·Published with the finaacial assistance from the Ministry
of Ideation and Social (Telfare, Gortacent of India
FIDYA
Sen
52739
BOMBAY-7.
B.S. Kumar
1976

·То
My Teacher
Dr. N. V. Smith

·PREFACE This book is intended to provide an analysis of the syntactic structures of Sanskrit in terms of the casegrainmar which has been suggested as a modification of the standard theory of the transformational generative grammar. Despite the appearance of an alternative proposal by Anderson (1971), the proposals, in regard to the the case-grammar, of Fillmore (up to 1971), have been accepted as the model of this analysis. The alternative proposal implies generalisations by way of reducing the number of case-catcgories to the extent of making the analysis uninteresting and less ret caling at the deep level. My intention in writing this work has been to analyse the syntactic structures of Sanskrit with a view to test the validity of the thcoretical assumptions of the casegrammar, examine the methodology proposed therein and expose problems, if any. Though no claim to originality is laid in regard to the theoretical base of this work the investigation docs break a new ground with reference to the analysis of Sanskrit syntax. The chapters which discuss the case-categories (specially Locative, Extent, Path, Essive, Translative), NP-astı-NP constructions and Genitive constructions are particularly notable from this point of view. They interpret the syntactic structures of Sanskrit in a way that has not been attempted so extensively and intensively before and proposes to be stimulating in its own right. The traditional Sanskrit grammar limits the number of the case-categorics to say Once thic dichotomy between the suffixation and the cases as semantically defined primitives, is accepted, thus limitation becomes questionable. Besides, the rocative which has been left out of this analysis has a ·V111 claim to its acceptability as a distinct case-category (Kumar, 1976). It will be gratifying to me, if this book provohes scholars to take interest in the case grammar for the description of the Indian languages For all errors and imperfections that may remain, I am to blame To my teacher, Dr N V Smith, of the Department of Linguistics, University of London, my debt is incalculable. Without his teaching and gudance, it would not have been possible for me to get through the intricacies of the transformational generative grammar This book is dedicated to him as a token of tribute from one of his pupils То my teachers in India and specially those in the Department of Phonetics and General Linguistics, School of Oriental and African Studies, London University Prof RH Robins, Prof E JA Henderson, Prof CE Bazell, Dr RK Sprigg, Dr JE Buse, Mrs EM Whitley, Dr Mes T Bynon, Mrs N Vaterson, Dr D C Bennett, Miss R Kempson, from whose teaching I have benefited immenselt, express my gratitude I can mention the names of Prof DN Sharma, Vice Chancellor, Patna University and Dr R K Sharma, Vice Chancellor, KSD Sanskrit University, who hate shaped and guided my academic interests, only with a profound sense of respect Thanks are duc to the Commonu calth Scholarships Commission in the UK for the award of a Commonwealth Scholarship (1969 1972) which enabled me to complete the courses in linguistics in the University of London I am gratcful to Dr KK Mandal, Vice Chancellor, Bihar University, those interest in linguistic studies has added a neu dimension to the academic life in the uniorsitt, De LP Sinha, Pro-Vice Chancellor and Principal KDP.

·1X Sinha for the help and encouragement I have received from them To Professor J Mishra and my other colleagues in the Department of Sanskrit in the University and its colleges, and in other departments, who have given an impetus to me by sharing my interest in linguistics, I am thankful To my elders, friends and well-wishers in India and abroad, whose affection and good wishes have sustained me during the preparation of the booh, I express my deep sense of gratitude I am indebted to my friend Dr P Singh, Reader in the Department of Ancient Indian History and Archaeology, Banaras Hindu University, Varanasi, who has given much of his time and talen great care during the pointing of this booh, and to Mrs Singh for her hospitality I received during my stay in Varanası I am grateful to Dr SS Prasad of the Depactment of Sanskrit, Bihar University, who has completed the difficult job of correcting the proofs and preparing the index so meticulously To my parents and wife, oue more than I can express I am thankful to Dr SN Chaudhary, Registrar, Shri TN Jha, Finance Officer, Shri GN Jha, Sectional Officer of the KSD Sanskrit University, who have taken heen interest in the publication of this book Shri Surendra Bhargava, Proprietor, Shri CS Pandey, General Manager and other members of the staff of the Bhargava Bhushan Press deserve my sincere thanks for coping with this difficult publication so cfficiently B. S. Kumar Bihar University Vuzaffarpur ·KEY TO TRANSLITERATION Velai Retrofle«. Palatal Dental Labial t с t 1 hh р ph th ch th Consonants Stops voiceless unaspirated aspirated voiced unaspirated aspirated Nasals Fricaties voiceless Toiced d dh d dh b bh jh n n m gh N Gloital 11 h ş Ś s Labio dental у 1 Long 1 1 J'rictionless continuant r Rolled Anusi āra m Vowels Short Simplc front close half-opene back closc u ralf-close o open central half open Diphthongs ū 2 11 שם ·ABBREVIATIONS AND CONVENTIONS A ab acc С Com Cop D d Es Eit F Agentie ablative inflevon accusative inflexion Case Comitative Copula Datite dative inflevon Essile Extent Tactitise feminine Genitive construction genitive inflexion Goal Instrumental instrumental infiction Case macher Locatie locative inflexion Modality masculinc fem g GO I L 1 M nas mas ncuter N Nom nom NP 0 Noun Nominalised nominative Noun Phrasc Objectic Proposition P ·PP Pa PR pres R Re S SO T V VP X-Y X-Y X+y XY X**Y Post-Position Path Prompter present Receptive Result Sentence Source Translative Verb Verb Phrase X is rewritten as Y X is equal to Y X is not equal to Y X is transformed into Y X is not transformed into Y X is an optional element X is a case in a case-frame X contains the feature Y [X-) [X] [+Y] [-Y) Y is a positively mathed feature Y is a negatively specified feature either X OL Y is selected {x} X+Y/V-Z X 1s rewritten as Yin the environment following V and preceding Z $ Zero : absence of an clcment/constituent abbreviation for a string dominated by a constituent X, where the detailed constituent- Structure below X is not seles ant indicates doubt regarding acceptability, indicates unacceptable construction 7 ·CONTENTS Page X ܝܕ Preface V11 Key to Transliteration Abbreviations and Conventions xi-x11 Contents X111-11X Chapter I Introduction 1 I.1. Sanskrit language and the scope of the present study 1.1.1 Sanskrit language 1 1.1.2 Sanskrit language and the views of Indian grammarians 2 1.1.3 Sanskrit language and the views of western scholars 4 1.1.4 Indian and western scholars and linguists on Sanskrit syntax 5 1.1.5 scope of the present study 9 1.2.1. Theoretical assumptions of FCG 10 1.2.1.1. works of Fillmore 10 I 2.1 2. the notion of case 10 1.2.1 3. Fillmore's case-grammar. 12 1.2.1.3.1. descriptive adequacy and FCG 13 I.2.1.32 centrality of syntar and covert categories in FCG 13 1.2.1.33 ordering and base-component in FCG 14 I 2.1.3.4 base-component and case in FCG 15 1.2.1.3 5 notation in FCG 18 1.2 1.4 summary of principles 21 ·xiv Page 1.2.1.5 ICG as a framic-stork for thic present study 22 1.2.2. limitations of thic present study 22 1.2.2.1 caia of the present study 23 Chapter II Casc.categories A, D, R, O, F, 1, 1., SO, GO, Ext, Pa 24 II.1 casc-categorics A, D, R 24 11.1.1. A, D distinguished as casccategories 24 11.1.2. R as a casc.category distinguished fron A/D 34 11.1.3. dcfinitions of A, D, Expcricncer as given by: Fillmorc 36 11.1.3.1 Experiencer as a casc-catcgory 37 11.1.3.2 Experiencer and D as defined by Fillmore compared with D, R as specificd in II.1.1 and II.1.2 40 11.1.4 intentional and non-intentional involvement of an actant 41 11.2 casc-categories O, F and cognato ccts 43 II.2.1 case-category O 43 II.2.1.1 O dcfincd 43 II.2.1.2 O in relation to different syntactic structures 47 II.2.1.3.1. sentences containing O and the process of passivisation with subjectivalisation 61 11.2.1.3.2 sentences not containing O and the passivisation with subjectivalisation 63 process of ·Page II 2 1 4 sentences with cbligatory deletion of A/D and the process of passivisation in relation to O 67 II 22 F/Re and cognate objects 69 II 221 case category F/Re 69 II 2 2 2 justification of F and the status of cognate objects 70 II 2 23 F as a distirct case category further examined 77 II 3 case category I defined 79 II 31 I distinguished from A/D/R/O/F 83 II 3 2 1 process of subjectit alıcation ard I 87 II 322 process of paceivisation ard I 90 II 331 NP's indicating natural forces and I 94 IT 3 32 I as a complex structure 96 II 34 NP s indicating emotie states and I 97 II 35 NP's dernied frcm ccmplex stru ctures and intentional/ncn intent ional unicitement of an actant 99 II 351 NP's derived from complex structures ard I 99 II 352 I and Laloff s positicn 105 II 353 intenticnal/non interticral in Solvement of an actant and I 108 JI 4 case categories L, SO, GO, Eit, Pa 110 II 41 Sanskrit grammarians on categories related to time/spacc 110 casc ·Xy1 Page II 42 linguists on case-categories related to time space 111 II 43 case-category in relation to the distinction between time and space 112 II 44 L as a case-category 114 1 441 L as a complex structure 116 II 4 42 L with different surface - rcalısations 117 II 443 L as the only case-category in a case frame and the process of subjectivalisation 119 II 4 5 SO as a case category 120 II 4 5 1 SO and NP's with ablative inflexion 122 II 4 5 2 SO as a complex structure and equating SO and A in certain constructions 127 II 46 GO as a case category 130 II 461 GO with different surface-realisations related to semantic distinct- 100S 132 II 462 GO and NP's with dative inflexion 133 II 4 7 Ext as a case-category 136 II 471 NP's derived from complex structures containing Ext 138 II 472 Ext with different surface realisatıons related to semanticdistinctions 139 II 48 Pa as a case category 141 II 49 case-categories related to time! space and Fillmore's views thercon 142 ·X1 11 Page 145 Chapter III NP-asti-NP Constructions III 1 III 1 1 III 12 III 1 3 III 14 INI 15 III 2 Sanskrit grammarians and linguists on NP-asti-NP constructions 145 NP-asti-NP constructions and the case-category Es 148 T and Es 152 L and Es 156 GO and Es 156 Ext and Es 157 NP-asti-NP constructions and Fillmore 158 the status of copula in NP-asti NP-constructions 161 copula and the derivation of NP asti-NP constructions 165 process of nominalisation and NP asti-NP constructions 170 rules operating in the derivation of NP asti-NP constructions 189 III 3 III 4 III 5 III 6 Chapter IV Genitive Constructions 191 IV 1 IV 2 on IV 3 IV 31 Genitive construction and genitive inflexion 191 Sanskrit grammarians Genitive constructions 192 types of Genitive constructions 194 nominalisation and Genitive constructions 197 derivation of Genitive constructions related to nominalisation 198 IV 311 ·Page IV 32 Genitic constructions ind zlicnablc and inalienablc re- Intionships 214 IV 321 Tillmore's schema and alıcn2- ble and inilicnablc rclation sluips 217 IV 3211 Genitric constructions and inalienable relationslup 217 IV 32111 derivation of Genitic constructions with inalicnable rclationship 219 IV 3 2 1 1 2 dernition of Genitive con structions with inalicnable relationship 221 IV 321 13 dernation of Genitive con structions with inalienable relationship 227 IV 32 11 4 derivation of Genitive con Structions with inalienable relationship 230 IV 32 11 5 derivation of Genitite con structions with inalienable relationship 240 IV 321 1 6 Genitive constructions taining directional indicator and inalienable relationship, and Ross's Performatite analysis 241 IV 32 117 Genitive constructions with inalienable relationship and directional NP's 244 con ·Page for case IV 322 schema for Genitive construction with inalienable relationship 246 IV 323 schema Genitive construction with inalienable relationship and the notion of 246 IV 324 Genitive construction with alienable relationship 247 Chapter V Lexicon and Rules 255 V1 lexicon in FCG, lexical entries for verbs and rules 255 V 11 lexicon in FCG 255 V 111 FCG and prior selection of V’s, subject object relation 255 V 112 lexion and frame features 256 V 113 specification of lexical entries 257 V 12 lexical entries for verbs and their case frames 258 V 13 rules Phrase Structure and Transformational 265 V 131 PS rules 265 V 1 3 2 T rules 266 Chapter VI Conclusion 269 Bibliography Indes 281 273 ·I INTRODUCTION I. The study of linguistics, as it has developed since the publication of Syntactic Structures of N. Chomsky (in 1957), has made it imperative, to take a fresh look at Sanshrit syntax, and examine the validity of the theoretical assumptions and the methodology involved in the post- Chomskyan linguistics. There is an attempt, in this study, to discuss the nature of case-grammar as proposed by Fillmorel and to analyse, within its framework, the syntactic structures of Sanskrit. 1.1. This section deals with the following points : in I. 1. 1. reference made to the Sanskrit language (the classical language), as distinguished from the Vedic language for Vedic Sanskrit); in I. 1. 2. and I. 1. 3. the investigations and studies of the Sanskrit language and grammar by the Indian grammarians and the western scholars respectively, are referred to; the treatment of case in relation to syntax, is taken up briefly in I. 1. 4, and the scope of the present study is discussed in I. 1. 5. and I. 2.2. 1. 1.1. Interest in the Sanskrit language whose precise antiquity still remuns unsettled,” was revived by the European scholars towards the end of the eighteenth century: 3 15 1. 2. Fillmore, C J., (1968a) Burrow (1955, p. 31) assigns the period 1200-1300 B C. to the composition of the Rgveda which is not likely to be many centuries out, either one way or the other.' Winternitz (1927, pp. 290-310), cn the other hand, mentions a period in the middle of the third millennnm BC Burrow, T., (1955), P6 3.

