 176 'come'. But with the ' f o r . . . t o . . . ' noun phrase complements, and i n general f o r noun phrase complements, the whole complement i s the object of the main verb. Thus i n : 12.3 John expected Mary to come back e a r l y . the whole noun phrase complement 'Mary to come back e a r l y ' i s the object of the main verb 'expected*. A f i r s t stab at a paraphrase f o r 12.1 then, would be something l i k e : 12.4 John persuaded Mary. Mary to come back e a r l y . This i s u n s a t i s f a c t o r y , however, because 'Mary t o come back e a r l y ' i s an a s s e r t i o n ; assuming John to have persuaded Mary to come back e a r l y , i t would not follow that Mary came back e a r l y , or w i l l come back e a r l y . So a second attempt would be: 12.5 John persuaded Mary. "Mary to come back e a r l y . " This v e r s i o n , however, does not s p e l l out the r e l a t i o n s h i p between the two sentences; how to represent the r e l a t i o n s h i p i s not immediately c l e a r . To gain some perspective, l e t ' s compare 12.5 with the paraphrase v e r s i o n of a ' f o r . . t o . . ' noun phrase complement l i k e 12.3. The paraphrase of 12.3 would be: 12.6 "Mary to come back e a r l y " The previous p r o p o s i t i o n r e f e r s to an event. C a l l that event 'A'. John expected event A to occur. In 12.6, 'event to occur', as usual, i s a clause permitted a maximum of once per sentence, and introduced by base r u l e s which permit non-recursive s e n t e n t i a l embedding. Now the 'event to occur' clause serves as what I e a r l i e r r e f e r r e d to as*an 177 appropriate o b j e c t ' ; here i t i s the appropriate object of the verb 'expect'. F i l l i n g out the p a r a l l e l , what i s missing from 12.5 i s reference to a s i m i l a r appropriate object of the verb 'persuade' which could then serve as a l i n k between the two sentences. Thus a b e t t e r paraphrase than 12.5 f o r 12.1 would read: 12.7 "Mary to_ come back e a r l y . " The preceding p r o p o s i t i o n r e f e r s to an a c t i o n . C a l l that a c t i o n 'A'. John persuaded Mary to perform a c t i o n A. In paraphrasing verb phrase complements, then, we could allow f o r the embedding of a maximum of one verb phrase comple-ment per sentence. The clause 'to perform a c t i o n A' f o r verb phrase complements would be a counterpart to clauses l i k e 'event A to occur' f o r 'that + S' noun phrase conp lements. L i k e ' f o r . . . t o . . ' noun phrase complements, verb phrase complements lend themselves to the formation of m u l t i p l e embeddings. Thus we can have sentences l i k e : 12.16 Mary persuaded Edna to encourage Tim to tempt John to r e t u r n . The s t y l e of paraphrase exemplified i n 12.7 can be used to disentangle m u l t i p l y embedded verb phrase complements. 12.16, for example, can be paraphrased by the f o l l o w i n g s e r i e s of sentences i n which verb phrase complements are permitted a maximum of once per sentence: 12.17 "John t_o r e t u r n . " The preceding p r o p o s i t i o n r e f e r s to an a c t i o n . C a l l that a c t i o n 'A'. "Tim to tempt John to perform a c t i o n A." The preceding p r o p o s i t i o n r e f e r s to an act of tempting. 178 C a l l that act of tempting 'B'. "Edna to encourage Tim to perform a c t i o n B." The preceding p r o p o s i t i o n r e f e r s to an act of encouraging. C a l l that act 'C . Mary persuaded Edna to perform act C. As with ' f o r . . . t o , . ' noun phrase complements and 'poss..inc. noun phrase complements, we need not r e l y on the c r e a t i o n of a new modal or a u x i l i a r y of tense. The actions r e f e r r e d to i n the verb phrase complements are not 'timeless' a c t i o n s . More-over, with verb phrase complements there i s u s u a l l y no ambiguit} of tense. Most verbs which take verb phrase complements, e.g. 'persuade', 'encourage', 'tempt', impose the r e s t r i c t i o n that the a c t i o n r e f e r r e d to i n the subsequent clause i s i n the f u t u r e with respect to the a c t i o n r e f e r r e d to i n the main clause. (I say 'most' because some verbs, l i k e 'force', do not impose such a r e s t r i c t i o n . ) In any case, with verb phrase complements, as with ' f o r . . t o . . ' noun phrase complements and 'poss..ing" complements, tense i s not s p e c i f i e d , but some time sequence i s imp l i e d . And the implied sequence can be represented using the tense system of E n g l i s h grammar as i t i s p r e s e n t l y c o n s t i -tuted . Thus f o r : 12.5 John persuaded Mary to come back e a r l y . we could paraphrase with: ^ 12.18 "Mary w i l l come back e a r l y . " The preceding p r o p o s i t i o n r e f e r s to an a c t i o n . C a l l that a c t i o n 'A'. John persuaded Mary to perform a c t i o n A. and f o r : 12.16 Mary persuaded Edna to encourage Tim to tempt J oh to r e t u r n . 179 we could paraphrase with: 12.19 "John w i l l r e t u r n . " The preceding p r o p o s i t i o n r e f e r s to an a c t i o n . C a l l that a c t i o n 'A'. "Tim w i l l tempt John to perform a c t i o n A." The preceding p r o p o s i t i o n r e f e r s to an act of tempting. C a l l that a ct 'B'. "Edna w i l l encourage Tim to perform act B." The preceding p r o p o s i t i o n r e f e r s to an act of encouraging. C a l l t h a t act 'C'. Mary persuaded Edna to perform act C. And i n 12.18 the convention would be that the tense s p e c i f i e d i n the p r o p o s i t i o n surrounded by quotes marks the r e l a t i o n s h i p between the event or a c t i o n r e f e r r e d to i n that p r o p o s i t i o n and the event or a c t i o n r e f e r r e d to i n the main clause of the l a s t sentence. In t h i s case, the ' w i l l ' of "Mary w i l l come back e a r l y " s p e c i f i e s that i n r e l a t i o n to the act of persuasion i n "John persuaded Mary to perform a c t i o n A", Mary's coming back e a r l y i s i n the f u t u r e . This convention thus makes e x p l i c i t what i s already i m p l i c i t i n 12.5 ("John persuaded Mary to come back e a r l y , " ) namely that Mary's coming back e a r l y i s i n the future i n r e l a t i o n to John's act of persuasion. And the convention would apply i n the same way to 12.19. Bearing t h i s convention i n mind, 12.19 should not be too d i f f i c u l t to fo l l o w . A l s o worth co n s i d e r i n g very b r i e f l y , i s the format exempli-f i e d i n the f o l l o w i n g paraphrase f o r 12.5: 12.20 John persuaded Mary to do the f o l l o w i n g t h i n g . Mary w i l l come back e a r l y . Using t h i s format, 12.16 would be paraphrased by: 180 12.21 Mary persuaded Edna to do the f o l l o w i n g t h i n g . Edna w i l l encourage Tim to do the f o l l o w i n g t h i n g . Tim w i l l tempt John to do the f o l l o w i n g t h i n g . John w i l l r e t u r n . This format i s a l i t t l e e a s i e r to follow, but i t i s not f r e e from problems. The a s s e r t i o n i n 12.20, for example, that Mary w i l l come back e a r l y i s too strong s i n c e i t only follows from John's having persuaded Mary to come back e a r l y that Mary intends to come back e a r l y , not that she w i l l . There are probably ways to s o f t e n these a s s e r t i o n s , (maybe even s y s t e m a t i c a l l y ) but s i n c e the s t y l e s exemplified i n 12.17 and 12.19 are p r e c i s e and systematic, (and t h e r e f o r e adequate, although d i f f i c u l t to follow) I w i l l not discuss t h i s t h i r d format any f u r t h e r . I claim, then, that r e c u r s i v e l y derived verb phrase complements are e l i m i n a b l e i n paraphrase. The p e c u l i a r i t y of verb phrase complements l i e s i n a kind of double object s t r u c t u r e which follows the verb. The paraphrase of a verb phrase comple-ment must take account of t h i s f a c t as w e l l as of the u n s p e c i f i e d tense of the i n f i n i t i v a l verb i n the o r i g i n a l . Two b a s i c approaches were developed, one using a new modal or a u x i l i a r y of tense, the other using only standard E n g l i s h grammar. Having t r e a t e d verb phrase complements, I have f i n i s h e d with the d i f f e r e n t kinds of s e n t e n t i a l embeddings. However, i n each case, i n demonstrating the e l i m i n a b i l f t y of m u l t i p l y embedded clauses, I looked at m u l t i p l e embeddings where each s t r u c t u r e 181 was of the same type. Thus sentences l i k e " E i n s t e i n s a i d that Newton s a i d that P l a t o s a i d that Socrates s a i d that sophists are ignorant" which contains a s e r i e s of embedded 'that + S' clauses were considered, and sentences l i k e , "Tom persuaded Mary to encourage Edna to tempt John to come back e a r l y " , which contains a s e r i e s of embedded verb phrase complements were a l s o considered. But sentences i n which the various kinds of s e n t e n t i a l embeddings are mixed were not considered, and a p r i o r i there i s no assurance that unique expressive functions are not undertaken by such mixtures. In the next s e c t i o n I w i l l examine some such 'mixed m u l t i p l e embeddings.' Mixed m u l t i p l e embeddings. Most cases of m u l t i p l e embeddings of mixed types do not present s p e c i a l problems; but r e l a t i v e clauses o c c u r r i n g i n s i d e noun phrase complements are i n t e r e s t i n g . Consider, f o r example: 13.1 Mary swore that the man who wore the hat l e f t f i r s t . Here we have a noun phrase complement, 'that the man who wore the hat l e f t f i r s t ' which contains the r e l a t i v e clause 'who wore the hat*. The d i f f i c u l t y l i e s i n the f a c t that Mary, i n swearing that the man who wore the hat l e f t f i r s t , i s not thereby swearing that some man wore the hat. Once again Quine's 'opaque' vs. 'transparent' d i s t i n c t i o n i s relevant (Quine, Word and Object, #30) On opaque reading 13.1 means that Mary used words to the e f f e c t 182 of "I swear that the man who wore the hat l e f t f i r s t . " But on transparent c o n s t r u a l , i t i s not implied that Mary h e r s e l f r e f e r r e d to the man who wore t h e h a t as t h e man who wore the hat. She may simply have pointed to someone and sworn that that person l e f t f i r s t . The i n d i v i d u a t i o n v i a 'who wore the hat' may be the speaker's d e s c r i p t i o n i n 13.1. The problem, however, i s by no means i n s o l u b l e . A paraphrase w i l l be a d e q u a t e i f i t leaves both the transparent and opaque readings open, and such paraphrases are not d i f f i c u l t to provide. A paraphrase l i k e : 13.2 "Some man wore the hat." "That man l e f t f i r s t . " Mary swore the immediately preceding p r o p o s i t i o n to be t r u e . leaves both opaque and transparent construals open. And f u r t h e r c l a r i f i c a t i o n s of 13.2 are p o s s i b l e without u s i n g r e c u r s i v e l y derived s t r u c t u r e s (for example, a c l a r i f i c a t i o n o f 13.2 s u c h as "Mary i n d i v i d u a t e d the man by p o i n t i n g " c a n be h a n d l e d a l o n g l i n e s d i s c u s s e d above.) Of course to show that mixed m u l t i p l e embeddings do n o t i n general i n v o l v e extra burdens for a paraphrase a n a l y s i s w o u l d be p r a c t i c a l l y impossible by way of enumeration o f t h e permu-t a t i o n s and combinations. The f o l l o w i n g example might serve as an i l l u s t r a t i o n , i n any case, of how a large s t r i n g of mixed m u l t i p l e embeddings might be approached: 13.4 Mary argued that what John s a i d about whether Tim had persuaded Edna to t a l k to the man who came e a r l y was t a c t f u l . Using the usual paraphrase approach appropriate to each t y p e of clause, we would have something l i k e : 183 13.5 Some man came e a r l y . "Edna w i l l t a l k to that man." The preceding p r o p o s i t i o n r e f e r s to an a c t i o n . C a l l that a c t i o n 'A' "Did Tim persuade Edna to perform a c t i o n A?" John s a i d something about that question. "What John s a i d was t a c t f u l . " Mary argued the preceding p r o p o s i t i o n . I t might be worth p o i n t i n g out here, s i n c e t h i s example combines so many types of clauses, that i n an informal and rough way even the conventions developed i n t h i s study are not e n t i r e l y necessary. Using a lower standard of paraphrase, a sentence l i k e 13.4 and, a f o r t i r i o r i , many of the sentences which have been considered, can be paraphrased with a more id i o m a t i c E n g l i s h than the E n g l i s h of 13.5. I suggest that the f o l l o w i n g s e r i e s of o r d i n a r y E n g l i s h sentences would do to express the thought of 13.4: 13.6 Some man came e a r l y . Edna may have t a l k e d to that man. Someone may have persuaded Edna. Therefore Edna may have t a l k e d to that man. "Did Tom persuade Edna?" John spoke about that question. "John's speech was t a c t f u l . " Mary argued the preceding p r o p o s i t i o n . D. S e n t e n t i a l Co-ordinations. In t h i s s e c t i o n I want to demonstrate that s e n t e n t i a l co-ordinations are e l i m i n a b l e i n paraphrase. S e n t e n t i a l co-ordina-t i o n s are e a s i e r to paraphrase than s e n t e n t i a l emdeddings. A s e n t e n t i a l l y embedded s t r u c t u r e such as a noun phrase comple-ment stands as subject or object of a verb; another type of s e n t e n t i a l l y embedded s t r u c t u r e , the r e l a t i v e clause, i n some 184 ways r e l a t e s to the noun phrase to which i t i s attached the way an a d j e c t i v e r e l a t e s to a noun. (This i s not s u r p r i s i n g , s i n c e a d j e c t i v e s are g e n e r a l l y held to come from r e l a t i v e clause s t r u c t u r e s . For a treatment of t h i s p o int, see J & R pages 204 to 206.) But s e n t e n t i a l l y co-ordinated s t r u c t u r e s , character-i s t i c a l l y , are rather independent from one another, and t h i s make the s e p a r a t i o n of s e n t e n t i a l co-ordinations i n t o s e p a r a t e sent-ences, i n most cases, a r e l a t i v e l y p a i n l e s s operation. Consider, f o r example, a sentence containing two co-or-dinated s e n t e n t i a l clauses l i k e : 14.1 When the sun rose, the sky was c l o u d l e s s . This can be paraphrased by: 14.2 The sun rose at some time. At that time t h e s k y was c l o u d l e s s . S i m i l a r l y , c o n s i d e r : 14.3 Before she could w r i t e the note, she had t o make a telephone c a l l . This might be paraphrased by: 14.4 F i r s t she had to make a telephone c a l l . Then s h e could w r i t e the note. (I am assuming here that 'had to' i s the past tense o f the modal 'must'; i f t h i s assumption i s i n c o r r e c t , then the f i r s t sentence of 14.4 would be broken down fu r t h e r along the l i n e s i n d i c a t e d i n the previous sections.) When two sentences are joined by 'and', they can u s u a l l y be separated without s u b s t a n t i a l changes. For example: 185 14.5 Mary went to the sto r e and John went to the park, can be paraphrased with: 14.6 Mary went to the s t o r e . John went to the park. I f there i s an i m p l i c a t i o n of s i m u l t a n e i t y i n 14.5 which i s missing i n 14.6, then i t can be made e x p l i c i t without d i f f i c u l t y . Thus in s t e a d of 14.6, we would have: 14.7 Mary went to the s t o r e . At the same time, John went to the park. I f the example chosen had been a l i t t l e d i f f e r e n t , the i m p l i c a t i o n with regard to time might a l s o have been d i f f e r e n t . Thus: 14.8 Mary went to the store, and she bought some f r u i t , implies sequence r a t h e r than simul t a n e i t y , i f i t implies anything. Sequence, of course, would be no more d i f f i c u l t than s i m u l t a -n e i t y to represent e x p l i c i t l y : 14.9 Mary went to the s t o r e . Then she bought some f r u i t . What about ' i f — then' co-ordinations? These are c o n s i -derably more ch a l l e n g i n g to paraphrase. Consider: 14.10 I f i t ' s r a i n i n g then i t ' s cloudy. And consider a l s o as a f i r s t attempt at a paraphrase, the f o l l o w i n g : 14.11 Sometimes i t ' s r a i n i n g . On those occasions, i t ' s cloudy. I think i t would be j u s t i f i a b l y objected here that 14.11 does not capture the i n f e r e n t i a l connexion which the ' i f — then' co-ordination expresses. Consider, then, as a second attempt at a paraphrase: 186 14.12 " I t ' s r a i n i n g . " " I t ' s cloudy." The f i r s t p r o p o s i t i o n implies the second p r o p o s i t i o n . Now i f there were to be a general o b j e c t i o n to t h i s s t y l e of paraphrase i t would be d i f f e r e n t from the o b j e c t i o n given to 14.11. I t would be that the r e l a t i o n s h i p drawn i n 14.10 i s a r e l a t i o n s h i p between two s i t u a t i o n s or between two s t a t e s of a f f a i r s , but i t i s not a r e l a t i o n s h i p between two p r o p o s i t i o n s . Once again, then, we are drawn i n t o using the terminology of ' f a c t s ' 'events' ' s i t u a t i o n s ' 'states of a f f a i r s ' 'propositions' and so on; to remedy the f a u l t of 14.12, we would be l i k e l y to t r y something l i k e : 14.13 " I t ' s r a i n i n g . " The preceding p r o p o s i t i o n describes a s t a t e of a f f a i r s • C a l l t h i s s t a t e of a f f a i r s 'A'. "I t ' s cloudy." The preceding p r o p o s i t i o n describes a s t a t e of a f f a i r s . C a l l t h i s s t a t e of a f f a i r s 'B'. State of a f f a i r s B i s a consequence of s t a t e of a f f a i r s A. Now t h i s v e r s i o n , although i t invokes the somewhat unwelcome terminology, i s , I b e l i e v e , a more accurate r e p r e s e n t a t i o n of the meaning of the o r i g i n a l , 14.10, than 14.12. But i t i s a l s o p o s s i b l e , I think, to give a p r e c i s e paraphrase f o r 14.10 without the cumbersome paraphanalia of 14.13. Instead of up h o l s t e r i n g 14.12 one might upholster 14.11: 14.11 Sometimes i t ' s r a i n i n g . On those occasions, i t ' s cloudy. and i n the f o l l o w i n g way: 14.14 Sometimes i t ' s r a i n i n g . On those occasions i t must be cloudy. 187 And i f to t h i s i t were to be objected that the modal 'must' expresses a d i f f e r e n t r e l a t i o n s h i p than that expressed by the ' i f - t h e n ' connective, or that i t would be i l l e g i t i m a t e to assume that the modal could take over a l l the expressive functions of the ' i f -- then' connectives without any consequent o b l i t e r a t i o n of semantic d i s t i n c t i o n s , then one could argue f o r the c r e a t i o n of a modified 'must' modal, say 'musti' which would be used wherever we now use the ' i f — then' connectives, and which would thus preserve the semantic d i s t i n c t i o n s , i f any, which e x i s t . We would then paraphrase 14.10 with: 14.15 Sometimes i t ' s r a i n i n g . On those occasions i t mustj_ be cloudy. The'musti' of 14.15 would be a modal reserved f o r the paraphrase of ' i f — then' c o - o r d i n a t i o n s . Thus, f o r ' i f -- then' co-ordinations there are two a v a i l a b l e s t y l e s of paraphrase, one represented by 14.15 and the other by 14.13. In a d d i t i o n to these two s t y l e s of para-phrase, there i s a t h i r d one p o s s i b l e which i s very u s e f u l , p a r t i c u l a r l y f o r the d i s e n t a n g l i n g of m u l t i p l e ' i f -- then' c o n s t r u c t i o n s . Consider: 14.16 Suppose i t ' s r a i n i n g . Then i t w i l l be cloudy. I b e l i e v e that 14.16 i s a paraphrase of the o r i g i n a l , 14.10 i n the sense that i t i s synonymous with i t ; but i t i s not a paraphrase i n our s p e c i a l sense of a paraphrase, i . e . i t i s not a paraphrase which contains no r e c u r s i v e l y derived s t r u c t u r e s . The f i r s t sentence of 14.16 i s a transformation of: 188 14.17 Suppose that i t ' s r a i n i n g . which contains the noun phrase complement 'that i t ' s r a i n i n g ' . One approach here would be to break 14.17 down i n t o the sentences 14.18 " I t ' s r a i n i n g . " C a l l the preceding p r o p o s i t i o n 'A'. Suppose p r o p o s i t i o n A to be t r u e . But t h i s approach ignores an i n s t r u c t i v e assymetry present i n 14.6. Although the ' i f — then' s t r u c t u r e would appear to be symmetrical with regard to the grammar of the two clauses (no l o g i c a l symmetry intended here), paraphrasing the " i f — then c o - o r d i n a t i o n of 14.10 r e s u l t e d i n two sentences, one of which contained a noun phrase complement, the second of which d i d not. This r a i s e s the p o s s i b i l i t y of t r y i n g to construct two sentences roughly along the l i n e s of 14.16 i n such a way that n e i t h e r of them would contain a r e c u r s i v e l y derived s t r u c t u r e . Imagine, then, that 'suppose' i n : 14.16 Suppose i t ' s r a i n i n g . Then i t w i l l be cloudy, were to r e l a t e i n grammatical terms to ' i t ' s r a i n i n g 9 the way 'then' r e l a t e s to ' i t w i l l be cloudy'. Since the transform-a t i o n a l a n a l y s i s of the 'then* i s not c l e a r , i t i s not easy to s p e l l out how 'suppose' could be construed as f i l l i n g a p a r a l l e l grammatical r o l e i n i t s sentence. Consider, then, the f o l l o w i n g attempt at a paraphrase f o r 14.10: 14.10a i t hypothecate be r a i n i n g . I t consequence be cloudy. The form of these sentences follows the form of sentences l i k e : 14.19a I t must be r a i n i n g . 14.19b I t may be r a i n i n g . 189 Thus'hypothecate' and 'consequence' would be modals, and ' i t hypothecate be r a i n i n g ' would mean something l i k e "Suppose i t ' s r a i n i n g ' while ' i t consequence be cloudy' would mean something l i k e 'then i t i s cloudy' or'consequently i t i s cloudy' or 'there-fore i t i s cloudy'. Let's look at t h i s s t y l e of paraphrase f o r ' i f — then' co-ordinations using another example: 14.20 I f John came home e a r l y then he didn't stop at the pub. This example i s more t y p i c a l i n that i t does not have ' i t ' as the subject of the cl a u s e s . Using the modal paraphrase, we would have. 14.21. John hypothecate came home e a r l y . John consequence di d n ' t stop at the pub. There i s s t i l l an unresolved problem of tense here s i n c e i t i s unclear whether the modals should be i n present or past tense, that i s , i t i s unclear i n what way the modals, or whether the modals, should be subject to laws of agreement with the verbs that f o l l o w them. Since modals l i k e 'may' and 'must' are subject to such laws of agreement, and con s i d e r i n g how, i f the modal had been 'must' or 'can', the second sentence would have read: 14.22a John mustn't have stopped at the pub. or: 14.22b John couldn't have stopped at the pub. i t would seem reasonable to pre f e r the fol l o w i n g v e r s i o n to 14.21: 14.23 John hypothecate have come home e a r l y . John consequence not have stopped at the pub. The novelty of the modal paraphrase should not be taken to 1 9 0 count against i t i n any way. My argument for the l e g i t i m a c y of the modal paraphrase, b r i e f l y r e s t a t e d , runs as f o l l o w s : the s t r u c t u r e of antecedent and consequent (although l o g i c a l l y d i r e c t i o n a l ) i s grammatically symmetrical. We can c l e a r l y express consequence by using a modal, as i n "John must have stopped at the pub', where 'must' might express consequence, and so there is no reason why we could not express antecedence by using a modal. Since we want to preserve a l l p o s s i b l e d i s t i n c t i o n s which now e x i s t we don't n e c e s s a r i l y want to use 'must' f o r the consequence; and since there i s no present modal expressing antecedence, we must invent one. I have chosen, rather a r b i t r a r i l y from a number of p o s s i b i l i t i e s , to borrow on the verb 'hypothecate' and the noun 'consequence" to represent these modals. I t remains to be shown that m u l t i p l y co-ordinated ' i f —-then' s t r u c t u r e s are eliminable i n paraphrase. Consider: 14.24 I f , i f Mary l i k e s John, then John l i k e s Mary, then, i f John l i k e s Mary, then Mary l i k e s John. This sentence i s of the form: 14.25 ((p =>q) ? (q ?»p) ) and i t can be seen that there i s an important d i f f e r e n c e between s i n g l e ' i f — then' sentences and sentences with m u l t i p l e ' i f — then' c o n s t r u c t i o n s . Sentences with a simple ' i f — then' s t r u c t u r e w i l l be s p e l l i n g out some r e l a t i o n s h i p between events of any s o r t , or between s i t u a t i o n s , or states of a f f a i r s of any s o r t . But only the smallest " i f — then' units i n a s t r u c t u r e 191 with m u l t i p l e ' i f — then' constructions can r e l a t e events or states of a f f a i r s of any kind; the only kind of things that the l a r g e r ' i f — then' s t r u c t u r e s can r e l a t e are i m p l i c a t i o n s . Thus, i n 14.25, the l a r g e r " i f — then' s t r u c t u r e r e l a t e s the i m p l i c a t i o n ' p — > q' to the i m p l i c a t i o n ' q > p ' . And to paraphrase 14.24, we can use something l i k e : 14.26 "Mary hypothecate l i k e s John." "John consequence l i k e s Mary." The preceding i m p l i c a t i o n hypothecate holds. The f o l l o w i n g i m p l i c a t i o n consequence holds . "John hypothecate l i k e s Mary." "Mary consequence l i k e s John." Perhaps someone might object that t h i s paraphrase makes reference to an i m p l i c a t i o n which i t s e l f r e f e r s to two separate sentences, or two separate p r o p o s i t i o n s . But i t i s no more o b j e c t i o n a b l e to l e t two separate sentences or p r o p o s i t i o n s c o n s t i t u t e an i m p l i c a t i o n than i t i s to l e t two separate people, say, c o n s t i t u t e a couple. So f a r I have argued f o r two s t y l e s of paraphrase f o r ' i f — then' c o - o r d i n a t i o n s . One of these uses references to s i t u a t i o n s and states of a f f a i r s , , the other uses new modals. I t i s p o s s i b l e , however, to paraphrase m u l t i p l e ' i f — then' co-ordinations using a maximum of one ' i f — then' c o - o r d i n a t i o n per sentence and therefore using neither cumbersome para-phenalia nor new (and hence suspicious?) modals. •If then' co-ordinations used only once per sentence 1 9 2 can be introduced using base ru l e s which permit non-recursive s e n t e n t i a l c o - o r d i n a t i o n i n the same way that embedded clauses can be introduced a maximum of once per sentence using base r u l e s which permit non-recursive s e n t e n t i a l embedding. The f o l l o w i n g set of base r u l e s , f o r example, would permit the c o n s t r u c t i o n of a s i n g l e ' i f — then' c o - o r d i n a t i o n but would not permit the c o n s t r u c t i o n of m u l t i p l y co-ordinated or embedded ' i f — then 1 c o - o r d i n a t i o n s : 14.27a S1 > I f + S + then + S b S ^ Noun phrase + Verb phrase (Here 'S' stands f o r 'Sentence'). The presence of the s u b s c r i p t on the ' S' i n the l e f t hand s i d e of 14.27a would b l o c k these r u l e s from being used r e c u r s i v e l y , so that only one ' i f -- then* c o - o r d i n a t i o n per sentence could be generated by them. That we can disentangle m u l t i p l y co-ordinated ' i f -- then' constructions using a maximum of one per sentence can be best seen from an example l i k e 14.24; 14.24 I f , i f Mary l i k e s John, then John l i k e s Mary, then, i f John l i k e s Mary, then, Mary l i k e s John. The key poi n t f o r t h i s a n a l y s i s has already been mentioned: only the innermost ' i f —• then' co-ordinations can r e l a t e any k i n d of s i t u a t i o n s or states of a f f a i r s ; a l l the others r e l a t e i m p l i c a t i o n s . Thus 14.24 can be paraphrased using a maximum of one ' i f — then' co-ordination per sentence by: 14.24b " I f Mary l i k e s John, then John l i k e s Mary." 193 C a l l the previous i m p l i c a t i o n 'A'. "I f John l i k e s Mary, then Mary l i k e s John." C a l l the previous i m p l i c a t i o n 'B'. I f i m p l i c a t i o n A holds, then i m p l i c a t i o n B h o l d s . This s t y l e of paraphrase, as i t does not involve references to c h a r a c t e r i z e r s l i k e ' s i t u a t i o n s ' and 'states of a f f a i r s ' , and as i t does not r e s o r t to the invention of new modals, i s perhaps the best of the three presented. I conclude that we can express a l l our thoughts without using r e c u r s i v e l y derived ' i f — then' c o - o r d i n a t i o n s . The s t o r y f o r 'either - or" co-ordinations, roughly speaking, runs p a r a l l e l t o that for ' i f — then' c o - o r d i n a t i o n s . The t h i r d and best technique developed f o r ' i f — then' co-ordina-t i o n s a p p l i e s e q u a l l y w e l l f o r 'either — or' c o - o r d i n a t i o n s . Thus consider the f o l l o w i n g sentence with m u l t i p l e 'either -- or' c o n s t r u c t i o n s : 14.28 E i t h e r , e i t h e r Johnson w i l l be there, or h i s proxy w i l l be there, or, e i t h e r Johnson's proxy w i l l be there, or Johnson's second proxy w i l l be there. Sentences such as 14.28 and m u l t i p l y co-ordinated ' e i t h e r — o r ' constructions i n general sound l i k e they come from l o g i c i a n s ' notebooks rather than from ro u t i n e conversations, but s i n c e they are meaningful and grammatical they are important specimens to consider. Now using a maximum of one 'either — or' co - o r d i n a t i o n per sentence we can paraphrase 14.28 by a s e r i e s of sentences l i k e : 194 14.29 " E i t h e r Johnson w i l l be there or h i s proxy w i l l be there. " C a l l the preceding d i s j u n c t i o n 'A'. "Either Johnson's proxy w i l l be there or Johnson's second proxy w i l l be there." C a l l the preceding d i s j u n c t i o n 'B'. E i t h e r d i s j u n c t i o n A holds or d i s j u n c t i o n B holds. I t i s a t e l l i n g f a c t that one would probably use something l i k e 14.29 to e x p l i c a t e 14.28. One might be a l s o tempted to argue f o r the e l i m i n a t i o n of a l l ' either - or' co-ordinations by, say, i n v e n t i n g a verb l i k e 'to or' so that " S i t u a t i o n A ors s i t u a t i o n B" would mean something l i k e " E i t h e r s i t u a t i o n A obtains or s i t u a t i o n B o b t a i n s . " This step i s not necessary, however, and I w i l l pursue i t no f u r t h e r . Now i n a d d i t i o n to the s e n t e n t i a l co-ordinators so f a r discussed there are co-ordinators l i k e 'but' 'whereas' 'not-withstanding' and 'because'. Some of these can be t r e a t e d by simply s e p a r a t i n g the clauses. Thus, examine the f o l l o w i n g p a i r s of cases: 14.30a John l e f t , but Mary remained, b John l e f t . But Mary remained. 14.31a John l e f t , whereas Mary remained. b John l e f t . However, Mary remained. 14.3 2a Notwithstanding that John l e f t , Mary remained, b John l e f t . Despite that, Mary remained. 'Because', however, i s more d i f f i c u l t than the others. Consider: 14.33 John l e f t , because Mary remained. We might t r y two avenues of approach here. F i r s t : 195 14.34 Mary remained. Because of t h i s , John l e f t . And second: 14.35 Mary remained. Therefore, John l e f t . I t ' s a l s o worth noting that these co-ordinators are not c l e a r l y r e c u r s i v e devices i n any case. Thus sentences l i k e the f o l l -owing are b i z a r r e i f not ungrammatical: 14.36 John l e f t , but Mary remained, but Edna a r r i v e d . 14.37 Notwithstanding that John l e f t , Mary remained, notwithstanding that Edna a r r i v e d . 14.38 Because Mary remained, John l e f t , because Edna a r r i v e d , because Tim telephoned. Thus, use of such co-ordinators to create m u l t i p l y co-ordinated s t r u c t u r e s i s not c h a r a c t e r i s t i c ; c e r t a i n l y being able to construct sentences l i k e 14.3 6 to 14.38 does not enhance the expressive c a p a c i t y of our language, and although such s t r u c t u r e s c h a r a c t e r i s t i c a l l y appear i n the form of 'mixed co-ordinations', these mixed co-ordinations do not appear to present s p e c i a l problems. To summarize t h i s s e c t i o n , I cla i m that r e c u r s i v e l y derived clause co-ordinations such as ' i f — then', 'either - or* 'and' and 'when' (or adverbial) clause co-ordinations can be eliminated i n paraphrase. The most i n t e r e s t i n g and cha l l e n g i n g problems surround the ' i f — then' and 'either - or' co-o r d i n a t i o n s . I have argued that, using a maximum of one such clause per sentence, m u l t i p l y co-ordinated ' i f — then' and 'either - or' constructions can be disentangled. i n general, we do not need r e c u r s i v e l y derived co-ordinations to express our thoughts. 196 E. Non-sentential r e c u r s i v e l y derived s t r u c t u r e s . In a d d i t i o n to the kinds of r e c u r s i v e l y derived s t r u c t u r e s which have so f a r been considered, there are other types of st r u c t u r e s whose o r i g i n s i n deep s t r u c t u r e are s t i l l obscure, but which at some l e v e l , i n a l l p r o b a b i l i t y , involve the r e c u r s i v e r u l e s of the base. According to some authors, i t i s not necessary — even to s a t i s f y the requirements of d e s c r i p t i v e grammar — to p o s t u l a t e r e c u r s i v e base r u l e s other than those which c o n t a i n the 'S' or 'sentence' category on the r i g h t hand s i d e . Thus, f o r example, Chomsky says, (Aspects,. page 137); Now the r e c u r s i v e property i s a feature of the base component, i n p a r t i c u l a r , of the r u l e s that introduce the i n i t i a l symbol S i n designated p o s i t i o n s i n s t r i n g s of category symbols. There are, apparently,, no other r e c u r s i v e r u l e s i n the base. And f u r t h e r , i n a footnote to t h i s paragraph which makes the po i n t even more e x p l i c i t l y (Aspects, footnotes. C h a p t e r 3 #11): ...In other words, the only way to form new deep s t r u c t u r e s i s t o i n s e r t elementary " p r o p o s i t i o n s " -- t e c h n i c a l l y , base Phrase-markers — i n other Phrase-markers. This i s by no means a l o g i c a l l y necessary feature of phrase s t r u c t u r e grammars. Notice that the schemata that u n d e r l i e c o - o r d i n a t i o n a l s o provide i n f i n i t e generative capacity, but here too the true r e c u r s i v e property can apparently be l i m i t e d to the schema S—> S#S# ...