·2 SANSKRIT SYNTAX The Sanskrit language spoken' and understood today, is strictly the classical Sanskrit distinct from the language of the Vedic texts, which is called the Vedic Sanskrit or Vedic language The classical Sanskrit language and the Vedic Sanskrit language are also referred to as 'bhāsā 5 and ‘chandas' respectively. It is maintained that the structure of the Sanskrit language (henceforth classical Sanskrit will be mentioned as such) assumed rigidity, after Pānını,? deviation from which could not be tolerated Scholars hold the view that the Sanskrit language remains, even today, as fixed by Panini's rules 8 I 1 2 The Sanskrit language has been studied in depth by the ancient Indian grammarians 9 The initial motivation was, most probably, the preservation of the sacred and ritual texts through oral tradition which made the study of phonetics imperative for them 4 5 6 7 There ace 2 460 speakers who accept Sanskrit as their mother tongue, accordırg to the Gazetteer of Irdia Vol 1 (1065), apper dix VI Panini rules II 2 108, VI 1 62,1V 1 181, VI 3 20, VIII 2 93 Panını rules 1 2 36, 1 4 9, 20, 81, II 33 62 CF Staal, JF, (1967), P 18 His grammatical treatise, Astadhyayi, corsists of about 4,000 rules (Pandega (1938) lists 3982 ni les) This work is assigned the period 700 600 BC by Goldstucher and Bharda rahar Burrow (1955, p 48) mentions 400 B C as the period of Panini s work Thus Panini's work has raciously been assigoed to around 600 B C and around 300 B (Robins, (1067), p 137) Burrow, T, (1955), p 36 For details Belvalkar, S K ,(1915), Cf Allen, WS, (1953) p 3 he quotes J R Firth 'without the Indian grammarians and phoneticiacs it is difficult to imag ne cuc n netceosh century school of phonetics 8 9 ·INTRODUCTION 3 'Niruhtam’, the ctymological composition by Yasha,10 embodies the 'carliest systematic discussions on questions of grammar' 10 This system of grammatical analysis reaches perfection in the hands of Pānını.11 The grammarian Kātyāyana,1. also known as the maker of notes (Vārttihahara), and Patañjalı, 13 the writer of the Mahabharya' (the great commentary) follow after Pānını and subject his grammatical work to their scrutiny There has been much controversy as to how Katyāyana and Patañjalı have assessed Panini's work 14 The grammarians who followed the above trio, did not contribute to the grammatical system by way 10 11 12 13 Burrow, T, (1955), p 48 Cf Sarup, L , (1920), p 54 his period is mentioned between 700 BC-500 BC Bloomfield, I , (1935), p 11 one of the greatest moruments of human intelligence 'it describes every syntactic usage of its author's speech' His period is considered to be near about 350 B C Cf Belvalkar, S K, (1915), pp 28 29 Scholars are definite that his period is sccord century BC (Burrow, T, (1955), p 49), cf Robins, RH, (1937), p 140, cf Belvalkar S, (1915), p 32) Goldsturker, T, (1961), pp 119 121 'Katyayana does not leave the impression of an admirer or friend of Panını, but that of an antagonist, often too, of an erfair antagonist Patañjali often refutes the strictures of Katyayana and defends Panini CF Kielhorn, F, (1876), p 52 Katyayana Is a follower and judicious admirer of Panini' 'Patanjali has defended Panini from some of the objections brought against him by Katyayana', but 'in many cases his criticism is much more thorough going and destructive than Katyayana's and Panini has suffered more at his hands than at those of the Varttihakara '(1 e Katyayana) 14 ·4 SANSKRIT SINTAX of facts or methodology.15 The only exceptions appear to be Bhartrhari and Vāmana Jayāditya The work of the former16 is more concerned with the metaphysical aspects of grammatical analysis, and that of the latter17, the carliest commentary aftec Mahābhāsya on Pānını's work, furnishes clamples for each of Panini's rules.18 The later grammarians are better to be regarded as commentators Ramacandra,19 the writer of the commentacy Prahriyāhaumudi', explains and rearranges Pānını's rules His method is closely followed and improved upon by Bhattojidiksita,20 the writer of the 'Siddhāntakaumudi' The latter work has virtually ousted, for pedagogical purposes, Panini's work from the curriculum 21 I 1.3 Among the western scholars 22 who have taken up and encouraged the study of the Sanskrit language and its grammar, the first is Sir William Jones,23 thougb Macdonell refers to two missionaries (1968, 1732) who were acquainted with the language 24 The earliest Sanskrit grammars published in English include that of V. Carey (1806, Serampore) and of C Burrow, T,(1955), p 50 they 'presert the material corta ced in Panını they contain little that is origiral' 16 Bhartrhası (700 AD), (ed ) Abhyaahar, KV 17 Vamana Jayaditya, (650 A D ), (ed ) Sbasta), S D D rd Pt KP Shukla (1955). It may be noted that Panını himself did not give examples for his rules CE Misra, V N,(1966), p 212 His period is 1500 A D , (Macdonell, 1911, p xi) His period is 1500 A D ,(Macdonell, 1911, p xu) Belvallar, S K ,(1915), p 46, al o Sascoa B R (1951), p 14 For details see Pedersen, H ,(1931) and Lehmann WP, (1967). Jones, Sir William, (1786) 'the Third Arousersary Discourse on the Hindus' in Lehmann, VP, (1967), pp 10 12 Macdonell, A A, (1911), p 4111 15 18 19 20 21 22 23 24 ·INTRODUCTION 5 Wilkins (1808, London) The three most notable figures among the Europeans are · (1) F Bopp, who was the first to attempt a comparative philological study of Sanskrit,25 (11) T Benfey, who investigated the language of both Vedic and epic tests, 25 and (111) V. D. Whitney who was the first to write a historical grammar of Sanskrit by treating the Vedic language 'more fully and explaining from it the development of classical Sanskrit.26 I 1 4 The Indian grammatians and commentators, in their analysis of cases, more or less followed Panini's rules which considered cases (hārahas) to be grammatical relations of underlying structures, and did not identify these with case-inflexions However, the western scholars of the nineteenth century missed this point in their analysis of the Sanskrit language The reason was that they followed the traditional approach in which case-infleuon was considered the fundamental criterion of distinction between noun and predicate (or verb) and the category of case became restricted to nominals They described the different case-forms in relation to their uses 27 The first to be different in his approach from the above, appears to be the English grammarian James Harris (eighteenth century) 28 25 26 27 Macdonell, AA, (1911), p vi Robins, RH, (1967), p 148 Pedersen, H ,(1931), pp 1922, Macdorell, 1 A ,(1911) p x1, Lehmann, W P, (1967), pp 38-45 Robins, R H ,(1967), pp 26 35, 50 52, 59 Robins, R H ,(1967), p 154, Hairis advocates 'idenufcation of those universal categories of grammar ard relations', be cause 'the same furction tras served by care irflexiors in Latin ard by prepositions in English 28 ·6 SANSKRIT SYNTA 31 Thus, Wilson in lus Sanskrit grammar, refers to eight cases, nominative, accusative, instrumental, dative, ablative, genitive, locative and socitive 29 Monier-Williams thinks that 'Sanskrit syatay, unlıhr that of Greeh and Latin, offers fewer difficulties "30 Perry devotes eight pages to nominal declension only, and mentions nothing about cases Edgren refers to cases and also to the 'function of cascs' in a footnote 32 According to Macdonell 'the absence in Sanskrit of the indirect construction is in itself a tolerably clear proof that the structure of complessentences in Sanshot must be far less involved than in Latin or Grech' He refers to eight cases and adds that 'the syntactic arrangement of the Sanskrit sentence is primitive and undeveloped, as compared with Greek and Latin "33 Thus, all these scholars appear to identify case-infiexion with cases, and, at best, they refer to uses or functions of the cases for which the most comprehensive analysis with extensive citations from the tests, appears to be that of Speijer 31 The Indian grammadians of the nineteenth and twentieth centucies follow, more or less, either the principles of the western traditional grammatical analysis or the system of the traditional Indian commentators 35 29 30 31 32 33 34 33 Wilson, H H , (1847), pp 377-402 Monter Williams , M , (1877), pp 354 361 Perry, E D (1885), pp 83 91 Edgren, H ,(1885), p 37(f n 90) he refers to eight cases, and in his treatment lays emphasis on morphology Macdonell, A, (1911), pp 178 209 Speijer, J S, (1886), pp 24 113 Taraporen al? I JS,(1967), this contains his Wilcon Philological lectures delisered at the Bombay Uansersity in 1937 Apte, VS, (1921) hale MR, (1918) ·INTRODUCTION 7 However, the period from the thirties of the present century, has witnessed a revival of interest in the work of Pānını, in the wake of advances both in theory and methodology of linguistic studies, and, consequently, scholars have better appreciation of the case-system in Sanskrit (an indication of which is given by Faddegon36) as explicated through the rules of Pānını Recently, linguists like Staal and Kiparshy, in their wouls, have re-analysed some aspects of Sanskrit syntax on the basis of the rules of Pānını Staal has discussed word-order in Sanskrit which he accepts, on the basis of the works of ancient Indian grammarians, to be free37, and advocates a word-order-free categorical component as a linguistic universal The word. order in Sanskrit which Staal cites as a support for having base-rules free from sequential ordering, has also been studied by scholars like Delbruck and Gonda who have analysed selected Sanskrit texts Their studies, however, have ended as statistical assessments which reveal little significant about the word-order in Sanskrit.38 36 Taddegon, B , (1936), P 18 referring to Tanini s karaba' theory, he says "evidently Panini tries, in this analysis, to separate the ideational aspect from the linguistic expression, an attempt which the occidental linguists of the latter half of the pigeteenth century have condemned (he refers to the criticism of Whitney b), misted as they were by the hope of being able to understand language through the cxculsive study of its phonal and morphological aspect, I c , its articulatie uiterance and the association system underlying declension and conjuga tion, as if, the application ard imitation of physics and a mecha nistic psychology were the last word of moral science Staal, J P, (1967), pp 16 50 As discussed in Staal (1967), p 59 37 38 ·8 SANSKRIT SYNTAX Indian scholars, like Lahıcı, also have undertaken such analyses, but their conclusions remain vague and, at best, mantain that the logical natural sequence of ideas is expressed in word-order 39 The possibilities of sequential ordering that such studies suggest, amount to 'saying yyz and yyz occur, but on the other hand, yız, yzı and zyı also occur' 40 Due to this characteristic of free word-order in Sanskrit, F(illmore's) Case) Grammar) commends itself as a framework for the analysis of Sanskrit syntactic structures Because, as pointed out later (I 2 1 3), FCG does not assume a sequential ordering in the base component As against FCG, Staal's proposal for a set system in the categorial component for making it universal" (as opposed to Chomsky's concatenation-system), does not appear to be acceptable for Samshtit As Staal retains the constituency analysis (ie S->NPVP), it is not clear how this VP is to be specified in Sanskrit unless some sort of word-order is implied Since, for the analysis of a language lihe Sanskrit, the choice remains between EST (of Chomsby) and FCG, the one proposed by Staal does not appear to be acceptable 4. Secondly, Staal appears to identify 39 Lahırı, PC, (1933) P 372 it is generally said that an in Aexional language like Sanskrit is almost absolutely free in its word order, the have found that usually the word-order follows the thoughts as they occur to the mind of the speaker Their order is occasionally disturbed for special reasons matnly to lay emphasis on a particular word' Dover K J (1960) p 1 Staal J r (1967) pp 60 90 42 CE Seuren, P AV, (1969), pp 173 174 he does not raise this point but questions Staal s position on some other ground ‘it has not app-ared so far that an unordered nucleus expan sion of the type proposed by Staal is useful or necessary for English, as opposed to languages with free word order Or writer 40 41 ·INTRODUCTION 9 case with inflexion, as he postulates a rule such as NP > {N, Num, Cas, (S)}43, despite his acceptance that case (käraha) - relation in Sanskrit is a deep grammatical relation. Thirdly, his treatment of Sanskrit syntax is confined to word-order only and does not go beyond the analysis of ten sentences. In view of these, Staal's framework and treatment of Sanskrit syntax appear to be highly inadequate. Kiparshy, along with Staal, has investigated the syntaclic semantic relations of Sanshat as accounted for by Panini's rules 44 But this treatment also appears to be inadequate for more than one reason In thc first instance, their treatment is not concerned with Sanskrit syntax in general, rather it shows to what extent Panini's rules capturc thc generalisations of Sanskrit syntactic structures Secondly, though it mentions genitivc, it does not discuss genitric constructions Thirdly, certain constructions which arc discussed in chapters II, IV of this study, are not discussed by them. Thus, their analysis of Sanskrit syntay, covers a highly restricted ground. I. 1.5 The present study may, thus, bc considered as the first attempt at an analysis of Sanskrit syntax within a transformational generative theoretical model The study is carried on within the framework of TCG in vicw of ICG's characteristics as specified in 1 2 132 and 1 2 1.3 3 No attempt is niadc, in this study, to present a comparative study of Pānını’s ‘hāraha'-theory and TCG 45 In fact, the cases of ICG and the härahas nccd not bc idcntrfied with each other, and, as such, this problem is considcrcd beyond the scope of this study, 43 Staal, J 1, (1967), p 65 Kiparsky, P2djr Staal, (1969) Ananthararayara II S, (1990) 45.