#S, hence to rul e s i n t r o d u c i n g " p r o p o s i t i o n s . " This formulation leaves unexplained some rather marginal phenomena (e.g. the source of such expressions as "very, very very A d j e c t i v e " and some more s i g n i f i c a n t ones (e.g., the p o s s i b i l i t y of i t e r a t i n g Adverbials and various kinds of parenthetic elements, the status of which i n general i s unclear.) From the p o i n t of view of t h i s a n a l y s i s we do not need to decide whether or not these non-sentential r e c u r s i v e l y derived 197 s t r u c t u r e s can be accounted for i n terms of r e c u r s i v e r u l e s which introduce the sentence category; I think i t can be shown, q u i t e independently of t h i s question, that these compound s t r u -ctures are not necessary f o r expression; i . e . , that sentences which c o n t a i n them can be paraphrased by sentences a l l of which can be generated by a grammar which contains no r e c u r s i v e base r u l e s . Coordinated noun phrases. I t ' s p o s s i b l e to co-ordinate noun phrases i n d e f i n i t e l y . Thus, i n : 15.1 Mary and John and B i l l . . . w e n t home together. one could f i l l i n the dots with any number of noun phrases, i n p a r t i c u l a r , with any number of proper names of people. I t i s not d i f f i c u l t , though, to separate the noun phrases. Thus, we could paraphrase: 15.2 Mary and John and B i l l and Edna went home together. with: 15.3 A group of people went home together. Mary was i n the group. John was i n the group. B i l l was i n the group. Edna was i n the group. (No one e l s e was i n the group.) (The sentence i n the parentheses is of course o p t i o n a l , i t s i n c l u s i o n depending on how much content one reads i n t o a sentence l i k e 15.2). I t i s not c l e a r to what extent a procedure f o r paraphrasing 198 co-ordinated noun phrases can be formalized, but roughly speaking, i t could be s a i d that noun phrase co-ordinations can be e l i m i n a t e d by c l a s s i f y i n g the objects being co-ordinated and then i n d i v i d u a l l y i n t r o d u c i n g the noun phrases i n separate sentences, with some s p e l l i n g out of a membership-tie between the c l a s s i f i e r and the noun phrase. To i l l u s t r a t e with a second example: 15.4 I f you saw John and Mary and Edna and B i l l then I could not have been there. 15.5 John i s a member of group A. Mary i s a member of group A. Edna i s a member of group A. B i l l i s a member of group A. I f you saw these members of group A, I could not have been there. And to i l l u s t r a t e with a t h i r d example, one i n which the co-ordinated noun phrases themselves contain embedded clajses: 15.6 The man who wore red and the woman who wore blue were married long ago i n England. 15.7 Some man wore red. Some woman wore b l u e . These people were married long ago i n England. Nested Possessives. Possessives e a s i l y lend themselves to the c r e a t i o n of m u l t i p l y nested s t r u c t u r e s . Perhaps a s i n g l e instance w i l l s u f f i c e to show the e l i m i n a b i l i t y of m u l t i p l y nested possessives 15.8 That man's f r i e n d ' s uncle's s i s t e r ' s dog's c o l l a r ' s r i n g i s b r i g h t . or: 1 5 . 9 The r i n g of the c o l l a r of the dog of the s i s t e r of the uncle of the f r i e n d of that man i s b r i g h t . To paraphrase such sentences one could use the following kind of approach: 199 15.10 That man has a f r i e n d . That f r i e n d has an uncle. That uncle has a s i s t e r . That s i s t e r (has)(owns) a dog. That dog (has) (wears) a c o l l a r . That c o l l a r has a r i n g (on i t ) . That r i n g i s b r i g h t . Many i n t e r e s t i n g problems come up i n such paraphrases, p a r t i c u l a r l y surrounding the choice of verbs to f i l l i n the r e l a t i o n s h i p between the two noun phrases which i n the o r i g i n a l are connected only by the rather contentless p o s s e s s i v e . But the b a s i c p o i n t that some verb w i l l f i l l i n the rather un-s p e c i f i e d content of the possessive is not open to question. I f no other verb w i l l do, the verb 'have' w i l l . An i n t e r e s t i n g f eature of t h i s paraphrase a n a l y s i s , by the way, i s that i t suggests that the possessive i s i t s e l f unnecessary as a s y n t a c t i c s t r u c t u r e . Compounded a d j e c t i v e s . There i s no l i m i t to the number of a d j e c t i v e s which can precede a noun. Thus for example, the dots i n the f o l l o w i n g sentence can be f i l l e d i n by any number of a d j e c t i v e s : 15.11 He i s t a l l , t h i n , dark-eyed, bald, Canadian,..., twenty-nine years old, and i n t e r e s t e d i n chess. A c t u a l l y , some q u a l i f i c a t i o n of t h i s point i s necessary. In a l l the previous examples of r e c u r s i v e l y derived s t r u c t u r e s , the number of p o s s i b l e occurences of the s t r u c t u r e i n a given sentence was t r u l y u n l i m i t e d . Thus with a f i n i t e , and even very l i m i t e d vocabulary (say four words) an u n l i m i t e d number of sentences can be created using r e l a t i v e clauses. For example, 200 with only the four words, 'the', 'man', 'who', and 'saw', one can create an i n f i n i t e number of sentences of which the fo l l o w i n g would be a t y p i c a l instance: 15.12 The man who saw the man who saw the man who saw the man who saw the man who saw the man who saw the man who saw the man saw the man. But using only a f i n i t e number of a d j e c t i v e s i t i s not c l e a r whe-ther one can create an i n f i n i t e number of sentences. Thus i t i s not c l e a r , f o r example,whether the f o l l o w i n g sentence i s grammat 15.13 He i s t a l l , t h i n , dark-eyed, dark-eyed, bald, Canadian, Canadian, dark-eyed, t h i n , t a l l , twenty-nine years o l d , dark-eyed, Canadian, Canadian, and i n t e r e s t e d i n chess. But i f there are only a f i n i t e number of a d j e c t i v e s i n E n g l i s h , which i s not an unreasonable assumption, then a d j e c t i v e s would have to be repeated i n order f o r the grammar to be able to generate an i n f i n i t e number of sentences based on using a d j e c t i v e s i n sequr<-However, even i f we do consider a d j e c t i v e s used i n sequence to be r e c u r s i v e l y derived s t r u c t u r e s , i t w i l l not be d i f f i c u l t to see that t h i s use of a d j e c t i v e s i s eliminable i n paraphrase. Thus we could paraphrase 15.11 by: 15.14 He i s t a l l . He i s t h i n : He i s dark-eyed. He i s ba . He i s Canadian. He i s . . . He i s twenty-nine years o l d . He i s i n t e r e s t e d i n chess. Or, using an example where the ad j e c t i v e s precede a noun: 15.15 The t a l l , t h i n , supple-limbed, nimble, a g i l e clown had a sad face. 15.16 Some clown i s t a l l . That clown i s t h i n . That clowr i s supple-limbed. That clown i s nimble. That clowr i s a g i l e . That clown has a sad face. 201 Compounded a d j e c t i v e s present a more complicated problem, however, when there are q u a n t i f i e r s present i n the sentence. Consider, f o r example: 15.17 Few people are t a l l , t h i n , dark-eyed, b a l d , Canadian, twenty-nine years o l d , and i n t e r e s t e d i n chess. We could not use the q u a n t i f i e r d i r e c t l y i n each sentence,or e l s e we would have something l i k e the f o l l o w i n g : 15.18 Few people are t a l l . Few people are t h i n . Few people are dark-eyed. Few people are b a l d . Few people are Canadian. Few people are twenty-nine years o l d . And few people are i n t e r e s t e d i n chess. Of course, 15.18 i s not synonymous with 15.17; 15.18, f o r instance, as s e r t s that few people are twenty-nine years o l d , which i s not i n any way an i m p l i c a t i o n of 15.17. However we could paraphrase 15.17 e i t h e r by using c l a s s i f i e r s and r e f e r r i n g to membership, or a l t e r n a t i v e l y by changing the q u a n t i f i e r f o r the i n d i v i d u a l p r e d i c a t e s and then c o l l e c t i n g the predicates together under the o r i g i n a l q u a n t i f i e r . The f o l l o w i n g two paraphrases, r e s p e c t i v e l y , i l l u s t r a t e these two approaches: 15.19 The f o l l o w i n g l i s t defines a c l a s s of people, " ' t a l l ' ' t h i n ' 'dark-eyed' 'bald' 'Canadian' 'twenty-nine years o l d ' ' i n t e r e s t e d i n chess'". This c l a s s has few members. 15.20 Some people are t a l l . Some people are t h i n . Some people are dark-eyed. Some people are b a l d . Some people are Cana.'ian. Sortie people are twenty-nine years o l d . Some people are i n t e r e s t e d i n chess. Few people are a l l of these t h i n g s . The second approach i s p r e f e r a b l e , I b e l i e v e , as i t involves no 202 m e t a t h e o r e t i c a l i n t r o d u c t i o n s ; i . e . , i t does not mention words which r e f e r to p r e d i c a t e s ; rather i t uses the predicates themselves. Paraphrases such as 15.19 are subject to such problems as are set f o r t h by Church i n On Carnap's A n a l y s i s of A s s e r t i o n and B e l i e f ; and i f one accepts Church's argument one may want to r e j e c t 15.19 as a paraphrase of 15.17. 15.20 on the other hand, might be c r i t i c i s e d on the grounds that the o r i g i n a l does not presuppose that some people are t a l l ; i t presupposes, r a t h e r that some or no people are t a l l , and ac c o r d i n g l y one might want to include t h i s i n the paraphrase v e r s i o n , s t a t i n g "Some or no people are t a l l . Some or no people are t h i n " e t c . A l t e r n a t i v e l y , one might r e p l y that the c r i t i c i s m i s not sound on the grounds that 15.17 does pre-suppose t h a t some people are t a l l , t h i n , dark-eyed, bald, Canadian, twenty-nine years o l d , and i n t e r e s t e d i n chess, other-wise the q u a n t i f i e r 'few' could not be a p p l i e d . And i f some people are a l l of those things, then i t i s presupposed that some people are t a l l , some people are t h i n , some people are dark-eyed, and so on. Whatever one's approach to the meaning of the q u a n t i f i e r 'few', paraphrases f o r asser t i o n s such as 15.17 can be constructed. I conclude, then, that a d j e c t i v e s used s e q u e n t i a l l y are elim i n a b l e i n paraphrase, and that there are s e v e r a l s t y l e s of paraphrase which can be used i n t h i s regard. 203 A d v e r b i a l sequences. F i n a l l y , there are a d v e r b i a l sequences which very much resemble a d j e c t i v a l sequences, from the point of view of paraphrase e l i m i n a t i o n . A d v e r b i a l sequences at some point w i l l begin to repeat, there being only a f i n i t e number of adverbs a v a i l a b l e f o r any sequence, and i t i s unclear as to whether a d v e r b i a l sequences with repeated elements are to be considered grammatical. Thus: 15.21 He walked s w i f t l y , determinedly, c o n f i d e n t l y , s w i f t l y , and determinedly. i s a b i z a r r e i f not ungrammatical sentence. In any case, a d v e r b i a l sequences are e a s i l y e l i m i n a b l e i n paraphrase, -- more e a s i l y , i n f a c t , than a d j e c t i v a l sequences. Thus: 15.22 He walked s w i f t l y , determinedly, c o n f i d e n t l y , and c h e e r f u l l y . can be paraphrased by: 15.23 He walked s w i f t l y . He walked determinedly. He walked c o n f i d e n t l y . And he walked c h e e r f u l l y . I claim, then, that a d v e r b i a l sequences are eli m i n a b l e i n paraphrase. F. F i n a l Remarks on Chapter Two. This chapter has been perhaps too long, but I chose not to break i t down in t o s e v e r a l chapters because I have only been t r y i n g to demonstrate a s i n g l e point, and a simple one at th a t . 204 The p o i n t i s that having r e c u r s i v e l y derived sentences i n our language i s not necessary f o r us to be able to say whatever we might want to say. We can express a l l of our thoughts using a language whose grammar generates only a f i n i t e number of deep s t r u c t u r e types f o r sentences. Unfortunately, the reader w i l l probably agree that the s i m p l i c i t y of the p o i n t has not been matched by s i m p l i c i t y of demonstration. Instead of p o i n t i n g out i n a general way why r e c u r s i v e l y d e r ived sentences are not necessary f o r thought expression, I have, ad hoc, made a l i s t of r e c u r s i v e l y derived c o n s t i t u e n t s and examined a sampling of cases belonging to these kinds of r e c u r s i v e l y derived s t r u c t u r e s . The question then a r i s e s , can there be a general explanation f o r why r e c u r s i v e l y d e r ived sentences are not necessary f o r expression — or b e t t e r s t i l l — a general demonstration that r e c u r s i v e l y derived sentences are not necessary f o r expression? I f the b a s i c moves made i n t h i s chapter are c o r r e c t , then i t w i l l be u n l i k e l y that there i s a s i n g l e general explanation f o r , or demonstration of, the e l i m i n a b i l i t y of r e c u r s i v e l y derived sentences i n paraphrase. The reason f o r t h i s i s that the expressive functions served by the d i f f e r e n t kinds of r e c u r s i v e l y derived constituents are " d i f f e r e n t . The c h a r a c t e r i s t f u n c t i o n of a r e s t r i c t i v e r e l a t i v e clause, f o r example, i s to in d i v i d u a t e , and so decid i n g whether we can eliminate r e s t r i c t i v e 205 r e l a t i v e clauses i n paraphrase amounts to d e c i d i n g whether we can i n d i v i d u a t e i n the way that r e s t r i c t i v e r e l a t i v e clauses enable us to i n d i v i d u a t e , without using r e s t r i c t i v e r e l a t i v e clauses (or other r e c u r s i v e l y derived constituents.) But not a l l r e l a t i v e clauses i n d i v i d u a t e i n the same way and other r e c u r s i v e l y d e r i v e d c o n s t i t u e n t s , say ' i f — then' s e n t e n t i a l c o - o r d i n a t i o n s , do not have any p a r t i c u l a r r e l a t i o n s h i p with the task of i n d i v i d u a t i o n . In f a c t , although we can say what i t i s t hat r e l a t i v e clauses c h a r a c t e r i s t i c a l l y do, we would hav< a good deal of t r o u b l e saying j u s t what i t i s that, say, verb phrase complements c h a r a c t e r i s t i c a l l y do, or j u s t what i t i s that 'whether' noun phrase complements c h a r a c t e r i s t i c a l l y do. The upshot of t h i s i s that the only way to f i n d out whether r e c u r s i v e l y d e r ived sentences are eliminable i n paraphrase i s to go through a l l the d i f f e r e n t kinds of r e c u r s i v e l y derived c o n s t i t u e n t s , examining re p r e s e n t a t i v e samples to discover what t h e i r c h a r a c t e r i s t i c functions are, and whether these functions can be undertaken with a language whose grammar does not contain r e c u r s i v e l y derived c o n s t i t u e n t s . In a word, since there i s no s i n g l e expressive f u n c t i o n undertaken by r e c u r s i v e l y derived c o n s t i t u e n t s , there i s u n l i k e l y to be a general explanation of why, or demonstration that, r e c u r s i v e l y derived sentences are eliminable i n paraphrase. This may be somewhat misleading, though, unless two further issues are at l e a s t b r i e f l y discussed. One suggests 206 that, given c e r t a i n assumptions a b o u t semantics, a general demonstration of the e l i m i n a b i l i t y o f recurs i v e l y derived sentences might be p o s s i b l e , but that these a s s u m p t i o n s are i n one sense, what i s at issue i n checking the s t a t e m e n t t h a t r e c u r s i v e generation of deep s t r u c t u r e types f o r s e n t e n c e s i s necessary f o r thought expression. The s e c o n d d i s t i n g u i s h e s between a general demonstration o f the n o n - n e c e s s i t y o f r e c u r s i v e generation of deep s t r u c t u r e types f o r s e n t e n c e s f o r thought expression on the one hand and a general d e m o n s t r a t i o n of the e l i m i n a b i l i t y of r e c u r s i v e l y derived s e n t e n c e s i n p a r a -phrase on the other. I f one assumes that the semantic c o n t e n t o f a complex sentence i s a f u n c t i o n of the semantic c o n t e n t c 1: i t s elementary p r o p o s i t i o n s and t h e i r s t r u c t u r a l r e l a t i o n s , t h e n i t may be p o s s i b l e to demonstrate that a l l r e c u r s i v e l y d e r i v e d s e n t -ences are paraphraseable by a sequence o f s e n t e n c e s none o f which i s r e c u r s i v e l y d erived. (The nature o f t h e d e m o n s t r a t i o n — and i n some respects i t ' s no more a ' u n i f i e d ' o r ' g e n e r a l ' one than that of Chapter Two —• i s discussed b e l o w , s e e page 240., However, i f we are t r u l y to c h e c k the statement o f n e c e s s i t y , that r e c u r s i v e generation of deep s t r u c t u r e types f o r sentences i s necessary f o r thought expression,' then the assumption that the semantic content of a complex p r o p o s i t i o n i s a f u n c t i o n o f the semantic content of i t s elementary p r o p o s i t i o n s and t h e i r s t r u c t u r a l r e l a t i o n s i s obviously too t i e d up i n t h e issue 207 to be an assumption which one could i n good conscience adopt. Rather, we whould regard the r e s u l t of Chapter Two as evidence f o r the p l a u s i b i l i t y of the assumption. Secondly, there i s a d i s t i n c t i o n between p r o v i d i n g a demonstration of the non-necessity of r e c u r s i v e generation of deep s t r u c t u r e types for sentences f o r the expression of thought on the one hand, and a demonstration of the e l i m i n a b i l i t y of r e c u r s i v e l y d e rived sentences i n paraphrase on the other. For although i t i s u n l i k e l y that we can do the l a t t e r i n a general way, i t i s p o s s i b l e to do the former i n a general way, i . e . i t i s p o s s i b l e to demonstrate i n a general way the non-n e c e s s i t y of r e c u r s i v e generation of deep s t r u c t u r e types f o r sentences f o r the expression of thought. The reader may object at t h i s point that i n Chapter One, part three, "On checking the statement of ne c e s s i t y " , 'the n e c e s s i t y of r e c u r s i v e d e r i v a t i o n s of deep s t r u c t u r e types f o r sentences 1 was taken to be equivalent to 'the n o n - e l i m i n a b i l i t y of r e c u r s i v e l y derived sentences i n paraphrase'. But t h i s i s not qu i t e accurate. In Chapter One, pa r t three, i t was argued (or shown) that a clai m of the n e c e s s i t y of r e c u r s i v e d e r i v a t i o n s of deep s t r u c -ture types f o r sentences f o r the f u l l expression of thought implies the n o n - e l i m i n a b i l i t y i n paraphrase of at f e a s t some types of r e c u r s i v e l y derived constituents i n sentences. And so Chapter Two approaches the statement of n e c e s s i t y by showing a l l r e c u r s i v e l y 208 derived sentences to be paraphraseable by sequences of sentences none of which i s a r e c u r s i v e l y derived sentence. But the statement that the statement of n e c e s s i t y implies the non-e l i m i n a b i l i t y of r e c u r s i v e d e r i v a t i o n s of sentences leaves open the question of whether the n o n - e l i m i n a b i l i t y i n paraphrase of r e c u r s i v e l y derived sentences would ( i f i t were true, which. Chapter Two argues, i t i s n ' t ) imply the n e c e s s i t y of r e c u r s i v e d e r i v a t i o n s of sentences f o r the expression of thought. I would l i k e to suggest here that even i f the a n a l y s i s i n Chapter Two f a i l e d , or were incomplete, there would be a general argument ag a i n s t the view that r e c u r s i v e d e r i v a t i o n of sentences i s necessary f o r thought expression. The argument however, leaves out too many c r u c i a l c onsiderations (considerations which the a n a l y s i s of Chapter Two includes) to be capable of r e p l a c i n g the a n a l y s i s of Chapter Two. Chomsky and other transformational l i n g u i s t s recognize that there are l i m i t s on performance such that o v e r l y - l o n g or exce-s s i v e l y complex sentences cannot be understood. However they s t a t e that the number of understandable sentences i s so large that i t would be absurd to view a generative grammar for a nature-language as anything but s y n t a c t i c a l l y r e c u r s i v e . I f t h i s i s t r i (which I have no reason to question) "then although we may not be able to understand i n f i n i t e l y many sentences, the grammar of our language can provide s t r u c t u r a l d e s c r i p t i o n s for an i n f i n i t 209 of formal objects or sentences. But now consider those formal objects or sentences which exceed performance l i m i t s ( i . e . , are too long and complex to understand) and yet which are s y n t a c t i c a l l y sound and can be give s t r u c t u r a l d e s c r i p t i o n s . The question i s , do they express thoughts which, i n v i r t u e of t h e i r t a x a t i o n of our performance c a p a c i t i e s , we can never grasp? i f there xs a way to eliminate the r e c u r s i v e l y d e r i v e d s t r u c t u r e s then those s t r u c t u r e s are not necessary f o r the f u l l expression of thought. On the other hand, i f there i s no way to eliminate r e c u r s i v e l y derxved s t r u c t u r e s i n paraphrase then we can never present those formal objects or sentences to ourselves i n such a way that we can grasp the thought 'expressed'. And i n that case, to generate the f i n i t e (though a s t r o n o m i c a l l y large) number of sentences which we can understand we could have an absurdly clumsy generative grammar which d i d not generate deep s t r u c t u r e types fo r sentences r e c u r s i v e l y , but was based on astronomically lar g e l i s t s of deep-structure-types of sentences. Despite the absurd clumsiness of the grammar, the statement that a grammar must contain devices which r e c u r s i v e l y generate deep s t r u c t u r e types for sentences would be blocked. For obvious reasons, the demonstration of the non-necessity of r e c u r s i v e generation of sentence s t r u c t u r e s f o r the f u l l expression of thought by demonstrating the p a r a p h r a s e a b i l i t y 210 o f r e c u r s i v e l y d e r ived sentences using no r e c u r s i v e l y derived sentences — although not a general argument ~~ i s by f a r the more p h i l o s o p h i c a l l y valuable. I n demonstrating the paraphra-s e a b i l i t y of r e c u r s i v e l y derived sentences using sentences which are not r e c u r s i v e l y derived one demonstrates that a small and f i n i t e number of deep s t r u c t u r e types f o r sentences c a n serve f o r the expression of a l l our thought., However, I must a l s o keep i t c l e a r as t o what p o r t i o n of the demonstration of Chapter Two has been accomplished: as mentioned at the outset of t h i s chapter, I have not o f f e r e d any t h e o r e t i c a l guarantees of the exhaustiveness of the l i s t , of r e c u r s i v e l y d e r ived c o n s t i t u e n t s , nor h a v e I given any assurances that there aren 11 other cases w i t h i n the l i s t , t h a t would present insurmountable obstacles to a paraphrase analysis,. Furthermore, I have not attempted t o show how, g i v e n a sentence to d e r i v e a s e r i e s of sentences S a...S n w h i c h express the thought of Sj_ using no r e c u r s i v e l y derived s t r u c t u r e : That i s , I have not attempted to give a d e c i s i o n procedure for the d e r i v a t i o n of the paraphrases. Nor, f o r that matter, h a v e I even discussed the problem of whether such a d e c i s i o n procedure e x i s t s . (But i t w i l l be discussed below, see page 241.) I have only argued that, as f a r as I could make out judging from a sampling of cases taken from a l i s t of r e c u r s i v e l y derived, c o n s t i t u e n t s , paraphrases of any sentences i n E n g l i s h could be constructed using sentences with no r e c u r s i v e l y derived constituen 211 CHAPTER THREE A. A T e c h n i c a l Problem. In Chapter Two I argued that we do not need r e c u r s i v e l y derived c o n s t i t u e n t s i n sentences to express our thoughts. In t h i s chapter I s h a l l explore some of the im p l i c a t i o n s of t h i s r e s u l t . The f i r s t implxcation to be examined i s the f o l l o w i n g : 3.1 Since we can have a f u l l y expressive language whose sentences contain no r e c u r s i v e l y derived c o n s t i t u e n t s we can have a grammar f o r a f u l l y expressive language whose base contains no r e c u r s i v e r u l e s . The i m p l i c a t i o n has been used i n previous remarks, but i t has never been c l o s e l y examined. In f a c t , there i s a b i t of a flaw i n the i m p l i c a t i o n which requires some patching. What I would l i k e to be able to say i s that si n c e we never need s t r u c t u r e s which are derived from the a p p l i c a t i o n of a set of r e c u r s i v e base r u l e s , that therefore we do not need any r e c u r s i v e base r u l e s . But here i s the h i t c h : while no r e c u r s i v e set of base r u l e s need be ap p l i e d i n any given d e r i v a t i o n , there i s nothing to say that the grammar does not requir e each of the base r u l e s i n the set for d i s t i n c t d e r i v a t i o n s . For example, the set of base r u l e s : 3.2a C L . C 2 + C 3 b c 3 — -> c 1 i s a r e c u r s i v e s et of r u l e s . (The n o t a t i o n a l conventions here are the same as i n Chapter One, so that C-j_. . .C n are grammatical 212 categories.) 3.2a and 3.2b c o n s t i t u t e a r e c u r s i v e set because they allow f o r a category, here C-^ , to dominate i t s e l f i n a p a r t i c u l a r deep s t r u c t u r e . Now the premise of 3.1 t e l l s us that we w i l l never need to have a s t r u c t u r e i n which Ci occurs dominating i t s e l f . But i t does not guarantee that we w i l l not need both r u l e s f o r use i n separate d e r i v a t i o n s . I f the base r u l e s which include 3.2a and 3.2b a l s o include; 3 .2c C 3 > C 4 + C 5 d c 4 > c 5 + c 6 e C 2 > C 3+ C 4 then we could have the f o l l o w i n g deep s t r u c t u r e types; 3.3 C x C2 c 4 C 5 and: 3.4 C 2 (This assumes that the procedure w i t h i n t h i s h y p o t h e t i c a l grammar f o r making a deep s t r u c t u r e type allows f o r more than one r u l e to be used as the f i r s t r u l e , as here we have two deep s t r u c t u r e types, one with C-j_ as the topmost category, the other with C 2 as the topmost category.) 213 Now n e i t h e r 3.3 nor 3.4 contain r e c u r s i v e l y derived c o n s t i t u e n t s ; that i s , i n both 3.3 and 3.4 no category dominates i t s e l f . But 3.3 contains an a p p l i c a t i o n of the base r u l e 3.2a, c l " "^c2 + c 3 ' a n d 3 - 4 contains an a p p l i c a t i o n of the base r u l e 3.2b, C 3 ~ >C1(- and thus w i t h i n the set of base r u l e s 3.2a...e there i s a r e c u r s i v e set of r u l e s c o n s t i t u t e d by 3.2a and 3.2b. To d e s c r i b e t h i s phenomenon, we could say that we have a set of base r u l e s 3.2a...e, and these r u l e s have generated deep s t r u c t u r e types 3.3 and 3.4. The whole set of base r u l e s i s a r e c u r s i v e set, but i n the generation of 3.4 and 3.5 the r e c u r s i v e s et has not been used r e c u r s i v e l y . And i n general we could dub t h i s phenomenon the phenomenon of 'unrealized r e c u r s i o n ' . Coming back to the i m p l i c a t i o n , 3.1, t h i s suggests that a m o d i f i c a t i o n i s r e q u i r e d . Roughly speaking, the r e q u i r e d m o d i f i c a t i o n would read something l i k e : 3.5 Since we can have a f u l l y expressive language whose sentences contain no r e c u r s i v e l y derived c o n s t i t u e n t s we can have a grammar f o r a f u l l y expressive language whose base contains no r e c u r s i v e r u l e s , or whose base contains r e c u r s i v e r u l e s which are never used r e c u r s i v e l y . But t h i s p a r t i c u l a r formulation w i l l not quite do. The concept of using r e c u r s i v e r u l e s r e c u r s i v e l y was defined i n terms of the r e l a t i o n s h i p between the r e c u r s i v e r u l e s and some deep s t r u c t u r e types derived from these r u l e s . But the c h a r a c t e r i z a t i 214 of the grammar should be r i c h enough so that the permanently u n r e a l i z e a b l e r e c u r s i o n i s st a t e d as a property of the r u l e s alone. F o r t u n a t e l y , i t i s not d i f f i c u l t to c o n t r i v e rules which b u i l d i n t o the grammar (or a grammar) the feature of perm-anently u n r e a l i z e a b l e r e c u r s i o n . One approach, for instance, would i n v o l v e s u b - c l a s s i f y i n g the base r u l e s i n t o a l l the non-r e c u r s i v e subsets and then adding to the d e c i s i o n procedure f o r c o n s t r u c t i n g d e r i v a t i o n s the s t i p u l a t i o n that every d e r i v a t i o n comes from only one of these subsets. Thus, using the base r u l e s 3.2a...e, we would have the two non-recursive subsets c o n s t i t u t e d by 3.2a, 3.2c, 3.2d, 3.2e, on the one hand and on the other by 3.2b, 3.2c, 3.2d, and 3.2e. Let me c a l l any such set of p r o v i s i o n s or r u l e s for b u i l d i n g i n t o the base of a grammar the feature of permanently u n r e a l i z e a b l e r e c u r s i o n , a 'recursion b l o c k ' . The i m p l i c a t i o n can now read: 3.6 Since we can have a f u l l y expressive language whose sentences contain no r e c u r s i v e l y derived c o n s t i t u e n t s we can have a grammar f o r a f u l l y expressive language whose base contains e i t h e r no r e c u r s i v e r u l e s , or r e c u r s i v e r u l e s with a r e c u r s i o n block. I b e l i e v e that t h i s formulation i s s u f f i c i e n t for our purposes. I f two r u l e s l i k e 3.2a and 3.2b c o n s t i t u t e a set of r e c u r s i v e r u l e s , but they are blocked from being used together i n any d e r i v a t i o n , then the base contains r e c u r s i v e r u l e s i n name 215 only. In f a c t , using the approach described above for the c o n s t r u c t i o n of a r e c u r s i o n block, i f base ru l e s are sub-class-i f i e d i n t o non-recursive sets and i t ' s these sub-sets which are used to generate the deep s t r u c t u r e types, then to say that the base comprises a set of r e c u r s i v e r u l e s would be vacuous. In Chapter One, a set of rul e s was defined as being r e c u r s i v e i n two ways: a set of r u l e s was s a i d to be r e c u r s i v e i f and only i f i t was capable of generating an i n f i n i t y of deep s t r u c t u r e types. And a s e t of r u l e s was s a i d to be r e c u r s i v e i f and only i f some category i n the set dominoed i t s e l f . The second def-i n i t i o n attempted t o make the f i r s t s p e c i f i c , or ' c l e r i c a l ' , so we should now modify the second d e f i n i t i o n of 'recursive' f o r a s e t of base r u l e s , i n the f o l l o w i n g way: a set of base r u l e s i s r e c u r s i v e i f and only i f some category dominates i t s e l f and the set of r u l e s (or procedure f o r de r i v a t i o n ) does not contain a r e c u r s i o n block. Once we have modified the c l e r i c a l d e f i n i t i o n of 'recursive set of r u l e s ' i n t h i s c l e a r l y a p p r o p r i -ate way we can see that the o r i g i n a l formulation i s indeed a good one: 3.1 Since we can have a f u l l y expressive language whose sentences contain no r e c u r s i v e l y derived c o n s t i t u e n t s , we can have a grammar f o r a f u l l y expressive language whose base contains no r e c u r s i v e r u l e s . And we have come f u l l c i r c l e . 216 B. Some r e s u l t s from Chapter Two. An important c o n c l u s i o n follows from 3.1. I t follows d i r e c t l y , and the j u s t i f i c a t i o n f o r i t w i l l be b r i e f . 