·10 SANSKRIT SYNTAX I. 2. This section contains two parts : 1. 2. 1. deals with the theoretical assumptions of Fillmore's case-grammar, and section 1. 2. 2. refers to the scope of the present study. 1. 2. 1. In this section, it is proposed to present the theoretical assumptions of case-gramma proposed by Fillmore. I. 2. 1. 1. The framework of FCG can be constructed on the basis of his undermentioned works : 1. 'A Proposal Concerning English Propositions' (1966) 2. "Towards a Afodern Theory of Case' (1966b) 3. "The Case for Case' (1968a) 4. 'Lexical Entries for Verbs' (1968b) 5. "Types of Lexical Information' (1969) 6. "The Grammar of Hitting and Breaking' (1970a) 7. "Subjects, Speakers and Roles' (1970b) 8. 'Some Problems for Case-Grammar' (1971) Of these, the third expounds FCG in detail and still remains the standard treatment, and the eighth sketches the present state of the theory, though regrettably in a very fragmentary manner, in the light of the inadequacies that have suggested themselves during the intervening period. However, it is assumed that the third and the eighth, between them, cover the exposition of FCG. The discussion in the succeeding chapters will be based on this exposition. I.2 1.2. The notion of case has been treated variously by grammarians and linguists, and equated with either ·INTRODUCTION 11 (1) 'semantic functions of inflexional affixes on nouns or the formal dependency relations which hold between specific nominal affixes and lexico-grammatical properties of neighbouring elements'; or (11) 'a statement of the morphophonemic reflexes of a set of underlying syntactic relations which themselves are conceived independently of the notion of case''46 In the first category, comes the description of case-systems of languages like Latin, Greek, and Sanskrit; and which have been accepted as the basis for the treatment of case-systems of other languages. In the second category, comes the treatment of case in some varieties of generative linguistics which identify 'case' with the inflexional realisation of particular syntactic relationships in the surface structure. 47 Some scholars have insisted on its being related to inflexion.49 46 47. Fillmore, C), (1968a), p. 2 Chomsky, N ,(1965): 'specified feature (case) is introduced by a rule that does not belong to the base sub-component of the syr tax at all but rather to its transformational pari" (p. 172),... the features involved in the case-dimersion are certainly added to a formative by rather late transformations (since case often depends on aspects of surface rather than deep structure (p. 177), case is usually determin by the position of the noun in surface-structure rather than in deep-structure' (p. 221 in 35). Cf. Ljors, ), (15C6), p 218. Case (in the language in which this category is to be fourd) is not prescrt in deep structure at all, but is merely the inflcxioral realisation of particular syntactic relatior tips. Cassidy, T. G, (1937), p 245 'Case will be picperly wed and Si continue to Fave some mcaring only if the association with inflexion be fully recognised.' Cf. Istmirr, 1.P. (1958), p. 167. 'a particular casc us roncxistent unless It is represeried by forms hich cor1123t in a system with others 48 ·12 SANSKRIT SYNTAX 1. 2 1 3. Fillmore, on the other hand, agrees with Hjelmslev in not accepting case as an 'expression in the form of affises on substantives'49 and uses case to identify the underlying 'syntactic-semantic relationship’ as a deep structure phenomenon in a natural language For Chomsky, Fillmore's case-grammar as 'a semantically justified universal syntactic theory'80 is not considered basically different from the earlier version of syntas as put forth in Aspects at the conceptual or theoretical level, but rather at the level of application to the particular grammar of English' 51 Within transformational generative grammar, basically three assumptions about FCG are possible : (1) it presents a theoretically different grammatical description which meets both descriptive and explanatory adequacy, (11) as an evaluation procedure it is to be preferred to other types of grammac for a natural language, (111) it is a grammar which is specific to English Chomsky has not argued that case-grammar is not applicable to other languages, and none of the linguists has given any argument to prove that casegrammar is language specific, and the plea that it is a "choice of grammar?52 for English, is untenable of the remaining two assumptions, even if the second assumption is demonstrated to be correct, FCG makes a significant contribution However, it will be examined below whether it meets the first assumption of descriptive adequacy or not. 49 Fillmore CJ ,(1968a), p 21, also Blahe, FR, (1930), p 35, 'the term case form will be employed for case in its usual sense, while the term case will be reserved for the relationship, itself without regard to its means of expression Fillmore, c ), (1968 a), p 21 Chomsky N, (1970) P, 38 Chomsky, N, (1957), pp 50_53 50 51. 52 ·INTRODUCTION 13 I 2 1 3 1. А grammar is descriptively adequate to the 'extent that it correctly describes the intrinsic competence of the idealised native speaker '53 Thus a descriptively adequate grammar is not concerned with a limited corpus, rather, it accounts for all the grammatical sentence-types as distinguished from utterance-type:54 which a speaker-hearer produces and understands on the basis of an 'internalised system of rules A grammar G, is more valuable than a grammar G, and reaches a higher degree of descriptive adequacy if it fulfils four conditions 55 (1) the rules (1 e description) are so explicit and exact that there is no confusion with regard to what is generated and what is not generated, (11) it gives only and all the combinations or strings of formatives which are sentences of a language, (111) it has more generalisations and expresses them with a smaller set of rules. In other words, it is as simple as possible 56 A description may be called simple if it mahes relevant generalisations and expresses regularities in the language concerned desciiptions teach the same level of generalisations, the description with the smaller set of rules and/or symbols is simpler, (iv) it is able to associate with a regular and straightforward semantic interpretation of these, the first two conditions are better left for verification when the description of the particular language is taken up The remaining two cond tions are fulfilled in FCG, as evident from the following discussion I 2 1 3 2 FCG assumes (1) the centrality of syntax wherein 'the form of words (formatives) are specified with If two 53 54 55 56 Chomsky, N, (1965), p 24 Seuren PAM , (1969) P 2 Seuren PAM , (1969) P 12 This condition of a grammar being 'simple, but ron trivial at the same time is difficult to be met ·14 SANSKRIT SINTAS respect to syntactic concepts' and (11) specifies the significance of covert categories' which exhibit the grimmatical properties on the basis of 'selectional restrictions and transformational possibilities '57 even if the grammatical properties are not morphemically realised FCG59 accepts the distinction between decp structure and surface structure, and trcats the base component (excluding the lexicon) 15 language-unis ersal, and not as language-specific FCG assumes a base component as a 'semantically justified syntactic deep structure, not as a syntactic deep structure of Chomsky 59 It advocates that the grammatical notion case60 deserves a place in the base-component of the grammar of every language 61 I 2 1 3 3 FCG assumec ? basc component without any sequential ordering of items This is opposed to the concatenation-system of the standard theory which implies a sequential ordering in the base-structure 62 Linguists have disergent viens in this respect Halliday would favour a sequence free base component and Lyons would argue for leaving this question for empirical investigation Agreeing with Lyons, even if the question 57 Filmore, C. (1968a), p 3 58 Fillmore, C ), (1968a) p 21 Fillmore CJ, (1968a) p 38 Cf id (1971), p 245 does not propose 'to eliminate deep structure altogether, to find a level of syötactic structure which tas deeper than that offered by the then standard theory Fillmore CJ ,(1971), p 245 'deeptructure cases their existence could be discovered and justified by syntactic criteria cf (19682), p 5 'semantically relevart spetactic relationships Fillmore CJ (1968) P9 Chomsky N (1965) pp 194–127 Halliday, MA (1966), p 227 Lyons J. (1966), p 227 63 59 FCG but 60 61 62 63 ·INTRODUCTION 15 is left open in respect of natural languages in general, and we do not concede sequential ordering to surfacestructure apart from 'stylistic inversion' (which Chomsky concedes to be a surface phenomenon), a sequence-free deep structure appears inevitable for languages like Sanskrit, Greeh, Latin. In the case of languages like German and Russian, linguists have tried to show that some order in the deep structure needs to be accepted.6.1 It appears that in order to meet the criterion of descriptive adequacy, a grammatical description of a language lihe Sanskrit, should have a deep structure free from word order phenomenon I 2 1. 3 4 According to FCG, the base-component specifies that a sentence consists of two constituents A(odality) and P(roposition) 65 The Modality) includes 'negation tense, mood and aspect. sition) is a 'tenseless set of relationships' involving a Verb) and one or more Nſoun)P(hrase) 67 The rela- '68 The P(ropo- 61 65 66 Chomsky, N , (1965), p 126 Ross, JR, (1970), pp 257-258 Fillmore, C J, (1968a), p 23 Tillmore CJ, (1968a), p 23 Cf Loyns, J (1966), PP 218, 223 he regards negation, tense, mood as associated directly 11th the sentence as a whole, and the perfect and progressive aspects as features on the v CE Robinson, J J, (1969), P63 she brings in interrogation under M(odality), but remos es negation and aspect from it Tillmore, C J ,(1968a), p 23 he later on (1971, pp 246 47) substitutes V with Predicator (verb, and adjective or noun). This revision in the constituents of P is unfortunate, as it is proposed without any discussion for or against In this study P as exparded abos e (1968a) is accepted 67 ·16 SANSKRIT SYNTAS tionships that the NPs contract with the V, are rcalised as different Cases) Rules (1) and (11)68 formalise thus (1) SM Pro (11) PVC, Cn A Case) dominates an NP, K(asus) (case inflexion) and P(ost) Position), which can be indicated by a rule lile (111) (111) CNP -K(+PP ) Fillmore has postulated (19682, 1969, 1971), the following Case)s with the semantically based definitions given 1 A(gentis e) 'the case of the typically animate perceived instigator of the action identified by the verb' (1968a), p 24, 'the instigator of the event' (1969), p 116, 'the principal cause' (1961), P 253 2 Dative) 3 E(xperiencer) 'the case of the animate being affe cted by the state or action identified by the serb' (1968a), p 24 'the entity which receives or accepts of experiences undergoes the effects of an action' (1969), p 116, then there is a genuine psycholo or 68 69 rotation 15 Fillmore C ) (1968a) p 24 FCG proposes a notation like the arrow used throughout but this should rot be interpreted as meafir8 that the proposal for a case grammar requires an assumpt on of a left to-night orientation of the constituent symbols of the Ten nating rules (Fillmore CI (16a), p 24 (f = 20) Fillmore CJ (1968a) p 33 he proposes a rule like C+k+NP perhaps in vier of the English larguage The PP Occurring in certain constructions of Sanskrit impo es a spec fc rcal sa tion of a 70 ·INTRODUCTION 17 gical event or mental state verb, we have the experiencer"...(1971) p. 251. 4. Instrumental) : 'the case of inanimate force or object causally involved in the action or state identified by the verb' (1968a), p 24, 'the case of the immediate cause of an event, or, in the case of a psychological predicator, the stimulus, the thing reacted to' (1971), p. 251; "the stimulus or immediate physical cause of an event' (1969), p 116. 5 Counter-agent 'the force or resistance against which the action is carried out' (1969), p 116. 6 F(actitive) 'the case of object or being resulting from the action or state identified by the verb, or understood as a part of the meaning of the verb' (1968a), p 25. R(esult) : 'the entity that comes into existence as a result of tlie action' (1969), p 116. 7. O(bjective) : 'the semantically most neutral case, the case of anything representable by a noun whose role in the action or state is identified by the semantic interpretation of the verb itself; conceivably the concept should be limited to things which are affected by the action or state identified by the verb. The term is not to be confused with the notion of direct object, nor with the name of the surface case synonymous with accusative' (1968a), P 25, 2 ·18 SANSKRIT SYNTAX in 'the entity that moves or changes of whose position or existence 15 consideration' (1969), p 116, 'the Object case 15 that of the cntity which moves or which undergoes changes, a vaste-bashet, sentences embedded to Objects can serve to identify the content of a psychological event as with the verbs of judging or imagining' (1971), p 251 8 L(ocative) 'the case which identifies the location or spatial orientation of the state or action identified by the verb' (1968a), p 25 9 SO(urce) 'the place from which something moves' (1969), p 116, 'starting point' (1971), p 259 10 GO(al) 'the place to which something moves' (1969), pp 116, 'the destination (1971), p 259 Tillmore has suggested additional cascs like (1) Com(tatue) (1968a, pp 81 83), (11) B(enefacune) (1968a,p 32), (111) Timc) (1968a, P 32), ('\) Es (sise) (1968a, P 84), (1) r(ranslatnc) (1968a, p 84), (1) Pa(th) (1971, pp 259 260), and possibls (111) T'» (1971, p 260 M) Nariempi as made, hert, 10 235c35 the justifiability or validity of ond or all of the cases mentioned abouc This question is taken up in Ilic sections concerned with spec fic case categories 1 2 1 3 5 of the cases, ID, , r, I, L, SO, GO, Ext, 1a, Ls, T are postulated in thic succreding chapters Cinc categori Ricap inc) is introduc-d The status of Ilic ·INTRODUCTION 19 V C Com (itative) is not cvamıncd, as It is linked up with coordination Other case-categories are discussed, but not retained 71 Now, for the present study, the Phrase Structure rules (1, 11 above) can be restated as follow (1) S MP (11) P Cn where C.=one of A, D, R, O, F, I, L, SO, GO, Ext, Pa, Es, T TCG muntains that the case frame(s) of a verb indicate(s) the relationship (s) in the form of cases that contract between the NP(s) and the V Thus, a verb like 'pathati' (reads) has a case-frame like pathati + [ A, (I), O (L) — ] The dash indicates the place where the verb occurs, the symbols to the right of it, show the cases that can co occur with the serb The parentheses indicate optionality The symbol '+' in front of the square brachets, indicates that the frame feature is positively specified with regard to the lexical item'pathati' According to FCG, the relationship obtaining between a V and the associated NP's in deep structure remains the same, irrespective of the final position of these NP's in surface Structure Thus, the case frame of the verb 'open', to take an example from English, is as follows Open + [ - O, (A), (I) ] (1) the door opened 71 The case category Coucter agert (1 134) is excluded from this discussion Because Fillmore does rot discu s it (1969), and he appears to have rejected it later on ·20 SANSKRIT SYNTAX (11) he opened the door A o (111) the key opened the door I O (iv) he opened the door with the key A O I [ Similar examples can be cited from other languages like Sanskrit : bālakah-pustakam-pathati А bālahena -pustaham -pathyate] A 0 In addition to the categorial component,72 the basecomponent of ICG comprises a lexicon As regards the selection of leucal items, FCG proposes that the selection of the verb should precede the selection of other levical items Secondly, FCG maintains that subject-object celations are surface-phenomena *3 Consequently, there does not appear to be any need to distinguish between strict sub-categorisation rules and selectional restrictions 74 As a result of thcsc, thc lcricon assumes a less complex form Besides, rCG proposes that onc case-catcgory occurs only once in a simple sentence, and occucrences of the same case-category can be conjoined in relation to a single verb " In so far as the sentence-embedding is concerned, 73 72 73 74 75 For Phrasc Structure rulcs sce V 3 1 or the details sce VI Cf Chomsh; N. (1965), pp 95, 113 Fillmore, CJ ,(1968a), p 21 although there can be compound instances of a single case (through roun phrasc conjunction), cach case rclatiorship occurs only once in a simplc sentence + ·INTRODUCTION 80 21 FCG makes a change in its earlier position78 and proposes that embedded S is dominated by some case-category.77 I 2 1.4 In sum, the FCG which, according to Its proponent, 'should be incorporated into the theory of transformational grammar',78 lays down the following principles for description (1) Case-systems are universal 79 (11) Cases are semantically justified syntactic primitives (111) 'Case-form'81 or 'surface-cases ’82 are language specific and may be realised by inflexion, pre-/post-positions, and other 'syntactic function indicators'82 like word-order. (1v) In a simple sentence, one case-relationship occurs only once 83 (v) Relations of subject and object are surface phenomena 84 (vi) Only noun-phrases representing the same case can be conjoined with a single verb 76 77 78 79 Fillmore, C ) , (1968a) P 32 complex sentences involve recursion through the category S(entence) ur der the case Category Objective' Fillmore, C ), (1971), p 247 Fillmore, C ) , (1971), p 245 Fillmore C J. (1971), p 247 'the cases identify the roles defined once and for all for human languages Fillmore, J ,(1968a) p 5,(1971), p 245 Fillmore, CJ, (19682), p 21 Fillmore, C J. (1971), p 47 Fillmore, C), (1971), p 248 'one instance per-clauseprinciple' Fillmore, C J, (1978a), p 32 (1971), p 247 80 81 82 83 84 ·22 SANSKRIT SYNTAL are (su) Sentences embedded in underlying representation are embedded as occupants of some case-role °85 I 2 1 5 FCG as specified, in brief, above, may need refincments and a detailed formulation of its rules both universal and language specific ones Its application to the analysis of natural languages may bring out imperfections which need to be rectified In fact, Fillmore is aware of these possibilities 86 Despite these reservations, FCG stands as a sufficiently detailed working hypothesis for empirical rc searcn to attempt a validation of its major claims 'The present study is an attempt to test the adequacy of ICG by applying it to the case of Sanshtit I 2 2 The following limitations have been imposed on this study (1) The first relates to case specification Sentences with metaphorical and suggestive interpretations excluded from this study Secondly, in addition to the above types of sentences, those containing negation have been left out (111) Thirdly, the intonational pattern of the sentences has been igpored The study of intonational pattern of Sanskrit sentences requires separate treatment Moreover, the intonation of the speakers of Sanskat mas largely be influenced by the intonation-pattern of the modern Indian languages which are, in most cases their first languages Besides, the study of accent in Sanskrit has been confined to the Vedic tests and is purely academic, and thus it has no relc vance for the classical Sansheit (11) Lastly, onls simple sentences, as a gencral rule, arc analysed in this study A departure from this is made in Allrrors ) (1971) p 247 Tillmore C} (106€a) p €8, (1991) P 246 85 86 ·INTRODUCTION 23 the case of genitive constructions and NP-asti-NP constructions wherein the process of nominalisation is involved. However, this departure is undertaken, more with a view to illustrating the points under consideration in simple sentences, than to explicating the characteristics of complex sentences. I 2. 2. 1 The data that are discussed in this study, are not based on some particular tests, but on the Loowledge of the classical texts of grammar and literature which I have acquired during the past years as a student of Sanskrit It is an attempt to analyse the principles or rules which underlie Sanskrit syntactic structures and which are internalised by a speaker-hearer of the language. Thus, this study is a "formal account of competence and makes 'no direct commitment vis-a vis performance '97 It is hoped that this study, in the sense of 'formal account of competence' would, given any particular sentence, spell out precisely the processes involved in its generation. Though, it cannot be claimed that this study exhausts the analysis of all possible simple syntactic structures in Sanskrit, it may reasonably be maintained that it does not exclude any structure which may materially affect the theoretical assumptions Lastly, this study is concerned with the Sanskrit syntactic structures, and not with the phonologyss of Sanskrit Besides, Sanskrit as spoken today in India, is influenced by the phonetic/phonological pattern of the speakers of different modern Indian languages So, there does not appear any need for giving phonetic details A systematic transcription may suffice for this study hing, RD, (1969), p 12 Zwich, A ,(1965) for 1 treatmert of the phonology within the fram" vork of generative grammar 67 88 ·II CASE-CATEGORIES II In thus chapter the case categories Agentire), D (ative), Receptive), Oſbjective), F(actitive), Instrumental), Locative), SO(urce), GO(al), Ext (ent), Pa(th) are discussed Of these, R1 is not mentioned by Fillmore It is proposed to discuss the case categories A, D, R in section II 1, 0, F and Cognate Objects in II 2, 1 in II 3, L, SO, GO, Est, Pa in II 4 II 1 In this section the case categories A, D, R are discussed and distinguished from each other Recently, the case-category D has been replaced with Experiencer, 3 but this substitution has not been followed by any discussion So, the case-category D, rather than Experiencer, is retained for the reasons given in II 1 3 below II 1 1 If the arguments of a verb are case categories A and DS, these must be realised by NP's having the feature [+animate] as in sentences 1, 2 If the NP acting as the argument of a verb (esen the same verb) is not animate, the case-category cannot be A/D, as in sentence 3 Vilson, W A A ,(1971), p 75 it appears that the case-category Ris mentioned, for the first time, in this thesis Fillmore, C J, (1968a), pp 85 86 Tillmore, C J, (1969) p 116 Fillmore, C. J ,(1968b), p 373 ‘ihe objects concerning which a predicato asserts something are the arguments of that predi. cate' The number of arguments is the number of souns a serb requires in 3 syntactically complete cspresion ci illmore, C ), (1969) pp 114 115 Case-catcgories A, D, R are cramired with refererce to their dcfinitions given by Til'rrore in II 12 ard II 13 1 2 3 4 on ·CASE-CATEGORIES 25 1. sevakah patat16 servant falls 2. mryate seyahah D servant dies 3. pustaham patati book falls The case-categories A, D are distinguished from R as follows. Whereas the case-categories and D may occur as the sole arguments of a verb, R (11.1.2) presupposes the presence of A/D in addition (to itself) Thus, sentences 4 and 5 are acceptable, but not 4' and 5'. 4. guruh dandena chättam tādayatı R teacher with stick student beats (the teacher beats the student with a stick) 5 mātā putram cintayari D R mothcc (the mother worries about her son) dandah chátram tādayatı I R Stich student bcats son -- Worrics dharmah putram R cintayant rcligion son wortics 6 The question of 170cr110-31 or ron interior al involcmert of oncor morcofthe participants in the cents described by the icrbs, ustalen upin II 14 and II 3 53 ·26 SANSERIT SYNTAX Vhen the case category A or D occurs in the caseframe of a verb in a simple sentence, it is subject, normally®, to a subjectivalisation-rule, and enters into concordial agreement with the V The subjectivalisation-rule which selects A/D for agreement with the V, applies optionally If case- Category R/O/F/GO also occurs in a case-frime, in addition to A/D, and A/D is not selected for subjectivalisation, R/O/F/GO9 is selected for subjectivalisation This fact is registered (1) in the from of adding the stem element 'ya? to the item under the V, (11) in taking on the features (of [+number] and [t-person]) of R/O/F/GO by the M, (111) and by a difference in the inflection selected by M[(i e an affia of the 'te'-class, rather than that of 'ti' class) This subjectivalisation of R/O/F/GO, in casc A/D is not subjectivalised, will be referred to, henceforth, as the process of passivisation If the case category R/O/F/GO does not occur in a case-frame, and the subjectivalisation-rule does not select AD for subjectivalisation, no other case-category is selected for agreement with the V, as a general rule, 11 and this fact is registered in the operation of the process of passivisation with the difference that the 11, invariably, tales on the features [+third] and [+-singular] This latter process 7 8 Sentences with the case category PR(ompter) in caucasus cors tructions are considered complex, and are not trcated in this analysis Sentences with OlI/L selected by the subjectnu alıcation rulc for agreement with the I, do not allow A D in their case frames (II 2 1 2 II 433) The subjectivalisation of GO is highly restricted (II 4 5) In traditional Sanskrit grammar, the 'tc class and the di class affires are referred 10 as 'itmancpada ard 'parimaipada' afixes respectuely Exceptions are mentioned in the scctions dcal ng with the Casc categories concerned 2 10 11 ·CASE-CATEGORIES 27 nom acc will be referred to as passivisation without subjectivalisation. Thus, a sentence like 6 corresponds to sentence 7, if A is not selected for concordial agreement with the V, and wherein the process of passivisation operates, and sentence 1 corresponds to sentence 8, in case the subjectivalisation-cule docs not select A, and, consequently, the process of passi visation without subjectivalisation operates 6 chátrah grantham pathati student book reads (the student reads a book) 7 chatrena granthah pathyate O by student book is read (the book is acid by the student) 8 sei akena patyate by servant is fallen ((it) is fallen by the servant) Sentences 6, 7 have the underlying structure (in relation to which P S rule,12 have already operated) as in figure I rig ire I S 1 nom 1 M [--present) path NP NP chatra frantha 12 Tor Srles ses 1 3 1 ·28 SANSKRIT SINTAX The subjectivalisation-rule13 applies and selects A for agreement with the V. The result is as represented in Figure II. Figure II S 1 - NI NP K +present tsing +thied 1 chātra +sing + third +mas. path NP gcantha 4-sing Emas The affix-substitution rule applics and replaces fcatures with affics, with the result as in Figure III Irgure III S NI } NP - chirra path NP 1 K 1 2171 frantha 13 ror T risasce V 132 ·CASE-CATEGORIES 29 The final form (aside from scrambling), 14 is as given in Figure IV Figure IV S chātrah pathati grintham Alternatively, at the stage of Figure I, the subjectivalisation rule does not select A, and, consequently, O is selected and the process of passivisation operates. The result is as in Figure V (from Figure 1). Figure V S MI Į present t.sing .+third ] NP } grantha t-sing +third +-mas path-ya- NP K rising. ] 14 chiatra -+-mas. Ross J R, (1970), IP 251 252 a scrambling rulc 'opuonally permutes major clements of a clausc, subject to various condı tions. In this study this problem is ignored, ard her ceforth, no reference to it will be made Also Ross, J R, (1967.)- pr 71 80, Lahof R T (1968), pp 95, 103 (f n 9)