3.7 Since we can have a grammar f o r a f u l l y expressive language whose base contains no r e c u r s i v e r u l e s , we can have a grammar for a f u l l y expressive language whose base generates only a f i n i t e number of deep s t r u c t u r e types. By d e f i n i t i o n , a set of base r u l e s i s r e c u r s i v e i f and only i f i t i s capable of generating an i n f i n i t y of deep s t r u c t u r e types. So i f a grammar contains no r e c u r s i v e base r u l e s , that grammar w i l l be capable of generating only a f i n i t e numberof deep s t r u c t u r e types. Now t h i s r e s u l t , that a f u l l y expressive language can e x i s t which contains only a f i n i t e number of deep s t r u c t u r e types, i s important. I t follows from t h i s t h a t : 3.8 I f the l e x i c o n i s a f i n i t e l i s t of items ( l e x i c a l e n t r i e s ) , then a f u l l y expressive language can e x i s t whose grammar generates only a f i n i t e number of deep s t r u c t u r e s . The d i f f e r e n c e between a deep s t r u c t u r e and a deep s t r u c t u r e type, i t w i l l be remembered, i s that a deep s t r u c t u r e type does not have l e x i c a l items included i n i t . A deep s t r u c t u r e has had l e x i c a l items i n s e r t e d . Thus a representation of a p a r t i c u l a r deep s t r u c t u r e type i s : 3.9 Sentence Noun ph"rase Verb phrase Determiner Noun Verbal Noun phrjase^ Determiner Noun 217 and a r e p r e s e n t a t i o n of a deep s t r u c t u r e (using t h i s deep s t r u c t u r e type) i s : 3.10 Sentence ^^Noion phrase Determiner Noun Verbal the g i r l Determiner I saw the Thus the c o n c l u s i o n that given a f i n i t e number of deep s t r u c t u r e types and a f i n i t e number of l e x i c a l items, there w i l l only be a f i n i t e number of deep s t r u c t u r e s follows d i r e c t l y , and again, i f there i s a f u l l y expressive language with only a f i n i t e number of deep s t r u c t u r e types, and i f the l e x i c o n i s a f i n i t e l i s t of l e x i c a l items, then a f u l l y expressive language can e x i s t whose grammar generates only a f i n i t e number of deep s t r u c t u r e s . And now we are only a small jump away from the f o l l o w i n g , f i n a l c o n c l u s i o n : 3.11 I f the l e x i c o n i s a f i n i t e l i s t of l e x i c a l e n t r i e s , then a f u l l y expressive language can e x i s t whose grammar generates only a f i n i t e number of sentences I f we continue to make the now standard assumption of trans-formational grammar that transformations operate only on i n d i v i d u a l deep s t r u c t u r e s , and do not combine deep s t r u c t u r e s — and there was nothing i n our examination of r e c u r s i v e l y derived s t r u c t u r e s to suggest that the assumption should be abandoned - then 3.11 follows d i r e c t l y from 3.8. But what i f that now standard assumption were to be 218 abandoned, someone might ask. In that case, the v a l i d i t y of the inference would depend on the grounds f o r the abandonment of the assumption. I f the grounds were the discovery of a sentence or type of sentence which had been ignored before, and which could not be as w e l l accounted f o r i n terms of r e c u r s i v e base r u l e s as i n terms of ge n e r a l i z e d transformations ( i . e . , transformations which combine separate deep structures) then the problem would be to discover i f the con s t i t u e n t s derived v i a the g e n e r a l i z e d transformations could be eliminated i n para-phrase. I f they couldn't then a considerable p a r t of my e n t e r p r i s e would be s c u t t l e d ; but i f they could, then the same concl u s i o n could be drawn using s l i g h t l y d i f f e r e n t formulations i n c l u d i n g both the non-necessity of s t r u c t u r e s derived from the a p p l i c a t i o n of the r e c u r s i v e base r u l e s , and of s t r u c t u r e s d e r i v e d from the a p p l i c a t i o n of gener a l i z e d transformations. In any case, u n t i l such problematic sentences are discovered, i t i s a f a i r l y remote xssue to be worrying about. Coming back to the point, we now have the f o l l o w i n g s t r i n g of i m p l i c a t i o n s : 3.12a A f u l l y expressive language can e x i s t whose grammar contains no r e c u r s i v e l y derived s t r u c t u r e s . b Therefore a f u l l y expressive language can e x i s t whose grammar contains no r e c u r s i v e base r u l e s . c Therefore a f u l l y expressive language can e x i s t whose grammar can generate only a f i n i t e number of deep s t r u c t u r e types. 219 d Therefore, given that a l e x i c o n i s a f i n i t e l i s t of items a f u l l y expressive language can e x i s t whose grammar can generate only a f i n i t e number of deep s t r u c t u r e s . e Therefore, given that a l e x i c o n i s a f i n i t e l i s t of items, a f u l l y expressive language can e x i s t whose grammar can generate only a f i n i t e number of sentences. I want to stop here, and dwell on the l a s t p o s s i b i l i t y , that a f u l l y expressive language could be constructed which contains (orwhose grammar generates) only a f i n i t e number of sentences. The inference that such a language can be constructed depends p r i n c i p a l l y on the premise that a l e x i c o n i s a f i n i t e l i s t of items. C. The size of the lexicon i n synchronic d e s c r i p t i o n . Now the f i r s t p o i n t to note i s that i t i s , i n f a c t , the standard conception of the transformation grammarians that the l e x i c o n i s a f i n i t e l i s t of l e x i c a l items, or d i c t i o n a r y e n t r i e s . Thus (Chomsky, Aspects, page 141); A grammar contains a s y n t a c t i c component, a semantic component, and a phonological component. The l a t t e r two are p u r e l y i n t e r p r e t i v e ; they p l a y no part i n the r e c u r s i v e generation of sentence s t r u c t u r e s . and f u r t h e r (Aspects page 142): The l e x i c o n c o n s i s t s of an unordered set of l e x i c a l e n t r i e s and c e r t a i n redundancy r u l e s . The c l a r i t y of the matter i s not helped any by the f a c t that i n some works Chomsky considers the l e x i c o n to be part of the s y n t a c t i c component, and i n some works he considers i t , or 220 appears to consider i t , to be part of the semantic component. Thus, i n Aspects, (page 141): The s y n t a c t i c component, c o n s i s t s of a base and a tr a n s f o r m a t i o n a l component. The base, i n turn, c o n s i s t s of a c a t e g o r i a l subcomponent and a l e x i c o n . Whereas, i n Topics (page 65): R e c a p i t u l a t i n g , we are proposing that the s y n t a c t i c component of a grammar c o n s i s t s of r e w r i t i n g r u l e s and tr a n s f o r m a t i o n a l r u l e s . Thus, i n Topics the base c o n s i s t s only of the r e w r i t i n g r u l e s . But even though the l o c a t i o n of the l e x i c o n i s a l i t t l e unclear, i t i s c l e a r that the l e x i c o n plays no r o l e i n the r e c u r s i v e generation of sentence s t r u c t u r e s , and hence must be a f i n i t e l i s t . Thus, even i n Aspects, where the l e x i c o n i s pa r t of the s y n t a c t i c component, Chomsky says, (Aspects, page 142): The i n f i n i t e generative c a p a c i t y of the grammar a r i s e s from a p a r t i c u l a r formal property of (the) c a t e g o r i a l r u l e s . So that the l e x i c o n , even as part of the s y n t a c t i c component, does not pl a y a r o l e i n the r e c u r s i v e generation of sentence s t r u c t u r e s . And a l s o , i n Topics, where the l e x i c o n i s not part of the synt-a c t i c component, Chomsky says, (Topics page 66): The r e c u r s i v e property of the grammar...is r e s t r i c t e d to the base component. Katz, furthermore, i s very e x p l i c i t on the subject of the l i m i t -ations on the s i z e of the l e x i c o n , (Philosophy of Language, page 154): Accordingly, we may regard the d i c t i o n a r y as a f i n i t e l i s t of r u l e s , c a l l e d ' d i c t i o n a r y e n t r i e s ' , each of which p a i r s a word with a representation of i t s meaning i n some normal form. 2 2 1 Thus, f o r Katz, the l e x i c o n i s a f i n i t e l i s t of d i c t i o n a r y e n t r i e s . And P o s t a l i s a l s o very e x p l i c i t on t h i s point (see Postal's Epilogue t o E n g l i s h Transformational Grammar by J , & R., page 279). But i t i s not c l e a r (and t h i s i s the second point) t h a t such a conception of the l e x i c o n i s adequate. My remarks h e r e are only t e n t a t i v e , but I b e l i e v e that considerations o f how t o represent, accommodate, and account f o r semantic change suggest that a l e x i c o n i s not best viewed as a f i n i t e l i s t , even w i t h regard to p r e s e n t i n g a synchronic d e s c r i p t i o n of i t . Now t h e r e are two kinds of semantic change, both of which are relevant, to c o n s i d e r a t i o n s of the s t r u c t u r e of the l e x i c o n , b u t o n l y one of which i s obviously and immediately r e l e v a n t to the number of l e x i c a l items. (a) There i s semantic change i n which a word acquires a new sense. And (b) there i s semantic change i n which a word i s added to the language, and thus, i n which a word i s added t o the l e x i c o n . The problems r a i s e d by (a) - type of semantic change are d i f f i c u l t indeed, and I s h a l l do no more here than to suggest that i t i s at l e a s t tempting to present a r e c u r s i v e and t r a n s -formational account of such phenomena as metaphor and metapho-r i c a l extensions of meaning. For an e x p l o r a t i o n of t h i s and r e l a t e d p o i n t s , see Botha, The Function of the Lexicon i n Generative Transformational Grammar, and Angel, A Transformational 222 A n a l y s i s of Metaphor. But t h i s p o i n t would only be r e l e v a n t here i f i t i s determined that i n counting the number of l e x i c a l items one must count the number of senses of the l e x i c a l items. As t h i s would need considerable conceptual digging before the p o i n t could be made, I w i l l not consider the problems r a i s e d by (a)-type of semantic change, and t h e i r i m p l i c a t i o n f o r a synchronic d e s c r i p t i o n of the l e x i c o n any f u r t h e r i n t h i s study. As f o r (b), there are three types of word-additions, the mechanisms f o r the i n t r o d u c t i o n of which r e q u i r e c l o s e i n s p e c t i o n . These are (1) neologisms i n the narrower sense, i . e . , words which are invented or borrowed from other languages to f i l l i n a gap i n the vocabulary of a p a r t i c u l a r language; (2) proper names; and (3) numerical nouns. E x a c t l y what c o n s t r a i n t s there are on the i n t r o d u c t i o n of neologisms i n t o a language i s a d i f f i c u l t question, and one which i s c e r t a i n l y worthy of i n v e s t i g a t i o n ; indeed, c o n s i d e r i n g that neologisms are so frequently borrowed from other languages, and that word-inventions so c h a r a c t e r i s t i c a l l y occur i n cases where there are s p e c i a l t e c h n i c a l requirements, no answer to the question as to whether the number of neologisms p o s s i b l e to introduce i n a language i s i n p r i n c i p l e u n l i m i t e d should be taken for granted. However, I think i t i s c l e a r that (2) proper names, and (3) 223 numerals are u n l i m i t e d i n number i n the E n g l i s h language. More or l e s s any permissable combination of phonemes can serve as a proper name, and there are an u n l i m i t e d number of such combinations. Thus: 3.13 I named my dog B r a k s o t e l a t i l i p o t l i c a m i t a n i s a t i s l e x . i s a p o s s i b l e sentence i n E n g l i s h , and an adequate grammar would have to provide i n some way f o r the f a c t that 3.13 i s a p e r f e c t l y comprehensible sentence i n E n g l i s h . Yet the person who encounters i t and understands i t with no d i f f i c u l t y may never have p r e v i o u s l y encountered one of the words i n the sentence. I t i s reasonable, t h e r e f o r e , to p o s t u l a t e that a synchronic d e s c r i p t i o n of the l e x i c o n i s not a f i n i t e l i s t , and that the grammar (perhaps the semantic component, or a part of the l e x i c o n i t s e l f ) contains some r u l e s which provide for the u n l i m i t e d a d d i t i o n i n t o the l e x i c o n of words which are subsumed under p a r t i c u l a r categories, l i k e the categories of proper noun and numerical noun. (I am using the terms 'proper name'and 'proper noun' interchangeably.) With numerical nouns, the account i s t r i c k i e r , as i t could come i n a number of ways. That i s , one could account i n a number of ways f o r the numerical noun i n a sentence l i k e : 3.14 There are 7,563,424 people l i v i n g i n Los Angeles. Among the p o s s i b i l i t i e s are (a) a s t r i c t l y 'mathematical' r e c u r s i v e system which the l e x i c o n borrows on i n d i r e c t l y . This approach would t r e a t the numerical nouns as a s i n g l e l e x i c a l item, 224 j u s t as the proper name i n 3.13 would, I think, have to be considered as a s i n g l e l e x i c a l item. The r e c u r s i v e processes which generate the numerical noun would stand outside of the l e x i c o n or outside of the semantic component i n the way that the p h o n o l o g i c a l r e c u r s i o n s which generate the i n f i n i t y of p o s s i b l e p h o n o l o g i c a l s t r i n g s , and hence the i n f i n i t y of s t r i n g s which can serve as i n d i v i d u a l proper nouns, stand outside of the l e x i c o n or outside of the semantic component. And (b) a set of r u l e s f o r the r e c u r s i v e generation of numerical nouns which mirror the mathematical formulations, and are considered as part of the l e x i c o n , or as pa r t of the semantic (or p o s s i b l y even s y n t a c t i c ) component of the grammar. I t i s not immediately c l e a r which of these i s p r e f e r a b l e , nor f o r that matter, how much the d i f f e r e n c e i s an a r b i t r a r y t e r m i n o l o g i c a l d i f f e r e n c e , and how much i t i s not. In general, from both the p o i n t of view of the philosophy of l i n g u i s t i c s and the point of view of the philosophy of mathematics, the r e l a t i o n s h i p s between l i n g u i s t i c representations of mathematical e n t i t i e s and the mathematical e n t i t i e s themselves are a subject worthy of serious i n v e s t i g a t i o n This problem centres around only one such r e l a t i o n s h i p , and undoubtedly there are others . i n any case, i t i s not my purpose"to dwell on these problems but to po i n t out that we should be very h e s i t a n t about deci d i n g that the way to conceive of the l e x i c o n i s as a f i n i t e l i s t of l e x i c a l e n t r i e s , a l i s t which plays no part i n the r e c u r s i v e 225 generation of sentence s t r u c t u r e s . D. The p l a u s i b i l i t y of e l i m i n a t i n g l e x i c a l r e c u r s i o n s . This does not mean, however, that we must r e j e c t the n o t i o n that some f u l l y expressive language might contain only a f i n i t e number of sentences. To do t h i s on the b a s i s that an adequate l e x i c o n f o r one of the n a t u r a l languages must comprise a p o t e n t i a l l y i n f i n i t e l i s t of l e x i c a l items would be to ignore the very d i s t i n c t i o n which Chapter Two i s based on: the d i f f e r e n c e between a grammar f o r a natural language, and the grammar f o r any f u l l y expressive language. J u s t as a sharp d i s t i n c t i o n can be made between asking whether we must po s t u l a t e r e c u r s i v e base r u l e s to adequately describe languages as they e x i s t , and asking whether we must p o s t u l a t e r e c u r s i v e base r u l e s f o r any language which i s to be f u l l y expressive, so we can make a sharp d i s t i n c t i o n between asking whether we must po s t u l a t e a l e x i c o n with an u n l i m i t e d number of e n t r i e s f o r a n a t u r a l language to account f o r various p r o p e r t i e s of that n a t u r a l language, and asking whether any f u l l y expressive language must have a l e x i c o n with an u n l i m i t e d number of e n t r i e s . i n f a c t , we might want to regard these as mutually implying d i s t i n c t i o n s ; thus, a complement to Chapter Two would be an essay which undertook to discover whether we need to have access to an u n l i m i t e d l e x i c o n i n order for us to be able to express 226 our thoughts. However, although the d i s t i n c t i o n s are p a r a l l e l the answers to the questions they r a i s e need not be. There-fore, I do not want to argue here fo r the view that a f u l l y expressive language need not have an u n l i m i t e d l e x i c o n . I only want to argue that such a view, at t h i s stage, remains p l a u s i b l e . The f o l l o w i n g points then, are not arguments f o r that view, but only prima f a c i e c o n s i d e r a t i o n s suggesting the p l a u s i b i l i t y of the view, or to put i t another way, prima f a c i e c o n s i d e r a t i o n s suggesting that one could not r e j e c t the view out of hand. a. I f a neologism i s introduced, i t i s g e n e r a l l y v i a some d e f i n i t i o n . Thus, i f a s c i e n t i s t says, "Let me c a l l such a s t r a l objects 'tropomorphs'", even i f we ignore the f a c t that the neologism u s u a l l y has some etymological j u s t i f i c a t i o n , we can see that the objects being named have already been designated — here by the phrase 'such a s t r a l o b j e c t s ' . There w i l l i n general be an i n t r o d u c i n g phrase fo r a neologism l i k e "I w i l l c a l l a s t r a l objects which emit Z - p a r t i c l e s only at the h i g h -po i n t of sunspot a c t i v i t i e s 'tropomorphs'." And i n that case, the neologism may be u s e f u l f o r economy, while being not a b s o l u t e l y necessary f o r expression. Instead of saying "The photograph records the presence of f i v e tropomorphs" one could say "The p i c t u r e records the presence of f i v e a s t r a l objects of the type which emit Z p a r t i c l e s only at the highpoint of sunspot 227 a c t i v i t i e s . " Considerations such as these would lead q u i t e d i r e c t l y i n t o a c o n s i d e r a t i o n of whether and how a minimal l e x i c a l l i s t could be constructed i n terms of which any concept could be introduced. There i s l i t t l e at the moment to suggest that the c o n s t r u c t i o n of such a minimal l e x i c a l l i s t would be a l t o g e t h e r impossible. Of course only a d e t a i l e d examination could attempt to show that neologisms i n general are e l i m i n a b l e , and only a d e t a i l e d examination could handle the problem of the i n t r o d u c t i o n of r e c u r s i v e d e r i v a t i o n s i n the course of e l i m i n a t i n g the l e x i c a l r e c u r s i o n s . In the above example the r e c u r s i v e l y derived c o n s t i t u e n t s are not d i f f i c u l t to e l iminate —• but what i s to guarantee that t h i s w i l l h o l d true i n general? Obviously, the question cannot be s e t t l e d with a cursory treatment but the attempt to construct a minimal l e x i c a l l i s t remains p l a u s i b l e . b. I t ' s p l a u s i b l e to argue that proper names can, i n general, f o r standard purposes of communication, be replaced by other expressions. Thus instead of saying "John saw the c a t " one could say something l i k e "The f e l l o w who you met on Saturday night saw the c a t . " And t h i s expression can be f u r t h e r para-phrased so as to eliminate the r e c u r s i v e l y derived r e l a t i v e c l a use. But could we eliminate a l l proper names? Once again, t h i s i s a question which i s not easy to answer b r i e f l y ; and even i f we couldn't dowithout proper names altogether, the 228 question of whether we need an u n l i m i t e d number of proper names would s t i l l be very much open to i n v e s t i g a t i o n . Obviously, proper names are not attached uniquely to o b j e c t s . 