·30
SANSKRIT SYNTAS
The affix-substitution rule applies and replaces
features with affises and the result ly as in Figure VI
Figme VI
M
-
te
NP
grantha
path ya
NP
chátra
ina
The final form is as in Figure VII
Figure VII
S
P
granthah
1
pathyatc
chatrena
Sentences 1 and 8 have the underlying structure
25 in Figure VIII

·CASI-CATEGORIES 31 Figure VIII S MI [+present) pat NP sevala The subjectivalisation-rule selects A for agreement with the V (as in Figure II), and the affix-substitution rule replaces features with affixes (as in Figure III), and the final form is as represented in Figure IX omitting intervening details) Figure IX S A sevahah patati Alternatively, at the stage of Figure VIII, the subjectivalısıtion cule does not apply, and the process of passivisation without subjectivalisation operates The result is as in ligure (from Figure VIII) ·32 SUNSARIT SYVTAI Figure X S JE *+present +sing third pat-ya- NP K - sevaha [+sing +mas The affis-substitution rule applies and the result is as in Figure XI Figure XT S te V pal-sapit-71- NP K seraha in3 The final form is 35 rcprc,cated in Ticure VII ·CASE-CATEGORIES 33 Figure VII S А . patyate sevahena It may be noted that the process of passivisation without subjectivalisation differs from the process of passivisation in that the M always selects (in the former case) the unmarked features [+sing] , [+third], as is evident from sentences 9 -12. 9. mürkhāh hasanti nom fo ols laugh 9. mūrhhaih hasyate А by fools is laughed ((tt) is laughed by fools) 10 mūrkhah hasati A nom fool laughs 10 mūrhhena hasyate А by fool is laughed ((it) is laughed by the fool) 11 balah sayane svapanti L girls in bed sleep (the girls sleep in bed) 11 balabhih sayane supyate L stet by girls in bed is slept ((it) is slept in bed by the girls) 3 1 nom ·34 SANSKRIT SYNTAT in bed 12. bälá sayane svapiti A nom L gic) sleeps (a girl sleeps in bed) 12'. balaya sayanc supyate L by girl is slept ((it) is slept in bed by the girl) 1 in bed The casc-category A is distinguished from D, as it occurs in a case-frame which is assigned to non-stative verbs. Non-stative verbs are those that may occur in reply to the query 'kim-karoti' (what does (x-do) 16 Thus, the verbforms in sentence 13 are non-stative, and have an A in their casc-frames, and those in sentence 14 are stative and have a D in their case-frames sa-lim-karoti 13 A sa pathati (reads) hc-what-docs hhelati (plays) (what is he doing) gāyati (sings) hasati (laughs)) sa-him-karoti. 14*. jivati (is alive) vidyate (exists) (is) bhavatı (becomes) cintayatı (worries) äste (stays) It may be noted that sentences (14) are deviant only as replies to the relevant question II I 2 In this section, R is distinguished from A and D All these three case categories are assigned to arguments of a verb which are realised by NP's with the feature [-t-animate] S2 asti 15 Thus, the case-categories A, D are distinguished from each other on syntactic aiterion as well ·CASE-CATEGORIES 35 1t It was mentioned above (II 1 1) that R presupposes the occurrence of A/D (cf sentences 4, 5) in its case-frame This case-category is distinct from A/D, because, though the entity specified as R undergoes or accepts the effect of an activity or state identified by the verb, it does not mutiate Fillmore has not postulated a case-category like R(fo 1). It may be that whatever is specified here as R, is subsumed under his Dto, which would mahe possible the elimination of a case-category lihe R Let it be accepted that R is not needed, and there is only D in addition to A Now, consider sentences 2 and 4 above, wherein both 'sevahah' (in 2) and 'chātram' (in 4) ate re-specified as D (rather D1 and Dº for convenience) The subjectivalisation-cule applies to the underlying structures of sentences 2, 4 and selects the D 'sevaha' (in 2) and the A'guru' (in 4) for subjectivalisation But, as mentioned earlier, subjectivalisation-rule operates optionally, and in case it does not apply, the Dl 'sevaha' (in 2) docs not enter into agreement with the V Now, the problem is, how to account for this difference in syntactic behaviour of Di and D* (in 2, 4); One may argue that D co-occurring with A (as in 4) must be subjectivalised, in case A is not subjectivalised But, this does not solve the problem. Consider sentence 5 Even Fillmore's definitions (IT 1 3) would specify‘mātă' (105) as D Now, what casc-category is to be assigned to 'putram' > It will be correct to assume that Fillmore would or will have to specify putram' as D (unless he invents another case-category) Apart from the fact that this occurrence of two Ds (in 5), in a simple sentence, goes against his principle of 'oneinstance-per-clause',17 it creates another problem Hot 16 17 for discussion sce II 1 3 Fillmore, C. ), (1971), p 248 ·36 SANSKRIT SYNTAX to account for the fact that the subjectivalisation-rule selects *mātā' (D?) for agreement with the V, and in case Di is not subjectivalised', 'putram' (D) neces cessarily enters into agreement with the V, with the consequential operation of the process of passivisation. To get round this problem, one may suggest that Fillmore would respecify D' ('putram') as O. But, this does not solve the problem. In II. 2.1.1 it is argued that there is need to distinguish R from O. Besides, this suggestion, if followed in a sentence like 15, would make the specification of case-categocies problematic. What case-category is to be assigned to 'putram' ini 15, Oor D? Following the assignation of O to 'putram in sentence 5, one would be inclined to specify 'putram'in 15 as O. The result is, again, the occurrence of two O's in one simple sentence, and the consequential problem of selecting one of the O's, in case the process of passivisation operates. 15. pitā putram dharmam brüte A 0 father religion speaks (the father speaks to his son about religion) It is difficult to anticipate the solution that may be suggested by Fillmore. But, there appears to be sufficient justification for the case-category R as postulated in this section. II. 1. 3. The three case-categories A, D, R have been distinguished above. Now, in the light of the above discussion (II. 1. 1., II. 1.2), the definitions18 given by Fillmore with reference to A, D, E(xperiencer) are examined below. The case-category A has been defined more or less consistently, and, in this analysis, the definition given Fillmore, C. J., (1968a), p. 24; (1969), P 116; (1971), pp. 251, 253. son 18.

·CASE:CATEGORIES 37 in 'the Case for Case' (1968a)19 is accepted. As specified above, sentences 1, 4, 6-13, 15 contain the case-categorv A. In so far as the case-category D20 is concerned, there have been proposals to replace it with Experiencer.21 It is proposed to discuss Experiencer first (as defined by Fillmore f. n. 21), and, then, take up Dative (f. n. 20). II. 1.3.11. The case-category Experiencer is not defined cxplicitly cnough to indicate which of the two NP's (in casc there are two NP's in the case-frame) in celation to a psychological event or mental state verb, is to be specificd as Experiencer. Consider a sentence like 5 above. Thc NP's mātā' (mother) and 'putram' (son) are both related to the verb 'cintayatı'. However, it may be interpreted (in favour 19. Terb 20 21. rillmore, C ,(1968a), p 24 'Agentnic, the case of the lypecally animate perceived instigator of the action sdertified by the Cf. Fillmors, C. ), (1969), p. 116 : Agent, the instigator of the crcnt'. Cf. Fillmore, C. J., (1971), p. 253 "she principal cause interpretation of the Agert case scem", for many sentences, to be quite adequate'. The problem of NP's indicating natural pheromena is discussed in II. 3. 3. I rillmore, C. ) , (1568a), p. 24 . 'Datic, the casc of the animale ang affected by the state of action ideatifed bs the serb.' Fillmorc, C. J. (1969), p. 116: 'Isperierces, the entity which receives or accepts or experiences or wsdergoes the effects of an action (carlice called by me 'Darne)'. Cf. Fillmoss, C.), (1971), p 251 : 'I have reanalysed the old Darne bs spread ng 11 arourd among the other cases. There there is a geru ne psychological crent or mental state Terb, ac hascthe Ispericrcer, where there is s ron-psychological scib which reais a change of stare, such as occ of dsrg or glok ng, clase the Obicct there there is a trarsfer or roretrert of some thing 10. per02, the reconstr as doit ta1 cn is les as the Goal!

·38 SANSKRIT SYNTAY S. of the definition) that the NP which indicates initiation of the psychological event or mental state verb, is to be specified as Experiencer. Thus, ‘mātā', and not 'putram', is Experiencer So far, there appears to be no difference between the old Dative (f. n 20) or Dative as analysed above (II 1.1) and Expeclencer except the terminological one Now, consider sentences like 16, 17, wherein the case-categories, following Fillmore (f. n 21), have been specificd as O, E 16 sohanah mriyatc O dies S. 17. sohanah glāyatı | cintayatı E feels weary / worries Whatever process, psychological or non-psychological, may be involved in feeling weary or worrying or dying, the way the NP sohanah' is affected in 16, is not different from that in 17 In other words, if no external agency affects 'sohanah' in 17, so is the case in 16 Thus, there is no reason why the 'sohanah' in both 16 and 17 should not be specified either as E or O. In so far, as syntactic justification is concerned, there appears to be none The subjectivalisation-rule applies to both (16, 17) optionally, and, in case OJE is not selected for agreement with the V, the process of passivisation without subjectivalisation operates in the case of both 16, 17 is is evident from 16, 17. 16 sohanena mryate 0 by s is died ( (it) is died by s) 17. sohanena glāyate / cintyate E ty s is felt weary / is worried ((it) is felt weary / worried by s) 1 - 1 ·CASE-CATEGORIES 39 nom acc son sees In sentence Again, take a sentence like 18, wherein Fillmore's definition of E, O{f n 21 (1971)) would lead to the postulation of two Os despite the explicit constraint against such a possibility 18 mātā putram pasyati O mother (the mother looks at her son) It is not clear, how the subjectivalisation-rule is to be specified in relation to its operation in sentence 18 15, the problem relates to the process of passivisation, and in 18, it relates to the process of subjectivalisation as well namely, how one of the two Os is to be selected for subjects valisation, and in case the subjectivalisation-rule doc, not apply, how one of the O, will be selected for passivisation The problem created by bringing in O as a partial substitute of D (f n 21 (1971)) is evident from sentences lıke 19 19 sevahah -prabhum--- mārgam- prcchatı way asha (the servant asks the master about the way) It is not clear, whether Eor Oshould be assigned to the unspecified NP in 19 If it is specified as E, it goes against the definition of Tillmore (f n 21 (1971)), and if it is O, then, again, thc problem in regard to passitisation (i e due to the occurrence of two Os) crops up In order to strengthen his position that Olihc L 15, at least parually, a replacement of D, and, thus, thcrcis some sort of link between 0 and E (f n 21 (1971), Fillmore suggests that cmbedding dominated by O may be related to a psychological cscnt (f n 21 (1971)) as in sentence 20 seti ant master ·40 SANSKRIT SYNTAX son - But, the embedding under O may relate to non-psychological events as well, as in sentences 21, 22. 20. pitā -- cintayati (yat -- putrah – kidịśaḥ - asti] OS S father - worries that - how is (the father worries as to how his son is) 21. sainikah - amriyata (yataḥ - deśaḥ no - parābhūyeta] OS soldier - died so that-country-not-may be subjugated (the soldier died so that his country might not be subjugated) 22. sainikah satrun hanti (yatah OS soldier enemies kills they - S te as ákramişyantis S will attaek (the soldier kills the enemy as they will attack) Considering sentences' like 16-22, iť is proper to ignore E, O (as defined in f. n. 21 (1971)) as replacements for either the old D{f, n. 20) or old E (f. n. 21 (1969)). Fillmore mentions that his old Dative (f. n. 20) includes what may be treated as Goal (f. 2. 21). But, this does not appear to be convinciag in view of his suggestion for a case-category like Benefactive, 22 in addition to D. However, it is proposed here to identify Goal as a case- Category distinct from AD (II. 4.5). II. 1. 3. 2. Now, that Experiencer as defined by Fillmore ((1971) f. n. 21) need not be accepted, it would be relevant to examine Experiences (f. n. 21) and Dative 23 (f. n. 20) in re- 22. 23. Fillmore, C. J. (1968a), p. 32. There appears to be no difference between D (19682) and E (1969) as defined by Fillmore, ·CASE-CATEGORIES 41 lation to D and R as specified in II. 1.1. and II. 1.2. above. There are three possibilities :D (1968a) and for E(1969) as defined by Fillmore (f.p. 20,21), includes (1) both D and R or (11) only D or (iii) only R, as explicated by me in II. 1. 1. and II. 1. 2. Whicheyer possibility is accepted, it indicates that D and R, as analysed above in II. 1. 1. and Tİ. 1. 2., are established on different grounds. As specified in II. 1. 1. and II. 1. 2., D initiates a state identified by the yerb (as opposed to A which is the initiator of an action identified by the verb), and R is the case affected by the state or action identified by the verb. As it has already been shown, D and R need to be distinguished from each other, and from A also, so there appears nothing to add on this point. II. 1.4. Before this section (II. 1) is concluded, one point needs to be considered. It has been suggested that there are sentences which indicate intentional or non-intentional involvement of an actant in the action identified by the verb. Now, the question is : should the NP specified as A, if there is non-intentional involvement, be rather considered a D? Consider sentences 1, 24-26. 24. bālakah pātrāni bhanakti 0 boy pots breaks (the boy breaks the pots ) 25. bālakahaNgulim --- churikayā -- chinatti A I boy finger with knifc- cuts (the boy cuts his finger with a knife) 26. rāmaḥ visam bhuNkte hc poison (he takes poison) eats ·42 SANSKRIT SYNTAX In all these sentences (1, 24-26), there may be an alternative interpretation which shows no intentional involvement of A Now, one may argue that as the actant is nonintentionally involved in the action, the case-category should be specified as D rather than as A. It is possible to disambiguate such sentences by adding adverbials like'sāvadhānam' (carefully) and 'nipunam' (cleverly / skilfully) In case either of these adverbials occurs in sentences 1, 24-26, the interpretation would be always intentional, and the actant would be specified as A. But, despite the above argument, thece appears to be no need to make this distinction in case-specification. The use of adverbials simply brings out the intentionality or non-intentionality of the actant, and it does not make a non-stative verb a stative one, even conceptually consider sentences 25, 26 specially. Whether the action performed is intentional or non-intentional, there is no difference in the way it is performed In fact, the vagueness in sentences 1, 24-26 may be attributed to our knowledge of the world 1 e we know that one would not normally perform an action like falling or cutting one's finger or swallowing poison. Again, there are verbs like hasatı' (laughs),ʻlikhatı' (writes), 'yāti' (goes) which may involve intentional or nonintentional participation of an actant, but no one would like to suggest that the case-category may be either A or D 10 the context of these verbs Besides, stative verbs like 'paśyatı' (sees) and Krśyatı' (becomes lean) may involve intentionality or non-intentionality of an actant, still, the case-category in their case-frames is specified as D Thus, it appears reasonable not to take account of intentionality or non-intentionality of an actant for casespecification.

·CASE-CATEGORIES 43 process of II 2 This section deals with the objective) and F(acti tive);R(esuit) cases and cognate objects As these have been considered related in one way or other, it is appropriate that they be discussed together, even though they appear, at first sight, to be different kinds of construct In sub section II 2 1 the case-category O is discussed. In sub section II 2 2, the F/R as a case-category distinct from O, and the relevance of cognate objects in relation to F/R are examined II 21 The case-category O is discussed in sub sections II 2 1 1, II 2 1 2, II 2 1 3 1,11 2 /3 2 In II 2 1 1 definitions of O are taken up In JI 2 1 2 senten ces containing O are shown through their derivations In II 2 1 3 I and II 2 1 3 2 the passivi sation which involves Receptive) as well, is discussed In II 2 1 4 sentences which allow obligatory deletion of A/D, are referred to II 2 I 1 According to Fillmore the case category O is assigned to the entity that moves or changes or whose posi tion or existence is in consideration 24 Neither thus nor an alternative difinition25 is followed up with exemplification or amplification The only change (from a to b) worth no ting, appears to be the elimination of the feature [animate] ? This is implied in Fillmore's remack about Experiencer and Object case categories wherein the two are considered to appear in complementary distribution, the Object being 24 Fillmore C), (1969) p 116 25 Fillmore, C ) .(1971) 251 "The Object case is that of the entity which moves or which undergoes changes, and I still use it as a waste basket sentences embedded to Objects can serve to identify for example the content of a psychological event as with the verbs of judging or imagining 26 Fillmore CJ (1971), p 251 'I no longer confuse selection restrictions to animates with true case like potion 26 ·O S. tām 44 SANSKRIT SYNTAX assigned in the context of a non-psychological velb indicating a change of state.27€ But, with this definition of O, sentences like 1, 2 would pose problem for Fillmore with' regard to case-assignment, as he would have to specify'two occurrences of O,'thus, breaking his 'one-instance [of one case-categóry bok] '-per-clausk principle.:29 1. sohanah tām paśyati 0 ,her looks (s. looks at her) 2. rāmah adriyate O, her respects (r. respects her) Secondly, there appears to be little syntactic justification (Fillmore has given none in 'Some Problems for Case Grammar'), ' for bringing o in line with A{gentive) and E(xperiencer) by eliminating the feature of (animacy] : As developed by Fillmore, the case-categories 'A, E are assigned to entities which are[+animate29] . Now, his definition of O (f. n. 27) indicates that it may be assigned to an entity which is [+animate]. Moreover, though' o is'assigned in relation to non-psychológical verbs (f. n. 27), the sentenceembedding under' O may be in relation to'verbs of judging: and imagining (f.' n. 25).' Thus, the demátcation between E and O becomez hazy to the extent that the, definition or explication ofthe O, as in'Some Problems for Case-Grammar', I. 1 27. Fillmore, C. J., (1971), p. 251 : 'Where there is a non-psychological verb s hich indicates a change of state, such as one of dying or groning, we have the Object." Fillmore, C. J., (1971), p. 248. Fillmorc, C. ) , (1971), pp. 249-250; 253-255, 261. t. 28. 29.

·CASE-CATEGORIES 45 ASE-C appears to be of no relevance to distinguish it as a case- Category. So, the elaboration of O, as in 'Some Problems for Case-Grammar', is ignored30 and, in the following discussions, it is proposed to refer to the case-category 031 as (1) the semantically most neutral case, the case of any. thing representable by a noun whose role in the action or state identified by the verb, is identified by the semantic interpretation of the verb itself; (ii) conceivably the concept should be limited to things, which are affected by the action or state identified by the verb." .:"32 This definition contains two parts (macked by me as) (i) and (ii). The first part accounts foc sentences like 3, 4, 5 and the second for sentences lıke 6. 3. gāyati sings vins balikā gånam А girl song (the gırl sings a song) atithih gamanam 33 A guest going (the guest leaves), 4. karoti 1 does 30. 31. 32. For some problems arising out of this postulation of O, see sub-sections 22., 2.3. The postulation of Oʻas a waste bashet' (Fillmofe, J., (1971), p 251) raises a problem for his case-theory, which is referred to in Chap VI. Pillmore, C ) , (1968a), p 25 : 'the term (Objective bsk) is not to be confused with the notion of direct object, nor with the name of the surface case, synonymous with 'accusative." Constructions like 'gamanam-karoti' are discussed in II. 2 2. 2. Towards the end of II 2 2 2, it is proposed that such cotstructions may be covered by the process of nominalisation. 33. i ·46 SANSARIT SYNTAS fIVCI - sun parıdhānam (11) - 1. 5 (1) nadi vahatı (11) váyuh vāti (til) suryah ~ divyatı 0 flows wind - blows - shines 6 (1) sohanah rañjayati A S dress dyes (s dyes the dress) Tāmah patram pathati А lettes reads (reads a letter) The case-category O is assigned to the entities which are [-animate] and, thus, it is distinguished from the case-category R(eceptive),34 assigned to the entities which are [-+-animate], and are affected by the action or state identified by the verb The case-category Rappears distinguished from O on three points in the first instance, R, as opposed to O (as in sentence 5 above),can never occur alone Its occurrence 10 a case-frame presupposes A/D Consequently, Omay be selected for subjectivalisation, even in the absence of the process of passivisation, but R can be subjectivalised only when the process of passivisation has operated Seconlly, in a sentence like 7, there is need to distinguish R and O ('one instance-per-clause principle' may be recalled) 7. yātrikah-nāgarikam -märgam-prcchati A R tourist-city dweller --way -asks (the tourist asks the city-dweller the way) Thirdly, if Rus distinguished from O which is assigned to inanimate entities, the embedding under o for processes like complementation appears semantically more appropriate This case-category is mentioned first (to my Lcon ledge) by Wilson (1971) non acc O acc 34 ·acc 1 1 CASE-CATEGORICS 47 11. 2. 1. 2. In this sub-section, the case-category Ois specified in relation to different syntactic structures. The caseframes in the sentences below, arc (A, I, 0-) in 8, 9; [14] in 10,10 ;(A, I, R-] in 11,12; (I–) in 13, 13'; [A,O.. in 14, 15, 17, 18;[0–) in 16, 16', 19, 19'. The derivations of8, 9, 10, 14-16 are given in detail, those of 11, 12, 13, arc limited to theic underlying structures. The derivations of 17-19 being similar to that of 14-16, are not indicated. 8. Satruh againā grāmam -- dahati A nom enemy with fire — village burns (the enemy burns the village with fire) 9. satrunā agnina grāmah dahyatc A I O nom by enemy-- with fire- village 1s burnt (the village is burnt with fire by the enemy) 10. agnih dahati (svayameva)35 Inom firc burns (itself) 10. agninā dahyate (svayameva) 1 is burnt (itself) ((it) is burnt by fire (by itself) ) 11. ākhetakah-dhanusa -- paśūn vidhyati 1 hunter with bow-animals-shoots (the hunter shoots animals with a bow) 12 akhetakena-dhanusā - pasavah, vidhyante А 1 1 R nom by hunter — with bow-animals--are shot (animals are shot with a bow by the hunter) 1 by fire nom 2 Racc 1 35 This sentence with a case frame [1-] is given here to compare with sentences 16,16',19,19' which have a case frame like [0–1- ·48 5 SANSKRIT SYNTAX 13. - itself) ( 13. 14. A nom F 20C 15. dhanuḥ vidhyati (svayameva) I nom bow shoots dhanuşă vidhyate (svayameva) I by bow is shot (itself) ((it) is shot by the bow (by itself) ) sudah odanam pacati cook bolied rice cooks (the cook cooks bolied rice) sūdena odanah pacyate А i by cook boiled rice -- is cooked (boiled rice is cooked by the cook) odanah Pacyate (svayameva) F nom boiled rice cooks (itself) (svayamera) F by boiled rice is cooked - (itself) ((it) is cooked by boiled rice (by itself) ) takşakah kāstham bhinatti A carpenter wood splits (the carpenter splits the wood) F nom 16. 16'. odanena-- pacyate 17. - nom ACC 18. is split taksakena kästham bhidyate nom by carpenter wood (wood is split by thc carpenter) kästham bhidyatc (svayamera) 0 nom wood splits (itself 19. - ·49 2 19 CASE CATEGORIES kasthena bhidyate (svayameva) O by wood is split (itself) both=(it) is split by the wood (by itself)) 1 Sentences 8,9 have an underlying structure as represented in Figure I Figure 1 S 1 M - [+pres ] V A dah NP K NP K NP 1 I satru agni grāma The subjectivalisation rule36 applies and selects A. The result is as gren in Figure II 36 It may be noted that the subjectivalisation rule in Sanskrit is optional But, whenes or the case-frame contains Rard/or Or, in addition to A/D, ard A/D is not selected for subjcctiralısa. tion, R or O or I, as a gencral rule, is selected for subjectvalisation This process is referred to as passivisation aith subjectivalisation (II 1 1) ·50 SANSKRIT SYNTAX Figure II 1 +pres NP K +sing .t-third s'atru +sing +third dah NP K NP K +-mas grāmar +-sing: 1+mas. agnir+sing.) +-mas The affix-substitution rule replaces the features with affixes, and the result is as represented in Figure III. Figure III S P IMla NP IKI Satru у dah am NP K NP 1 agni ina grāma The final form ( aside from scrambling ) is as given in Figure IV.

·CASE CATEGORIES 51 Figure IV S Р satruh I dahati agnina grāmam Alternatively, if the subjectivalisation rule does not select A(see f n 36) O is selected for subjectivalisation, and the representation is as in Figure V (from Figure I) Figure V S Lo M | - 1 | r+pres NP K +sing +-third grama +-sing dah +-third NP K NP K mas $atruIsning ]agni[imas ] The affix-substitution rule replaces the features with affires It may be noted that if Oſor R/F) is selected for subjectivalisation (and A/D is also present in the case frame), M selects an affix which belongs to the 'te'-class, 37 rather than the 'ti? 37 See f n 10 ·52 SANSKRIT SYNTAX class, the stem element 'ya’ıs added to the V 39 The result is as in Figure VI. Figure VI S - M P - te NP grāma dah-ya- NP NP K satru ina ina agni The final form is as represented in Figure VII Figure VII S grāmah Šatruni dahyate agninā Sentences 10, 10 harc the underlying structure as in Tigurc VIII 38 The whole process may b: referred to as that of passitisation If the form under the V contains the 'ga' element already, no igal element is added 35 in divgati dit gate' ·CASE-CATEGORIES 53 Figure VIII S NI VI --- dah NP к PP svayameva agni It may be noted that I is the only case-category in the caseframe of sentences 10,10 (like the FjO in 16, 16', 19, 19 and the I in 13, 13) If there is another case-category in the case-frame, I cannot be selected for subjectivalisation 39 The P(ost)-P(osition) 'svayameva'(by itself) bars the occurrence of any other case category in the case-frame of a simplex I is selected for subjectivalisation The P(ost)P(osition) "svayameval is deleted optionally The result is as in Figure IX 39 A sentence like agnih gramam dahatı' (fire village burns fire burns the village) appears doubtful Even if it is argued that in this sentence, the NT agnih indicating a natural phenomenon (cf II 3 3 1) behaves idiosyncratically an I, sentences like 10, 13, 16, 19 do not allow the occurrence of any other case-category in their case frames, if they have to convey the sense of 'svayameva' (by itself) (cf f n 41) as ·54 SANSKRIT SYNTAX Figure IX S M NP K +pres +third +sing 1 dah - agni t-sing +third tmas The affix-substitution zule replaces the features with affixes, and the result is as in Figure X Figure X S M NP } dah agni The final form is as represented in Figure XI Figure XI S agnih dahati ·55 CASE-CATEGORIES It may be recalled that the subjectivalisation rule being optional, it may be that no case-category is selected for subjectivalisation In such a case, (1) the Mindicates [t-third], [+singular], (11) the affis under M shows 'te' rather than *ti', (111) and the stem-element 'ya' is added to V 40 I is not selected for subjectivalisation and the result is as represented in Figure XII (from Figure VIII) Figure XII S M +pres +sing +-third] dah PP NP 1 1 agni svayameva +-sing L+mas The affix-substitution rule replaces the features with affixes. The Post Position is deleted optionally The final form is as in Figure XIV via Figure XIII 40 Recall the process applied to figure VI (f n 38) This process is similar to the earlier with the differerce that no case-category is subjectivalised and consequently the M as it does not take on the features of (number) ard (person) of any case-category, always indicates the features [+third] [+person) [+singular) This process may be referred to as passivisation without subjectivalisation ·56 SANSKRIT SYNTAX Figure XIII S NI - te dah-ya- NP K agni ina Figure XIV S Þ as in dahgate agninā! Sentences 11, 12 have the underlying structurc Figure XV. Figure XV S NI [+pres ] $t R sidh NP NP K NP alhetaka dhanus pašu llorc therc would be sodificsence of R is replaced by o But as mentioned earlier (II 2. 1. 1) there is need to distinguisha cuc-category lilc R from O.