'John', f o r example, i s the name of more than one person, so the question would s t i l l be open as to whether a f i n i t e stock of names could do f o r a l l the objects we could need to name. This too i s a matter which cannot be s e t t l e d b y r e a d i l y a v a i l a b l e c o n s i d e r a t i o n s . c. I t i s p l a u s i b l e to argue that at l e a s t t h e o r e t i c a l l y we could do without an u n l i m i t e d number of numerical nouns. Instead of saying, "There were f i v e apples on the t a b l e . ", we could say, "Add one to one. Add one to your r e s u l t . Add one to your r e s u l t again. Add one to your r e s u l t again. C a l l your r e s u l t 'your f i n a l r e s u l t ' . A number of apples were on the t a b l e . Your f i n a l r e s u l t i s that number." Here we say that there were f i v e apples on the t a b l e using the s i n g l e numerical noun 'one'. This s u b s t i t u t e s an i n d e f i n i t e number of i d e n t i c a l sentences f o r an i n d e f i n i t e number of numerical nouns, and i t i s i n e f f e c t l i k e using a numerical system with a unary modulus, only i n the place of each column, we have a sentence. N a t u r a l l y , the argument would depend on a demonstration of the o p e r a t i o n a l equivalence of such a system with a numerical system such as we p r e s e n t l y use. These three points are very t e n t a t i v e , however, and while the view that we can have a f u l l y expressive language which 229 contains (or whose grammar generates) only a f i n i t e number of sentences i s at t h i s stage a p l a u s i b l e one, only a d e t a i l e d study could demonstrate that i t i s a view which does j u s t i c e to the f a c t s of language use. E. Impl i c a t i o n s of Chapter Two f o r semantic and p s y c h o l i n g u i s t i c theory. But even i f the semantic i n v e s t i g a t i o n were to conclude that l e x i c a l r e c u r s i o n s are not necessary f o r thought expression, s t i l l the check on the o r i g i n a l statement of n e c e s s i t y wduld not be exhausted. Returning to the task of checking Chomsky's theory t h a t the e x p l i c i t formulation of r e c u r s i v e generative r u l e s c o n s t i t u t e s a "formulation of the c r e a t i v e processes of language," and more concretely, that i f a language i s to accommodate the c r e a t i v e aspect of language use then i t s grammar must r e c u r s i v e l y generate sentences, i f the semantic i n v e s t i g a t i o n produced a negative conclusion, we would be i n a p o s i t i o n to conclude that a language could accommodate the c r e a t i v e aspect of language use ( i . e . i t could be f u l l y expressive) whose grammar could generate only a f i n i t e number of sentences. The statement of n e c e s s i t y then would be,in a s t r i c t sense, f a l s e . But the question could be r a i s e d as to whether or not the s t a t e -ment of n e c e s s i t y could be rendered true by i n t e r p r e t i n g i t i n a somewhat more general way as a s s e r t i n g that the r e c u r s i v e generation of l i n g u i s t i c s t r u c t u r e i s necessary f o r thought 230 expression. On such an i n t e r p r e t a t i o n , i t would be true, because 'paragraphs', or sequences of sentences impose no p r i n c i p l e d l i m i t s on thenumbers of sentences they contain, and thus paragraphs can be regarded as r e c u r s i v e l y derived con-ju n c t i o n s of sentences. Since paragraphs — some paragraphs, a t any rate — are very much l i k e s i n g l e sentences consisting of smaller sentences con-j o i n e d to each other, the t h r u s t of Chapter Two, i n one sense i s a demonstration that wholly nested r e c u r s i v e l y derived s t r u c -tures can be paraphrased using only 'conjunctive' r e c u r s i v e l y d e r i v e d s t r u c t u r e s . And so the question could be r a i s e d as to whether the r e s u l t of Chapter Two i s i n any way a s i g n i -f i c a n t r e s u l t — indeed, whether i t attacks the thought behind Chomsky's statement of n e c e s s i t y at a l l . S t r i c t l y from the po i n t of view of semantic theory, the r e s u l t that r e c u r s i v e nestings introduce no unique semantic e f f e c t s , i . e . , that thoughts expressed using r e c u r s i v e n e s t i n g can always be expressed using r e c u r s i v e conjunctions, has, i t seems to me, some i n t r i n s i c i n t e r e s t . In p a r t i c u l a r , i t appears to have a great deal of bearing on the question of whether the semantic content o f complex p r o p o s i t i o n s can be r e c u r s i v e l y s p e c i f i e d i n terms of the semantic content of t h e i r elementary p r o p o s i t i o n s and t h e i r s t r u c t u r a l r e l a t i o n s . And t h i s matter w i l l be touched upon l a t e r . But the question here i s , can t h i s r e s u l t have any bearing on p s y c h o l i n g u i s t i c theory, l i n g u i s t i c s as psychology, or philosophy of psychology. 231 The issue i s o b v i o u s ly a complex one, and so I s h a l l not suggest here that the statement of n e c e s s i t y even i f i t i s i n t e r p r e t e d generously i s f a l s e . I t i s c l e a r that the statement as formulated by Chomsky and as r e f l e c t e d i n the w r i t i n g s of Fodor, Katz, and others i s g r o s s l y inadequate. For the only r e c u r s i o n s p o s t u l a t e d and discussed by these l i n g u i s t s are not necessary f o r thought expression. Whether Chomsky's statement of n e c e s s i t y can be given a p r e c i s e and accurate s i g n i f i c a t i o n w i t h i n the general framework of 'Cartesian' l i n g u i s t i c s i s l e s s c l e a r . Moreover, i t would be unwise to t a c k l e t h i s question i n advance of the semantic i n v e s t i g a t i o n . I've argued that to h o l d l e x i c a l r e cursions to be unnecessary f o r thought express i o n i s p l a u s i b l e , but that a d e t a i l e d i n v e s t i g a t i o n i s required, an i n v e s t i g a t i o n which might w e l l be roughly on the s c a l e of Chapter Two here, and i n v o l v i n g perhaps even more than Chapter Two d i d , independently c o n t r o v e r s i a l i s s u e s . In t h i s f i n a l s e c t i o n , then, I would l i k e to make some explo r a t o r y remarks on the d i f f e r e n c e s between the f o l l o w i n g claims: 1. Some r e c u r s i v e device (s) i s (are) necessary f o r thought expression. 2. Some sentence-generating recursions are necessary f o r thought expression. 3. Recursive r u l e s generating i n f i n i t e l y many deep s t r u c t u r e types f o r sentences are necessary for thought expression. 4. L e x i c a l recursions are necessary f o r thought expression. 5. Paragraph-generating recursions are necessary f o r thought expression. 2 3 2 My aims here are modest. I simply hope to e s t a b l i s h that there are enough explorable d i f f e r e n c e s between these claims, d i f f e r e n c e s , that i s , with enough p h i l o s o p h i c a l content of one s o r t or another, to warrant having taken the trouble to devote a mono-graph to r e j e c t i n g one of them (namely #3) q u i t e apart from the f a c t that i t was a claim apparently advanced by a number of prominent l i n g u i s t s . Even a cursory examination of these claims w i l l r e v e a l , I think, that they are p h i l o s o p h i c a l l y d i s c r i m i n a t i n g . Consider the d i s t i n c t i o n between a s s e r t i n g (3) that recursions i n t r o -ducing an i n f i n i t y of deep s t r u c t u r e types are necessary f o r thought expression and a s s e r t i n g (4) that l e x i c a l r ecursions are necessary f o r thought expression. In Chapter Two we saw that the semantic e f f e c t s of p r o p o s i t i o n a l r e c u r s i o n can a l l be handled using n o n - r e c u r s i v e l y derived p r o p o s i t i o n s i n sequence, l a r g e l y through the use of c h a r a c t e r i z e r s and i n t e r -s e n t e n t i a l l y r e f e r r i n g devices of one s o r t or another, t y p i c a l l y , pronouns and proper names. In Chapter Three parts C and D i t was not c l e a r whether we would be able to do away with an i n f i n i t y of proper names and do away with an i n f i n i t y of deep s t r u c t u r e types at the same time (see page 227). But one point, to my mind, appears l i k e l y : i_f i t were to come to a choice between p r o p o s i t i o n a l r e c u r s i o n and l e x i c a l r e c u r s i o n , i . e . , i f t h e i r semantic e f f e c t s were discovered tobe s u b s t i t u t e a b l e , we would 233 have to say that the l e x i c a l recursions are primary, the p r o p o s i t i o n a l recursions secondary. The reason i s that (once again) we can construct sentences which are too long f o r us to understand, and yet which we can understand by using a s e r i e s ci sentences with i n t e r - s e n t e n t i a l r e f e r e n c e s . Thus, to express a l l our thoughts we do need the i n t e r - s e n t e n t i a l l y r e f e r r i n g devices — whose underlying recursions may or may not be e l i m i n a b l e — ; but we do not need the p r o p o s i t i o n a l r e c u r s i o n s , the r e c u r s i o n s i n t r o d u c i n g an i n f i n i t y of deep s t r u c t u r e types f o r sentences i n order to be able to express our thoughts. This p o i n t might be f u r t h e r a m p l i f i e d by r e f l e c t i o n s on the f a c t that the paragraph versus sentence d i s t i n c t i o n could not be e l i m i n a t e d from the theory of discourse; to represent conversations as s i n g l e sentences, or to represent s e r i e s of sentences i n which there are questions, imperatives, and i n c o n s i s t e n c i e s , say, as s i n g l e sentences, would be most p e c u l i a r . From t h i s p o i n t of view as w e l l , i t would seem to be u n f r u i t f u l to think of the p r o p o s i t i o n a l recursions as primary. Secondly, from the p h i l o s o p h i c a l standpoint we are not only i n t e r e s t e d i n which recursions are necessary f o r thought i expression, but why. Now i f we assume, for the sake of d i s c u s s i o n that l e x i c a l recursions i n t r o d u c i n g an i n f i n i t y of proper names are necessary, we might be able to e x p l a i n t h i s by s t a t i n g that we need an i n f i n i t y of -symbol sequences i n order to 234 be able to r e f e r to the i n f i n i t y of objects and classes of objects i n the universe. But t h i s i s a very p a r t i c u l a r i n f i n i t y . Only by performing an i n v e s t i g a t i o n such as was performed i n Chapter Two could we s t a t e whether there are prop-o s i t i o n a l connectives which introduce otherwise uncapturable semantic 'values' (such as the i n t r o d u c t i o n of s i n g u l a r terms, the naming of c l a s s e s of o b j e c t s ) . And Chapter Two demonstrates that there are no semantic e f f e c t s or values which can be i n t r o -duced only by having access to an i n f i n i t y of deep s t r u c t u r e types f o r sentences. Perhaps i t should be s t r e s s e d here that the assumption i n the previous paragraph that we need an i n f i n i t y of proper names to r e f e r to an i n f i n i t y of c l a s s e s of objects without using r u l e s which would introduce an i n f i n i t y of deep s t r u c t u r e types f o r sentences was made for the sake of d i s c u s s i o n . Chapter Three p a r t C and D suggests that the opposite i s s t i l l p l a u s i b l e . In any case, fo r the two reasons sketched here, the d i s t i n c t i o n between claims 3 and 4 about the n e c e s s i t y of recursions i n t r o d u c i n g an i n f i n i t y of deep s t r u c t u r e types as opposed to the n e c e s s i t y of l e x i c a l recursions would be a valuable one f o r future p h i l o s o p h i c a l e x p l o r a t i o n . Now consider the d i f f e r e n c e between a s s e r t i n g paragraph-generating recursions to be necessary fo r thought expression as opposed to a s s e r t i n g sentence-generating recursions to be 235 necessary for thought expression. Once again the d i f f e r e n c e i s p h i l o s o p h i c a l l y d i s c r i m i n a t i n g . I s h a l l present two c o n s i -derations to t h i s e f f e c t , the f i r s t based on our o r i g i n a l i n t e r e s t i n formulations of r e c u r s i v e systems of r u l e s , the second p i c k i n g up on a t e c h n i c a l d i f f e r e n c e between para-graph as opposed to sentence generating recursions i n v i r t u e of which a statement of the n e c e s s i t y of paragraph generating r e c u r s i o n s could not have a most important i m p l i c a t i o n of a statement of the n e c e s s i t y of sentence-generating r e c u r s i o n s . ( 1 ) Our o r i g i n a l i n t e r e s t i n examining the statement that sentence generating recursions were necessary f o r thought expression was i n the l o c a t i o n of formal devices of our grammar which i n some way r e f l e c t the unique p r o p e r t i e s of the human i n t e l l e c t . I t has already been argued that a r e c u r s i v e generation of messages on the part of some creature does not guarantee that the creature shares what Chomsky and transformational l i n g u i s t s i n general regard as the unique and d i s t i n c t i v e p r o p e r t i e s of human mind (see the d i s c u s s i o n of the 'mathematical' h y p o t h e t i c a l b i r d songs, page 4 7 ) . Here an even stronger p o i n t can be made; i n order to describe the barking of a dog we need to p o s t u l a t e r e c u r s i v e generation of messages — there i s no t h e o r e t i c a l l i m i t of the number of barks i n any given sequence of barking. S i m i l a r l y , to describe such a c t i v i t i e s as walking, running, and breathing, we w i l l have to employ r e c u r s i v e r u l e s . Chomsky c o r r e c t l y observes that r e c u r s i v e generation of such items does 236 not c o n t a i n s t r u c t u r a l complexities i n any way resembling the semantic complexities of human speech. And i n Language and Mind page 62, Chomsky observes: There i s nothing u s e f u l to be s a i d about behaviour or thought at the l e v e l of a b s t r a c t i o n at which animal and human communication f a l l together. This i s a ra t h e r strong statement, but even a s o f t e r v e r s i o n of i t would be strong enough to r a i s e the question: i n v i r t u e of what would we say t h a t the paragraph generating r e c u r s i o n s r e f l e c t the unique p r o p e r t i e s of the human i n t e l l e c t ? Of course, such a question can't be e a s i l y r esolved; f o r one t h i n g , the r o l e of i n t e r - s e n t e n t i a l l y r e f e r r i n g expressions would have to be explored. One d i f f i c u l t y f o r the Ca r t e s i a n l i n g u i s t who would want to maintain Chomsky's focus on syntax i s that i n t e r - s e n t e n t i a l r e f e r r i n g devices are a sub-class of r e f e r r i n g devices i n general. Thus, -.if one i s i n t e r e s t e d i n the formal devices which r e f l e c t the unique q u a l i t i e s of the human i n t e l l e c t , i t would be much more to the point to launch