·CASE-CATEGORICS 57 The subjectivalisation-rule applies and selects A, the process as applied in the case of Figures II, III, IV, follows and gives the sentence 11 Alternatively, A is not selected, and R being in the case-frame, is selected for subjectivalisa- The result is as in Figure XVI It may be noted that M indicates the feature [+plural] taken from R Figure XVI tion S R M -- NP K +pres +plural .+third paśu 1 +plural +third +-mas vidh NP NP K -- ākhetaka [Emnas Jdhanus Emas ] -fem The process of passivisation as in the case of Figures VI, VII, applies and gives the sentence 12 Sentence 13 has a derivation similar to that of sentence 10 (figures VIII, IX, X) If I is not selected for subjectivalisation (as in figure XII), the process of passivisation without subjectivalisation (f n 40) applies and, through configurations similar to those in figures XII, XIII, XIV, we get sentence 13 above ·58 SANSKRIT SINTAX Sentences 14, 17 and 15, 18 have derivations similar to those of 8,11 and 9, 12 respectively The derivation of sentence 16 is shown. Sentence 19 has a derivation similar to that of 16, and hence is not shown. Sentence. 16 has the underlying structure as in Figure XVII. Figure XVII S M [+pres ] 1 NP PP odana siayameva It may be noted that the subjectivalisation-rule is optional. If A/D also occurs in the case frame (as in Figure I), in addition to R/O/F/GO, and the subjectivalisation-rule does not select A/D, R/O/F/GO must be selected for agreement with the verb In the above configuration, only F occurs (recall sentence 10 has only I in its case frame)42, and the sub jectivalisation-rule applies, and selects it for agreement with the V. The Plost) Position) 'svayameya' is deleted optionally The result (omitting intervening details) is as it Figure XVIII 42 fr 39 may be recalled "agnih'(fire) as a natural phenomenon may be subjectivalised, iodicating an idiosyncratic characteri stic in a construction like 'agnih gramam dahatı 'But, in no case would a sentence like 16 allow the occurrence of another case category Besides, if the sense of 'by itself is to be conveyed, even I in sentence 10 would not allow the occurrence of another case category ·CASE-CATEGORIES 59 Figure XVIII S A - *+pres NP K +sing -+-third] pac odana +sing +-mas +third The process of passivisation applies and the affix-substitution cule operates The result is as in Figure XIX. Figure XIX S - te V NP K рас-yaodana The final form is as in figure XX Figure XX S odanah pacyate ·60 SANSKRIT SYNTAS Again, the subjectivalisation-rule need not apply at the stage of figure XVII and, consequently, the process of passivisation without subjectivalisation (as in figure XII) applies, and the final form (aside from scrambling) (leaving out the intertening details) is in figure XXI Figure XXI S pacyate odanena Sentences 19, 19, will have a derivation similar to that of sentences 16,16 above (through figures XVII–XXI) It may be mentioned that sentences 10, 10, 13, 13, 16, 16 , 19, 19, all indicate that the action process is accomplished easily, as indicated by the occurrence of the Post Position 'svayameva' (by itself) in their underlying structures Such constructions allow only one case category (1 e 1/0) in their case frames In this sense, sentences 10* (recall f n 39,42), 134, 16" are unacceptable (sentences 13", 16" are, anyway, unacceptable) 10 gramam dahati I fire village burns (fire burns the village) dhanuh pas un vidhyatı I R bow animals shoots (the bow shoots the animals) agnih Stam 13' * ·rice in pot process of CASE-CATEGORIES 61 16". odanah sthālyam pacyate F L cooks (rice cooks in the pot) In the case of [O-], but not in the case of (1–), the passivisation also operates (cf. sentences 10, 13, 16, 19) Sanskrit grammarians would try to distinguish the verb-forms 'pacyate', 'bhidyate' in 16, 19 from those in 16', 19, on the ground that there being difference in their pitch accent, 43 they need to be derived differently. As mentioned in the beginning, no note of pitch-differentiation is taken in this analysis, and, therefore, there appears to be no necessity to propose different derivational processes for verb-forms 'pacyate' and 'bhidyate' in sentences 16, 19 and dahyate' and 'vidhyate' in sentences 10', 13. This is further supported by the verb-forms 'pacyate' and 'bhidyate'in 16', 19 wherein it would be difficult to separate the two processes II 2 1 3 1 In this sub-section, a set of sentences is considered, in which, there is subjectivalisation of a case-category other than R/O/F (which are present in the caseframe), in case the subjectivalisation-rule does 'not select A/D Consider sentences 20", 21", 22", 23" wherein a casecategory co-occurring with R/O/F enters into agreement with the V as A/D is not selected for subjectivalisation 20 gopah goh dugdham dogdhi SO ab F milkman from cowmilk mulks gopah gām dugdham dogdhi A SO F milkman COW milk milks both (the milkman mulls (milk from) the cow) Singh, J D, (1970), pp 33 34 A nom acc 20. nom асс асс 43 ·62 SANSKRIT SYNTAX 20. gopena ŠO nom 2CC 1 Anom $usyaya for pupil 22 gauh dugdham duhyato A 1 F асс by milkman COW - milk is milled (mill is milked from the cow by the milkman) 21 sūdah tandulebhyah ** odanam pacati A nom SO ab F асс cook from fice boiled rice cooks 21'. sudah tandulan odanam pacati A nom SO acc F cook rice boiled rice cooks both=(the cook cooks boiled rice out of rice) 21'. sūdena tandulah odanam - pacyantc A SO nom F 2CC by cook rice (s) boiled rice - are cooked (boiled rice is cooked out of rice by the cool) 22 guruh dharmam brūte R d 0 acc teacher religion talks guruh śisyam dharmam- brüte A nom O асс teacher pupil religion talks both=(the teachec talks on religion to the pupil) 22. gurunā śisyah dharmam ucyate A R nom by teacher pupil religion (the pupil is talked to by the teacher on religion) 44 Later Sanskrit commentators consider it to be harana' (INSTRUMENTAL upper face indicates that it is not to be identified with I/1), which appears to be wrong CF Roy, S, (1920) p 12 he says that it could be either 'Larana' or 'apa dana' (SOURCE) Cf Bhattojıdıksıta (p 523) tattvabo dhını commentary says it is 'karana’and (by implication) adhikarana' (LOCATIVE) On the other hand Patanjali mentions 'edhan' (fuel) as "karana', and not 'tandula' (rice), p 325 He gives a Genitive form of 'tandala, as in 'tandulanam-odanam pacatı' (of rices boiled rice (he) cooks (he cooks boiled rice)), P 332 Racc 1 асс is talked ·CASE-CATEGORIES 63 acc - 23. acc 23. bālakah — sārameyam vraje runaddhi Anom R LI boy dog in enclosure holds bälakah sārameyam vrajam runaddhi nom R L. acc boy dog in enclosure holds both=(the boy holds the dog in an enclosure) 23*. bālalena särameyam - vrajah - rudhyate A R acc L nom by boy dog in enclosure is held (the dog is held in an enclosure by the boy) -- 1 me It may be noted that in sentence 22, R and O co-occur and it is R, and not O, that enters into agreement with the V in sentence 22". This selection of a case-category other than R/O/F for agreement with the V, if A/D occurs in the case-frame and is not selected for subjectivalisation, is limited to a set of twelve verbs. Of these, four are given above (20-23). There has been some disagreement among the traditional grammatians regarding the exact number of such verbs. 45 IJ. 2. 1.3. 2. In this sub-section, a further set of sentences which allow a case -category other than R/O/F to be in 45. Roy, S, (1920), p 11. he quotes Bhattojidikşita's list of 12 verbs . (1) dogdhi(milks), (11) yācate (begs), (111) prcchati (asks), (1v) tunaddhi (holds), (v) brūte (speaks), (vi) śăsti (instructs), (v1) pacati (cooks), (v111) cinoti (gathers), (1x) jayati (wins), (x) mathnāti (churns), (x1) muşnăti (steals), (x11) dandayatı (punishes). Cf. Subrahmanya Sastri, P. S., (1970) (vol. V) he refers to Patañjali's list which contains the first sıx of Bhattojidıkşıta's list and (911) bhikṣate (begs). Kaiyata adds (1x),(1), (x11) of Bhaſtojidiksita's list to that of Patañjalı. Cf.Bhojadeva (1100 A. D.), p 13 he refers to the list of Patañjali, and adds two to his lists : (1) ‘katoti' (does), (11) 'gehnätı' (tales). p. 110 ·64 SANSKRIT SYNTAX A nom 1 acc agreement with V, in case A/D is not selected for subjectivalisation, is considered. In sentences 24", 26' L and I are in agreement with the V respectively. This selection of L and I induces the operation of the process of passirisation and is conditioned by the occurrence of the pre-verb 'adhi' with the verbs'sete/tişthatiſāste' (sleeps/sits/stays) in sentence 24", and by the occurrence of the verb 'divyati' (gambles) in sentence 26" (cf. in sentences 24", 26", L and I are not in agreement with the V respectively, and passivisation without subjectivalisation takes place). 24. sundari śayane seteſtişthatiſáste L beautiful lady in bed sleeps/sits/stays 24'. sundari sayanam-adhi -- Seteſtisthatijāste A nom L beautiful lady- in bed sleeps/sits/stays bɔth=(a beautiful lady sits/sleeps/stays in bed) 24". sundaryā śayanam-adhi-sayyate/sthiyatelāsgate A L nom by beautiful lady-bed—is slept in/sat into stayed in (the bed is slept in sat into stayed in by a beautifullady) 24''. sundaryā sayane - sayyte/sthiyateſāsyate L by beautifullady- in bed - is slept in sat into stayed in (the bed is slept in/sat into stayed in by a beautiful lady) It may be noted that seatences 25, 25' are unacceptable. 25.* sundary śayane-adhi-sayyatejsthiyatejāsyate A 1 by beautiful lady-in bed - is slept insat into; stayed in 25',* sundarya śayanam - śayyatejsthiyatejāsyate A by beautiful lady-bed - is slept indsat into/stayed in 1 i L 1 L nom ·CASE-CATEGORIES 65 ) 1 - divyati nom 1 26. dyūthatah nom 1 In sentence 26 I co-occurs with A. In sentence 26", A is not subjectivalised and I being optionally selected for agreement with the V, the process of passivisation operates. But, in sentence 26', I is not selected for agreement with the V, and passivisation without subjectivalisation takes place. 26. dyūtaharah ak saih А I (a gambler gambles with dice) 26. absan divyati A I acc gambler dice gambles ( a gambler gambles with dice) 26." dyūtakarena aksāh divyante А I nom by gambler dice are played (dice are played by a gambler) 26", dyūtakarena aksaih divyate А Ι by gambler with dice is played ((lt) is played with dice by a gambler) Before concluding this sub-section, it is worth noting that the process of passivisation may operate, but R/O need not be selected for agreement with V due to cooccurrence of some post-position. In sentences 28, 29, 30, the occurrence of prati' and 'antarena' blocks the selection of R and O for agreement with the V as in sentences 28', 29, 30. However, the process of passivisation without subjectivalisation applies. 28 guruh śisyam prati vadati Racc teacher pupil speaks (the teacher speaks to the pupil) 1 1 - ) А nom to 5 ·66 SANSKRIT SYNTAX 28. - to Owner 29'. is looked acc son gurunā śisyam prati ucyate A 1 Racc by teacher pupil is spoken (pupil is spoken to by the teacher) 29. svami geham prati paśyati D nom o acc house to looks (the owner looks at the house) svāminä geham prati drśyate D 1 асс by the owner house to (the house is looked at by the owner) 30. mātā putram-antarena cintayatı D nom R mother about worries (the mother worries about (her) son) 30'. mātrā putram prati cintyate D Racc by mother son about is worried ((it) is worried about son by the mother). It may be noted that sentences 28"",29",30" which undergo the process of passivisation, along with the selection of R and O for agreement with the V, are unacceptable. 28'* gutuna Šisyah prati ucyate A i R nom by teacher pupil to is spoken 29 * svāmina geham prati døyatc 0 nom by owner house 30" * mátra putrah untarena cintyatc AT R by mother about is worried 1 to 15 secn nom son ·CASE-CATEGORIES 67 1 1 son In case the post-positions pratı' and 'antarena' do not occur, the R/O agrees with the V, and the process of passivisation applies, and we obtain sentences 31, 32, 33 31 gurunā sisyah ucyate A R nom by teacher pupil is spoken (the pupil is spoken to by the teacher) 32 svaminā geham drsyate D O nom by the owner house is looked (the house is looked at by the owner) 33 matra putrah cintyate D R nom by mother 15 worried (it) is worried about son by the mother) It may be noted that sentences 31, 32, 33 with R/O not in agreement with the V and the process of passivisation operating, are unacceptable 31 * guruna sisyam ucyate A R 100 by teacher pupil is spoken 32 * svamina geham drsyatc D O acc by owner house is looked 33 * putram cintyate Racc by mother 1s worried II 2 1 4 In this sub section, a set of sentences, in which A/D is obligatorily deleted (in fact, A/D is a dummy), and O is obligatorily subjectivalised, along with the process of passivisation operating, is considered The case category O may dominate an embedded S as well, as in sentence 35 Consider sentences 34, 35 1 1 matra 1 son ·68 SANSKRIT SYNTAX 34. Tāstragánam giyate 0 nom national anthem is sung (national anthem is sung) 35. śrūyate-ſpātaliputre-asokaḥ avasat] OL A nom is heard in Patna lived (it is heard, a. lived in Patna) Sentence 34 has the underlying structure as in figure XXII and 35 has as in figure XXIII. Figure XXII S - N [+pres.] NP K rāsţra-gāna Figure XXIII S M [+pres.] śru pāfaliputre-síokah-prasat ·CASE-CATEGORIES 69 II. 2. 2. In this sub-section the case-category F(actitive) Result), as distinct from O is discussed, and its relation to cognate objects is examined. II. 2. 2. 1. The case-category F is proposed in'the Case for Case’.46 In "Types of Lexical Information',4F is replaced with Re. But, this change in nomenclature is not followed by any discussion at all, and thus, it is of no relevance. Besides, Re, as defined above, is covered by F. Recently, the elimination of F/Re as a distinct case-category has been proposed.4 According to this latter proposal, F/Re may be covered by GO(al). But, consider sentences like 36, 37. 36. sūdah bhaktam pacati A F cook boiled rice cooks (the cook cooks boiled rice) 37. kayih gitam racayati A F poet song composes (the poet composes a song) 46. 47 48. Fillmore, C. J., (1968a), p 25 the case of the object or being resulting from the action or state identified by the verb, or understood as a part of the meaning of the verb' This definition has two parts which are taken up later (II. 2. 2. 2). Fillmore, J ,(1969), p 116 'the entity that comes into ex- 1stence as a result of the action.' Fillmore, C J, (1971), pp 251-252 . 'Since the Goal case is used to indicate the later stage or end result of some action or change, it can absorbs hat I used to call 'Resultative' or 'Tactative', that is, it specifies the end result role of a thing which comes into existence as a result of the action identified by the predicator, as in 'I wrote a poem'or 'I constructed a bridge. As Fillmore's above proposal occurs in 'Some Problems for Case-Grammar', I would refer to the earlier position (as in fn 46, 47 above) as 'Pre-Problems' position.