an i n v e s t i g a t i o n i n t o the n e c e s s i t y or non-necessity of various r e f e r r i n g devices, f e a t u r e - p l a c i n g devices, p r e d i c a t i n g devices, and so on rather than to observe that conjunction of messages with no p r i n c i p l e d l i m i t s on the s i z e of the conjunctions i s (though there might be some argument - here too) a necessary c o n d i t i o n f o r a r a t i o n a l l y useable language. In p a r t i c u l a r a study would have to be done explo r i n g the r e l a t i o n s h i p between i n t e r - s e n t e n t i a l r e f e r r i n g devices ("that argument", f o r example, 237 or "the previous p r o p o s i t i o n " , or "the point j u s t made",) and other demonstratives r e f e r r i n g to a b s t r a c t objects (such as "that landscape" or "the shape of that c h a i r " ) . A f u r t h e r d i f f i c u l t y f o r those i n c l i n e d to i n t e r p r e t the statement of n e c e s s i t y more broadly so as to include paragraphs rather than to l i m i t the statement to the r e c u r s i v e generation of sentences i s t h i s : there may w e l l be no c l e a r sense to the notion of a r e c u r s i v e system for the generation of a l l and only the well-formed paragraphs of E n g l i s h ; on the other hand, the notion of formulating a r e c u r s i v e system for the generation of a l l and only the well-formed sentences of E n g l i s h i s at the heart of modern generative l i n g u i s t i c s . (2) The second c o n s i d e r a t i o n i s complex and has many important t h e o r e t i c a l underpinnings. Here I s h a l l do no more than attempt to sketch a few connections, l e a v i n g some of them open to future confirmation or r e j e c t i o n . The issue I take to be important enough to merit treatment here even i f at l e a s t one c e n t r a l connection w i l l not be f u l l y j u s t i f i e d . The over-a l l d r i f t i s t hat an argument may be a v a i l a b l e to the e f f e c t that i f p r o p o s i t i o n a l recursions (in a r e s t r i c t e d sense to be s p e c i f i e d below) are necessary fo r thought expression, then no mechanical explanation of the human use of language i s c o n s t r u c t i b l e But no such i m p l i c a t i o n could be drawn from the statement that paragraph-generating recursions are necessary fo r thought 238 expression. I f so, the two statements of n e c e s s i t y have c r u c i a l l y d i f f e r e n t i m p l i c a t i o n s f o r p s y c h o l i n g u i s t i c s and p h i l o s o p h i c a l psychology. Accordingly, r e j e c t i n g one statement of n e c e s s i t y while admitting the other i s p h i l o s o p h i c a l l y i l l u m i n a t i n g . In p a r t i c u l a r , I w i l l attempt to show that i_f a c e r t a i n c l a s s of r e c u r s i v e l y derived sentences were necessary f o r thought expression then no machine could ever simulate the human use of language. Thus i n demonstrating the non-necessity of a l l r e c u r s i v e l y derived sentences f o r thought expression (Chapter Two) hopes may be r e v i v e d f o r c o n s t r u c t i n g a machine which can use language as humans can. Throughout the d i s c u s s i o n which immediately fo l l o w s , then, i t should be borne i n mind that I w i l l be attempting to j u s t i f y an i m p l i c a t i o n whose premise has already been demonstrated to be f a l s e . Let me begin by d i s c u s s i n g what I s h a l l r e f e r to as 'the i m p l i c a t i o n ' or 'Implication I': I m p l i c a t i o n I: I f p r o p o s i t i o n a l recursions are necessary f o r thought expression then no machine can be given a programme for simulating the human use of language. Here ' p r o p o s i t i o n a l r e c u r s i o n s ' r e f e r s to the r e c u r s i v e devices whereby an i n f i n i t y of deep s t r u c t u r e types are generated based on the r e c u r s i v e embedding (or nesting) of p r o p o s i t i o n a l content i n t o l a r g e r p r o p o s i t i o n s . Now t h i s i m p l i c a t i o n , as b r i e f l y mentioned i n Chapter One, part two (see pages 38 - 39) 239 lurks i n many places i n Chomsky's w r i t i n g s , and i n a c e r t a i n sense may provide the 'ultimate j u s t i f i c a t i o n ' for Chomsky's frequent a s s o c i a t i o n of the presence of r e c u r s i v e generation of sentence s t r u c t u r e s and the possessing of s p e c i a l q u a l i t i e s of i n t e l l e c t and mental s t r u c t u r e which cannot be explained on mechanical models. As p r e v i o u s l y observed, the i m p l i c a t i o n when i t appears i n Chomsky's work i s mysterious because no reasoned argument i s presented i n support of i t . But the i m p l i c a t i o n i n i t s e l f , I s h a l l suggest, i s q u i t e p o s s i b l y sound. Although a passage has already been c i t e d with reference to t h i s i m p l i c a t i o n (page 38), i t w i l l do no harm to c i t e a second such passage here; i n Current Issues, Chomsky s t a t e s , (page 50 as anthologized i n The Structure of Language, ed. Fodor & K a t z ) ; The c e n t r a l f a c t to which any s i g n i f i c a n t l i n g u i s t i c theory must address i t s e l f i s t h i s : a mature speaker can produce a new sentence of h i s language on the appropriate occasion, and other speakers can understand i t immediately, though i t i s e q u a l l y new to them...It i s evident that r o t e r e c a l l i s a f a c t o r of minute importance i n ordinary use of language... the r e a l i z a t i o n that t h i s ' c r e a t i v e ' aspect of language i s i t s e s s e n t i a l c h a r a c t e r i s t i c can be traced back at l e a s t to the seventeenth century. Thus we f i n d the C a r t e s i a n view that man alone i s more than mere automatism, and that i t i s the possession of true language that i s the primary i n d i c a t o r of t h i s . . . The a s s o c i a t i o n here, as before, i s quite d i r e c t between a b i l i t y to produce and understand an i n d e f i n i t e number of sentences, and being a creature whose f a c u l t i e s are not e x p l i c a b l e on mechanical models. I f the i m p l i c a t i o n does not l u r k here, then the a s s o c i a t i o n between novelty of sentence production and not being 240 an 'automaton' i s u t t e r l y mysterious. However Chomsky makes no e f f o r t to j u s t i f y the i m p l i c a t i o n ; perhaps f o r t h i s reason many readers are i n c l i n e d to simply r e j e c t the i m p l i c a t i o n out of hand. This, however, would not do j u s t i c e to the i s s u e . I s h a l l suggest here that there may w e l l be a s o l i d j u s t -i f i c a t i o n f o r the i m p l i c a t i o n , based on the f a c t that statements about r e c u r s i v e embedding, or nestin g of p r o p o s i t i o n a l content are not s u b s t i t u t e a b l e i n c e r t a i n respects by statements about r e c u r s i v e conjunction of p r o p o s i t i o n a l content. In p a r t i c u l a r , c o n j o i n i n g p r o p o s i t i o n s so as to form conjunctive sentences i s qu i t e obviously i n at l e a s t some cases very much l i k e c o n j o i n i n g sentences to form paragraphs. But i t i s not c l e a r (without an a n a l y s i s such as i s given i n Chapter Two, of course) whether embedding p r o p o s i t i o n s to form complex sentences with various dependent and independent clauses i s or i s not ever very much l i k e c o n j o i n i n g sentences to form para-graphs. Through the d i f f e r e n c e between statements about embedding of p r o p o s i t i o n a l content as opposed to conjunction of p r o p o s i t i o n a l content, I s h a l l t r y to j u s t i f y the i m p l i c a t i o n as s y s t e m a t i c a l l y and yet as b r i e f l y as p o s s i b l e . Consider the f o l l o w i n g i m p l i c a t i o n : I m p l i c a t i o n II I f A: the semantic content (or truth-cond-i t i o n s ) of a complex sentence i s a ( c l e r i c a l l y computeable fu n c t i o n of the semantic content (or t r u t h - c o n d i t i o n s ) of i t s elementary propositions and t h e i r s t r u c t u r a l r e l a t i o n s 241 then B: there i s a d e c i s i o n procedure ( c l e r i c a l computation) f o r d e r i v i n g simple sentence conjunctions from complex sentences with i d e n t i c a l semantic content (or t r u t h conditions) and v i c e v e r s a . I cannot f u l l y j u s t i f y i m p l i c a t i o n I I , but consider i t to be eminently p l a u s i b l e . For i f A, then one can represent the semantic content or t r u t h - c o n d i t i o n s of a complex p r o p o s i t i o n as an ordered s et of i t s elementary p r o p o s i t i o n s , followed by a second ordered s et of pr o p o s i t i o n s which s p e c i f y i n gramm-a t i c a l language the s t r u c t u r a l r e l a t i o n s among the pr o p o s i t i o n s of the f i r s t s e t . Already the ordered set of both sets very much approaches the d e r i v a t i o n spoken of i n B. The p r i n c i p a l problem here i s the reduct i o n of the m e t a l i n g u i s t i c elements of the second set of pr o p o s i t i o n s i n t o f i r s t - o r d e r terms, and t h i s i s a problem which would appear to be s o l v a b l e . To i l l u s t r a t e , consider the sentences SI The woman who entered the press club i s the author of se v e r a l books. The two ordered sets here might be: A. 1. The woman entered the press c l u b . 2. The woman i s the author of se v e r a l books. and: B. 1. P r o p o s i t i o n (A.l) i s "wh—" i n s e r t e d i n p r o p o s i t i o n (A.2). 2. P r o p o s i t i o n (A.2) i s an - independent clause, (or main clause) . And the problem here i s whether the information i n (B.l) and (B.2) can be supplemented with a d e c i s i o n procedure f o r the d e r i v a t i o n of a paraphrase for SI such as: 242 S2 Some woman entered t h e press c l u b . That woman i s the author of s e v e r a l books. •Preliminary c o n s i d e r a t i o n s suggest that d e c i s i o n procedures of that kind might be p o s s i b l e t o c o n s t r u c t . F o r a c e r t a i n c l a s s of sentences of which S I i s a member t h e f o l l o w i n g d e c i s i o n procedure w i l l be adequate: C. 1. Under c o n d i t i o n t h a t some p r o p o s i t i o n Pj_ i s "wh-~ " i n s e r t e d i n some o t h e r p r o p o s i t i o n P 2 ( i t r a n s f o r m the Det of leftmost Np of P]_ to 'Some! , and; 2. Under the same c o n d i t i o n , transform t h e D e t of t h e leftmost Np of P 2 to 'that'. The c l a s s of sentences covered by such a d e c i s i o n procedure xs obvi o u s l y q u i t e a l i m i t e d c l a s s , but t h i s example may s e r v e to i l l u s t r a t e the nature of the problem and i n what way i t can be t a c k l e d . S i m i l a r p r e l i m i n a r y considerations w i t h other-recurs i v e l y d e r i ved c o n s t i t u e n t types suggest, t h a t t h e problem i n general may be s o l v a b l e , i . e . , that there i s a d e c i s i o n procedure f o r the d e r i v a t i o n of the paraphrases of C h a p t e r Two., I t i s p l a u s i b l e , then, to e n t e r t a i n the c o n s t r u c t i o n o t sue! a d e c i s i o n procedure, and fu r t h e r to assume t h a t I m p l i c a t i o n I I h o l d s . Consider, now, the f o l l o w i n g i m p l i c a t i o n : I m p l i c a t i o n I I I I f C: p r o p o s i t i o n a l recursions are necessary f o r thought expression then B: there i s a d e c i s i o n procedure ( c l e r i c a l computation) f o r d e r i v i n g simple sentence conjunctions from complex sentences with i d e n t i c a l semantic content (or t r u t h - c o n d i t i o n s ) and v i c e - v e r s a i s f a l s e . 243 That i s , i f (C), then (B i s f a l s e ) . I m p l i c a t i o n I I I follows a f o r t i r i o r i from the i m p l i c a t i o n argued i n Chapter One part three that i f some c l a s s of sentences i s necessary f o r thought expression then there are at l e a s t some members of the c l a s s which cannot be paraphrased using sentences which are not members of the c l a s s , i . e . , there are some members of the c l a s s which are non-eliminable i n paraphrase. Obviously i f p r o p o s i t i o n a l r e c u r s i o n s are non-eliminable i n paraphrase, then the c l a i m that e l i m i n a t i o n of r e c u r s i v e l y derived s t r u c t u r e s can be done through some d e c i s i o n procedure i s f a l s e . But i f B i s f a l s e , then A i s f a l s e through c o n t r a p o s i t i o n on i m p l i c a t i o n I I Thus, i f C, then A i s f a l s e , which we can w r i t e up as: Im p l i c a t i o n IV: I f C: p r o p o s i t i o n a l recursions are necessary fo r thought expression then A: the semantic content (or t r u t h -conditions) of a complex sentence i s a ( c l e r i c a l l y computeable) f u n c t i o n of the semantic content (or t r u t h conditions) of i t s elementary p r o p o s i t i o n s and t h e i r s t r u c t u r a l r e l a t i o n s , i s f a l s e . (I.e: i f (C) then (A i s f a l s e ) . ) But A, that the semantic content of a complex p r o p o s i t i o n i s a fun cuntion of i t s elementary p r o p o s i t i o n s and t h e i r s t r u c t u r a l r e l a t i o n s i s a p r e c o n d i t i o n f o r any mechanical explanation or simulat i o n of our use of language. For a machine must be a f i n i t e device, using only c l e r i c a l r o u tines and r e l a t i n g them to input of various kinds. In f a c t many semanticists assume that A: 244 the semantic content of complex p r o p o s i t i o n i s a f u n c t i o n of i t s elementary p r o p o s i t i o n s and t h e i r s t r u c t u r a l r e l a t i o n s i s true anyway, as a p o i n t i n semantic metatheory. Davidson's co n s i d e r a t i o n s i n On l a y i n g That (see pages 159 - 161, as anthologized i n Words and Objections,) would apply a l l the more so to the statement that A, the semantic content of a complex p r o p o s i t i o n i s a f u n c t i o n of the semantic content of i t s elementary p r o p o s i t i o n s and t h e i r s t r u c t u r a l r e l a t i o n s , i s a p r e c o n d i t i o n f o r any mechanical s i m u l a t i o n of human language use. I f we assume that A, that the semantic content of a complex p r o p o s i t i o n i s a f u n c t i o n of the semantic content of i t s elementary p r o p o s i t i o n s and t h e i r s t r u c t u r a l r e l a t i o n s , i s a p r e c o n d i t i o n f o r a mechanical explanation of human use of language, then i f A i s f a l s e , by c o n t r a p o s i t i o n , there can be no mechanical explanation or s i m u l a t i o n of the human use of language. And so from I m p l i c a t i o n IV, we can derive the o r i g i n a l i m p l i c a t i o n : I m p l i c a t i o n I: I f p r o p o s i t i o n a l recursions are necessary f o r thought expression, then there can be no mechanical explanation or s i m u l a t i o n of our use of language. Now observe that the i m p l i c a t i o n : I m p l i c a t i o n V: I f D: paragraph-generating recursions are necessary for thought expression then there can be no mechanical explanation or simulation of our use of language. 245 simply cannot be constructed. In p a r t i c u l a r , there i s no p a r a l l e l to i m p l i c a t i o n I I I , because paragraphs are conjunctions of sentences and so the d e s i r e d i m p l i c a t i o n - p a r a l l e l cannot be constructed. I m p l i c a t i o n I I I ' I f C' paragraph-generating r e c u r s i o n s are necessary f o r thought expression then B' there i s a d e c i s i o n procedure f o r d e r i v i n g sentence—conjunctions from paragraphs with i d e n t i c a l semantic content and v i c e versa i s f a l s e : ( i . e . , i f (C') then (B' i s f a l s e ) . ) i s p l a i n l y absurd as the paragraphs are already sentence conjunctions. Thus a s s e r t i n g that p r o p o s i t i o n a l recursions are necessary f o r thought expression appears to have q u i t e d i f f e r e n t i m p l i c a t i o n s f o r p s y c h o l i n g u i s t i c theory and the philosophy of mind as compared against the a s s e r t i o n that paragraph-recurs ions are necessary f o r thought expression. The d i s t i n c t i o n between the two statements of n e c e s s i t y may prove upon fu r t h e r a n a l y s i s to be fundamental, the l a t t e r being true but r e l a t i v e l y un-informative about the d i s t i n c t i v e l y human use of language, the former being r i c h i n i m p l i c a t i o n s f o r p s y c h o l i n g u i s t i c theory, but, i f the argument of Chapter Two i s c o r r e c t , f a l s e . Once again, though, I would l i k e to s t r e s s that i n t h i s f i n a l chapter I have only been concerned with sketching a number of d i r e c t i o n s f o r future i n v e s t i g a t i o n based on the work of Chapter Two. I t might a l s o be added at t h i s point that sema-n t i c i s t s i n t e r e s t e d i n j u s t i f y i n g what remains for them an 246 assumption, namely that A: the semantic content of a complex p r o p o s i t i o n i s a f u n c t i o n of the semantic content of i t s elementary p r o p o s i t i o n s and t h e i r s t r u c t u r a l r e l a t i o n s might w e l l regard the a n a l y s i s of Chapter Two as a d e t a i l e d p r e s e n t a t i o n of the p l a u s i b i l i t y of the assumption. In demonstrating the non-necessity of r e c u r s i v e d e r i v a t i o n s of sentences one seems to be c o v e r t l y demonstrating the p l a u s i b i l i t y of p r o v i d i n g r e c u r s i v e s p e c i f i c a t i o n s f o r the semantic content of complex p r o p o s i t i o n s ; concomitantly one also seems to be demonstrating the p l a u s i b i l i t y of c o n s t r u c t i n g a semantic theory i n which the semantic content of a p r o p o s i t i o n , or message, no matter how complex, i s determined by conjunctions of elementary p r o p o s i t i o n s which are e i t h e r f i n i t e i n number or of a small and f i n i t e number of types. F . Conclus i o n . To summarize, our concern has been to r a i s e some questions with regard t o the r e l a t i o n s h i p between d e s c r i p t i v e and explana-t o r y l i n g u i s t i c s as conceived by Chomsky, and i n p a r t i c u l a r to check the statement that i f a language i s to be ' c r e a t i v e l y useable' i t must contain a r e c u r s i v e device f o r the generation of sentences, or more broadly, of l i n g u i s t i c s t r u c t u r e s i n general. I have argued that the only recursions which Chomsky postula t e s and attempts to formulate are not required f o r thought 247 expression. That i s , the r e c u r s i v e generation of deep s t r u c t u r e types f o r sentences i s not necessary f o r thought expression. I suggest that a d e s c r i p t i o n of the l e x i c o n of E n g l i s h must cont a i n some r e c u r s i o n s , but that i t i s at l e a s t p l a u s i b l e that these too are unnecessary f o r thought expression. Paragraphs or sequences of sentences must be described at r e c u r s i v e l y d e r i v e d objects s i n c e there i s no l i m i t to the number of sentences i n any paragraph, but s i m i l a r l y there i s no l i m i t to the number of steps i n an act of walking, or barks i n an inst a n c e of a dog's barking. I t may be, rather, that the devices i n language which are t r u l y d i s t i n c t i v e and e s s e n t i a l as p r o p e r t i e s of human i n t e l l i g e n c e are embodied i n the devices of reference and p r e d i c a t i o n whereby i n d e f i n i t e l y many a b s t r a c t as w e l l as concrete objects may be re f e r r e d to by i n d i v i d u a l symbols and have predicates ascribed to them. I t appears, concomitantly, that an important i n v e s t i g a t i o n remains to be done i n t o the nature of references to p r o p o s i t i o n s and events and so f o r t h r e f e r r e d to i n previous and subsequent p r o p o s i t i o n s on the one hand, and the nature of references to other a b s t r a c t e n t i t i e s such as a landscape depicted i n some p a i n t i n g , or the shape of some object, on the other, to see whether there are e s s e n t i a l d i f f e r e n c e s between them. Also, i t appears that the n e c e s s i t y of p r o p o s i t i o n a l embedding 248 implies the non-mechanical nature of language use, but that no such i m p l i c a t i o n could be drawn from the n e c e s s i t y of c o n j o i n i n g sentences to form paragraphs. I f i t i s i n some sense an ultimate i n t e r e s t of Chomsky to draw from the f a c t that our competence extends over an i n f i n i t e range of sentences, the conclusion that our mental operations are q u a l i t a t i v e l y d i s t i n c t from those of other animals, and are not e x p l i c a b l e on mechanical models, then the p a r t i c u l a r a n a l y s i s of Chapter Two prevents the ' u l t i m a t e l y i n t e r e s t i n g ' c o n c l u s i o n from being drawn. 249 LIST OF WORKS CITED L. Angel A Transformational A n a l y s i s of Metaphor, MA t h e s i s U n i v e r s i t y of B r i t i s h Columbia, 1968. Jonathan Bennett, R a t i o n a l i t y , Routledge & Kegan Paul, 1964. Max Black, Comment on Chomsky's Problems of Explanations i n L i n g u i s t i c s i n Explanation i n the Behavioural Sciences, ed. by R. Borger and F. C i o f f i , Cambridge U n i v e r s i t y Press, 1970. R. Botha, The Function of the Lexicon i n Transformational Generative Grammar, Mouton & Co., 1968. N. Chomsky/ Three Models f o r the D e s c r i p t i o n of Language, IRE t r a n s a c t i o n s on Information Theory V o l . IT-2, 1956. On the Notion Rule of Grammar, i-n T h e S t r u c t u r e of Language, ed. Fodor & Katz, P r e n t i c e H a l l , 1964. Current Issues i n L i n g u i s t i c Theory, i n The S t r u c t u r e of Language, ed. Fodor & Katz, P r e n t i c e H a l l , 1964. P e r s i s t e n t Topics i n L i n g u i s t i c Theory, Diogenes #51, Mario C a s a l i n i Ltd., Montreal, 1965. Aspects of the Theory of Syntax, MIT Press, 1965. Topics i n the Theory of Generative Grammar, Mouton & Co., 1966. C a r t e s i a n L i n g u i s t i c s , Harper and Row, 1966. Language and Mind, Harcourt Brace & World Inc. 1968, Form & Meaning i n Natural Languages, i n Language & Mind, enlarged e d i t i o n , Harcourt Brace & World Inc. 1970. Reply to Black's Comment on Chomsky's Problems of Explanation i n L i n g u i s t i c s i n Explanation i n the Behavioural Sciences, ed. by Borger and C i o f f i , Cambridge U n i v e r s i t y Press, 1970. ! Some Remarks on Nominalization, Readings i n E n g l i s h Trans formational Grammar, #323 . Selected Readings, A l l e n & Van Buren eds. Oxford U n i v e r s i t y Press, 1971. A. Church, On Carnap's A n a l y s i s of Statements of A s s e r t i o n and B e l i e f , Analysis 10, A p r i l , 1950. D. Davidson, On Saying That i n Words and Objections, Harman and Davidson eds., D. Reidel P u b l i s h i n g Co, 1969. J . Fodor, P s y c h o l o g i c a l Explanation, Random House, 1968. 250 P. Geach, J . Hochberg & R. Zimmerman, R. Jacobs & P. Rosenbaum, J - Katz, P.&C. K i p a r s k i , K. Lashley, Reference and G e n e r a l i t y , C o r n e l l U n i v e r s i t y Press, 1962-P i c t o r i a l Recognition i n the Infant Monkey, Proceedings of the Psychonomics Society, 1963. E n g l i s h Transformational Grammar, B l a i s d e l l P u b l i s h i n g Co., 1968. The Philosophy of Language, Harper & Row, 1966. Fact, i n Semantics, an I n t e r d i s c i p l i n a r y Reader, ed. D. Steinberg and L.Jakobovits.Camb.U. Press, The Problem of S e r i a l Order i n Behaviour, i n L. J e f f r e s s i.9 / ed, Cerebral Mechanisms i n Behaviour: New York, Wiley, 1951 D. Langendoen, J . Lyons, P. P o s t a l , W.V.O. Quine, Ross, P.F. Strawson, Zeno Vendler, V. H. Yngve, The Study of Syntax, H o l t Rinehart & Winston Inc. 1969 Chomsky, Fontana-Collins, 1970. Epilogue to E n g l i s h Transformational Grammar, B l a i s d e l l P u b l i s h i n g Co., 1968. Word & Object, MIT Press, 1960. A u x i l i a r i e s as Main Verbs, Grammar and Philosophy i n L o g i c o - L i n g u i s t i c Paperu, Methuen & Co., 1971. Truth, i n Logico-Linguis t i c Papers , Methuen & Co . I \) /1 L i n g u i s t i c s i n Philosophy, C o r n e l l U n i v e r s i t y Press, 1967. A Model and an Hypothesis f o r Language Stru c t u r e , T e c h n i c a l Report 3 69, August 15, 1960, MIT Research Lab. of E l e c t r o n i c s -2 5 1 LIST OF ABBREVIATIONS Aspects Aspects of the Theory of Syntax C.L. C a r t e s i a n L i n g u i s t i c s Current Issues Current Issues i n L i n g u i s t i c Theory P e r s i s t e n t Topics P e r s i s t e n t Topics i n L i n g u i s t i c Theory L & M Language and Mind, 1st e d i t i o n S e l . Read. Chomsky, Sele c t e d Readings, A l l e n and Van Buren eds. Topics Topics i n the Theory of Generative Grammar j & R E n g l i s h Transformational Grammar, Jacobs & Rosenbaum. ) 252 APPENDIX: Recursion and the Predicate Calculus . Some r e c u r s i v e devices are c l e a r l y present i n the c o n s t r u c t i o n o f well-formed formulas i n the predicate c a l c u l u s . There i s obviously no l i m i t on the number of i m p l i c a t i o n s , d i s j u n c t i o n s o r conjunctions o f predicates i n well-formed formulas (or sentences) of the pre d i c a t e c a l c u l u s . This p o i n t i s obvious i f we compare two sentence of the pre d i c a t e c a l c u l u s such as: ( E x ) ( y ) ( ( P x y implies Qxy) implies (Qxy implied Pxy)) and (Ex)(y)(((Pxy implies Qxy) implies (Qxy implies Pxy)) or (Rxy)) Moreover, there i s no d e c i s i o n procedure whereby general sentences of the pr e d i c a t e c a l c u l u s may be broken down i n t o conjunctions of smaller sentences. General sentences of the pr e d i c a t e c a l c u l u s cannot be regarded as conjunctions of smaller sentences. Based upon t h i s f a c t , one might suspect that an o b j e c t i o n to the t h e s i s that r e c u r s i v e l y derived sentences can be paraphrased by sequences of sentences none of which i s r e c u r s i v e l y derived, could be constructed. In t h i s appendix, I w i l l not t r y to construct such an ob j e c t i o n , but rather to poi n t out that any such o b j e c t i o n must f a i l u n t i l two important questions are d e a l t with by proponents of any t h e s i s such as 'the p r e d i c a t e c a l c u l u s provides the l o g i c a l a n a l y s i s of the na t u r a l languages' or 'the predicate c a l c u l u s provides the 253 grammatical deep s t r u c t u r e for the n a t u r a l languages'. Answering these questions would r e q u i r e a f u l l - s c a l e study i n i t s e l f , and the problem of recursions i n the p r e d i c a t e c a l c u l u s v i s a v i s recur s i o n s i n the n a t u r a l languages cannot be d e a l t with u n t i l such a study i s at l e a s t w e l l under way. (1) To begin with, i t appears that a presupposition of the theory that the pr e d i c a t e c a l c u l u s t r a n s l a t e s the n a t u r a l languages i s that r e c u r s i v e l y derived sentences are paraphraseable by s e r i e s of sentences none of which i s r e c u r s i v e l y d e r i v e d . Consider the problems posed by having t o t r a n s l a t e the three sentences which f o l l o w i n t o the pr e d i c a t e c a l c u l u s : a. John persuaded Jack to go to the s t o r e . b. John persuaded Jack to persuade J i l l to go to the s t o r e . c. John persuaded Jack to persuade J i l l to go to the park. Let's begin with (a). I f we l e t 'persuaded Jack to go to the s t o r e ' be covered by a s i n g l e p r e d i c a t e ' , then (a) could be t r a n s l a t e d by: a' . PJ (where 'P' stands f o r the pr e d i c a t e 'persuaded Jack to go to_. the s t o r e 1 and 'J' i s a proper name fo r John.) But then how would we t r a n s l a t e (b) and (c)? I f we l e t the p r e d i c a t e be a s i n g l e placed p r e d i c a t e over John again, then the pr e d i c a t e w i l l have to be d i f f e r e n t , as 'persuaded Jack to go to the st o r e ' i s not the same as 'persuaded Jack to persuade J i l l to go to the s t o r e ' and not the same as 'persuaded Jack to persuade 254 J i l l to go to the Park'. So we w i l l have t r a n s l a t i o n s l i k e : a 1 ' P j j b' P 2 J C P 3 J But t h i s i s c l e a r l y inadequate as a t r a n s l a t i o n , because the t r a n s l a t i o n preserves none of the s t r u c t u r a l r e l a t i o n s which e x i s t between the predicates themselves. One might as w e l l regard the n a t u r a l numbers as a t r a n s l a t i o n f o r the sentences of E n g l i s h , s i n c e given any sentence of E n g l i s h one can a s s o c i a t e i t uniquely with a n a t u r a l number (the sentences of E n g l i s h being denumerably i n f i n i t e ) . But the problem i s that given one of the n a t u r a l numbers one w i l l have no idea how i t r e l a t e s i n semantic terms to the sentences represented by other n a t u r a l numbers. S i m i l a r l y here, we would have an i n f i n i t y of predicates i n which a l l the semantic s t r u c t u r e has been o b l i t e r a t e d i n the predicates as they o r i g i n a l l y e x i s t e d , and one would have no way of knowing, for example, whether 'P-^ 2 3 9 ' i s the negation of, say "P276'> o r i s synonymous with i t , or r e l a t e s to i t as, say 'persuaded Jack to persuade J i l l to go to the s t o r e ' r e l a t e s to 'persuaded Jack to persuade J i l l to go to the park'. Nor can the problem be swept under the carpet by using two-place predicates instead of one-place p r e d i c a t e s , or any s i m i l a r devices which now e x i s t in the p r e d i c a t e c a l c u l u s . The o b l i t e r a t i o n of semantic content occurs j u s t as much (or j u s t as s y s t e m a t i c a l l y over an i n f i n i t y of predicates) when one represents "John persuaded Jack to persuade 255 J i l l to go to the s t o r e " by "P 2 John Jack" as by "P 2John". On the other hand the problem disappears by assuming the e l i m i n a b i l i t y of r e c u r s i v e l y derived s t r u c t u r e s i n paraphrase. Given the e l i m i n a b i l i t y of r e c u r s i v e l y derived s t r u c t u r e s i n paraphrase, a f i n i t e stock of pred i c a t e s can be a p p r o p r i a t e l y a s s o c i a t e d with the f i n i t e stock of 'verbals' or, simple n a t u r a l language p r e d i c a t e s . Of course even then important problems appear such as how to represent grammatical aspects l i k e tense or mood i n terms of the p r e d i c a t e c a l c u l u s . But assuming — and perhaps only assuming -- the e l i m i n a b i l i t y of r e c u r s i v e l y derived s t r u c t u r e s , can the n a t u r a l languages be represented by well-formed formulas or sentences of the p r e d i c a t e c a l c u l u s . This involves the n o t i o n of 'series of sentences', and leads d i r e c t l y to the second question. (2) Assuming that the p r e d i c a t e c a l c u l u s t r a n s l a t e s the n a t u r a l languages (or provides the deep s t r u c t u r e f o r i t ) what does the t r a n s l a t i o n of a paragraph i n the n a t u r a l language look l i k e i n the p r e d i c a t e c a l c u l u s ? Is the s e r i e s of sentences "Only one man came home l a t e . That man was George" to be t r a n s l a t e d as two sentences i n the pre d i c a t e c a l c u l u s , or one? There simply i s no theory at the moment which deals with t h i s question, and yet i t i s c r u c i a l to deal with i t , since the paragraph as opposed to sentence u n i t i s t h e o r e t i c a l l y sound 256 i n grammar and we need to have i n t e r s e n t e n t i a l l y r e f e r r i n g devices i n order to be able to e x p r e s s a l l o u r t h o u g h t s . This p o i n t has a p p l i c a t i o n i n a number o f d i r e c t i o n s . For example, one of the s u p p o s e d v i r t u e s o f t h e p r e d i c a t e c a l c u l u s reading of s e n t e n c e s i s the e l i m i n a t i o n o f truth-value gaps. Representing "The p r e s e n t K i n g o f F r a n c e i s b a l d " as "Only one man i s p r e s e n t l y K i n g o f F r a n c e . That man i s b a l d " leaves us p u z z l i n g over the t r u t h v a l u e o f t h e s e n t e n c e " t h a t man i s b a l d . " Representing i t i n a s i n g l e s e n t e n c e of the p r e d i c a t e c a l c u l u s , using some v e r s i o n o f t h e t h e o r y of d e s c r i p t i o n s , does not (or i s not s u p p o s e d to.) B u t the u n i f i c a t i o n of separate sentences i n n a t u r a l languages i n t o s i n g l e sentences of the predicate c a l c u l u s must have i t s l i m i t s , s i n c e sentences beyond a c e r t a i n length and complexity c a n only be understood by breaking them down i n t o s e p a r a t e sentences. And conversations use the i n t e r - s e n t e n t i a l r e f e r e n c e s without being s i n g l e sentences , How then can i n t e r - s e n t e n t i a l references be dropped, and i f t h e y are not dropped, how would the t r u t h -value gap problem be solved through the d e v i c e s of the. p r e d i c a t e c a l c u l u s ? The device of i n t e r - s e n t e n t i a l reference appears to be a c e n t r a l and uneliminable device of the n a t u r a l languages; u n t i l i t i s c l e a r how they are to be t r e a t e d in the p r e d i c a t e c a l c u l u s l i t t l e of s i g n i f i c a n c e can be s a i d i n a general way about the adequacy of the devices of the predicate c a l c u l u s f o r 257 the l o g i c a l t r a n s l a t i o n or deep s t r u c t u r e p r e s e n t a t i o n of the n a t u r a l languages. Once again i t should be borne i n mind that i n t e r - s e n t e n t i a l l y r e f e r r i n g devices are a sub-class (whose borders may be undrawable) of r e f e r r i n g devices i n general and no adequate account of them could t r e a t them i n i s o l a t i o n from r e f e r r i n g devices i n general. 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            data-media="{[{embed.selectedMedia}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0106613/manifest

Comment

Related Items