·70 SANSKRIT SYNTAX The case category F as specified in sentences 36, 37 (according to 'Pre Problems' definitions) would be replaced by GO(al) following the ‘Post-Problems' position of Fillmore But, even accepting his definition of Goal, it does not appear to assign the GO to the NP, S specified as I above Secondly, if his contention is accepted, even Oin sentence 38 could be a GO 38 mātā patram pathati A nom o acc mother letter reads (mother reads the letter) However, he gives only six lines to his suggestion (f n 48) Considering all these, it is appropriate that the dışcussion of F/Re (and also of O and cognate objects) is confined to their treatment in the Case for Case' (f n 46) II 2 2 2 Now the question is is the case category F syntactically justified ? Is it related to cognate objects, and does the concept of cognate objects support its postulation > Consider sentences 39, 41 and 40, 42 in which the NP's 'patram' and 'asandı' are assigned the case categories F and O respectively 39 sitā patrm likhati A F letter writes (s writes a letter) 40 patram pathati O letter reads (s reads a letter) -- S sita S 49 See I 4 6 ·CASE-CATEGORIES 71 him 41. taksakah asandım nirmāti F carpenter chair mahes (the carpenter makes a chair) 42 tak salah asandim bhanakti 0 carpenter chair breaks (the carpenter breaks the chair) Sentences 40, 42, but not 39, 41, are possible in reply to sentence 43, because they (40,42) indicate the existence of the letter/chair prior to the initiation of the action 43 sitā/tahsah ah patrasyalāsandyāh haroti 5, Carpenter - of letter, of chair -- what does (what does she carpenter do to the letter, chair ?) Fillmore beings in the notion of cognate objectsso to justify the postulation of F Sentences with cognate objects are constructions in which at the very least, there is a high selectivity between a specific V and an object N, and in which the V+ N combination in one language might well be matched by a V alone in another '51 Consider sentences like 44, 45 44 sā asanam aśnāti she meal (she eats (her) meal) 45 să a Sanam karoti she eating does (she eats) According to Fillmore, sentences parallel to Fillmore, J ,(1968a), pp 85 86 under the section problems and suggestions 51 Tillmore, C J, (1968a), p 85 But even in the same language a VAN construction may match with a V construction (sentences 44, 45 above) eats 50 ·72 SANSKRIT SYNTAX 44, 45 have the underlying structure as in figure XXIV with F dominating a dummy (irrelevant details are eliminated) Figure XXIV Si À [+pres ] aśnā NP sā A copying rule, then, fills the dummy under F with the item under V. The result is as in figure XXV Figure XXV S M 1 [-+-pres ] í F asnā aśnā NP K 1 sa By another rule the V under F assumes the form of N (as in figure XXVI) ·73 CASL-CATEGORIES Figure XXVI S M V aśnā NP NP - sā aśana Fillmore proposes onc rules to combinc the two processes: 'copy the N-represcntative of the V under the F'. At the stage of figurc XXVI, sentence 44 is derived. Alternatively, by a rule, 53 the V assumes 2 pro-V form, and the result is as in figure XXVII. Figure XXVII S NI [+pres.] karo NP K NP к aśana sā 52. 53. Fillmore, C. J. (1968a), p. 85. Fillmore, C. J., (1968a), p 85, 'Replace the V bythe designated pro-V.' In Sanskrit, there is no verb-form comparable to 'have' in English which acts as a pro-verb among others. However, the verb-form 'karoti' (sentences 45, 46) functions as a proverb in a large number of constructions.

·74 SANSKRIT SYNTAX Sentence 45 is derived from a configucation as represented in figure XXVII. Fillmore's rule (11) appears to be an oversimplification. "Traditional grammars call objects like song, sleep, breath'cognates' because the noun in the socalled direct object position is morphologically related to the verb '54 This definition of cognate objects is covered by the first rule (of Fillmore), but not by the second rule Thus, according to this, sentence 45 would not be related to cognate objects Fillmore would like to extend the definition of cognate objects 65 Thus, he would consider a sentence lihe 45 as well as a sentence like 46 related to cognate objects it may be be noted that sentence le 46115 unacceptable. Sentence 46° is acceptable, but, then, sentences 46, 46' are not synonymous) 46 sā śayanam karoti A she sleeping does (she sleeps) 46 * sā sayanam sete she sleeping sleeps 46". sä sayane Šete A L she in bed sleeps (she sleeps in bed) Fillmore's extension of cognate objects to cover sentences like 45, 4615 not discussed here further, as I do not 54 55 Baron, NS, (1971), p 71 Fillmore, C J, (1968a), p 85 'Some words may be treated as cogaate-object V's even though the rule for replacing the pro V is obligatory' His example is 'nightmare nightmarc' => 'have nightmare' (p 86) F ·CASE-CATEGORIES 75 Weaver Weaves support the position that the concept of cognate objects is relevant for the postulation or the case-category F, as it is evident from the discussion below, Now, examine the relationship between cognate objects and F. Recall the definition of F (f. n. 46) and consider sentences 47-50 along with 44, 45. 1.47. patam vayati F cloth (the weaver weaves a cloth) 48. mūrtikārah mūitim racayati F idol-maker idol makes (the idol-maker makes an idol) 49. sa-sbhuNkte : cats vs bhojanam : çating-2-karoti he gacchati : goes does gamanam : going pathatı : reads pathanam reading 50. sundari madanalekham likhati F beautiful lady love letter (the beautiful lady writes a love-letter) ܐ . writes Sentences 44 and 45 are not parallels. Sentence 44 cannot have a continuation as in 44', but 45 can have as in 45', rotikāyah asanam āśnāti 44'.* să А she of bread food eats 45'. så roţikāyāh aśanam karoti А she of bread eating does It may be noted that 44" is acceptable.

·76 SANSKRIT SYNTAX 44". sā rotikām aśnāti eats she bread food (she eats bread, food) Thus, in 44, in accordance with the criterion given above (II. 2. 1. 1), the NP 'a śanam' will be assigned the casecategory O, and not F. Some may say that sentence 45 is an instance of a construction with deletable object. The notion of deletability of objects does not appear to be of significance for postulating case-categories, as, it is for the distinction between transitive and intransitive verbs 56, and in a case-grammar, the distinction between transitive and intransitive verbs need not be retained, as it is accounted for differently. In sentences 47, 48 the NP's marked as F are not cognate objects. In 50 the cognate object is also F. The part one of F's definition by Fillmore (i. e. 'the case of the object resulting from the action or state identified by the verb' (f. n. 4) accounts for sentences 50 (which has a cognate object), 47, 48 (which do not contain cognate objects), and leaves out sentence 44 (which contains cogaate object). Thus, the concept of cognate objects neither fully accounts for the case-category F nor makes the cognate objects a subset of F. Sentences 45, 46, 49 can be considered to contain cognate objects (this is an extension of the concept of cognate objects by Fillmore as mentioned above), and can be specified F only when the part II of the definition of F(i. e. 56. Lchrer, Adrienne, (1970), p. 227; 'some linguists have treated verbs which occur with or without objects as transitive verbs that allow their objects to be deleted.'

·CASE-CATEGORIES 77 "understood as a part of the meaning of the verb' (f.n. 46)) is taken into account. But, if this is so, how is F distinct from O, at least partially (cf. II. 2. 1. the first part of the definition of O) ? As explicated above, the concept of cognate objects does not appear to account forthe case-category F. And as this is so, there is no need to retain the part two of the definition of F as its sole purpose has been to account for instances like 45, 46, 49 above. The process of nominalisation (see IV. 3. 2.) may account for the phenomenon of cognate objects. II. 2. 2. 3. In this sub-section I revert to the problem of the status of F as a distinct case-category. There has been a suggestion 57 to distinguish the performance of a song which is already in existence from the creation of a new song by singing, as in sentences 51, 52. 51. sitā jayadevasya gītam gāyati А 8 song sings (s. sings the song ofji.e, composed by i) sitā gitam gāyati A FO song sings (s. sings a song) Sentences like 52 are ambiguous because the song may be already composed (as in sentence 51) or is being created While singing. Consider sentences 53-56 which indicate the indeterminacy of any semantic criterion to distinguish F from 0. 53. sudah dugdham pacati А cook milk boils (the cook boils milk) Lehrer, Adrienne, (1970), p. 249. S. of j. 52. S. 57.

·78 SANSKRIT SYNTAX 54. pacati S. cooks sohanah tandulān/māmsam А rice / meat (s. cooks rice meat) sita rotikām pacati bread cooks (s. cooks bread) sitä bhojanam pacati 55. S. 56. S. food cooks (s. cooks food) In sentences 53-56, the verb-form 'pacati' occurs in the context of the NP's whose status as FIO cannot be specified on the basis of semantic-criterion only. In 53, 'dugdham' is 0, and not F, as its resultative aspect is insignificant. In 54, which stage do the NPs specified as F indicate ?": rice and meat when they are put on the oven or when they are taken off the oven ? In sentence 55, the NP 'rotikām' (bread) is F to an Indian, but an O to an Englishman. . And, in 56, the resultative aspect of 'bhojanam' is definitely relevant. All these might suggest that resultative and nonresultative are points along a semantic continuum. 58 However, in this situation, the relevance of a syntactic criterion cannot be ignored. In the beginning of the sub-section II. 2.2.2. reference is made to one syntaccic criterion furthe distinction between Faad O, and whose application successfully specifies the case-categories in sentence 55, the case-category F is assigned to frotikām' (bread), because, in Sanskrit, sentence 55 is not possible as a reply to the query 'sitā-roţikāyāh - kim - karoti' (s. - of bread - what does). 58. Baron, N. S., (1971), p. 75.

·CASE-CATEGORIES 79 There has been a suggestion 58 to eliminate the case-category F. The reason given is the indeterminacy arising out of the application of a semantic criterion. But, this situation may obtain in the case of other case-categories as well, and it is precisely in such situations that syntactic criterion plays a crucial role. It has been further suggested 60 that the motivation for postulating a case-category like F lies in the adherence to the rule that there should not be more than one occurrence of one case-category in a simple sentence. Thus, in a sentence like 'he has a dream about Jayā' the case-frame, as suggested, is [D+F+0–) rather than (D+0+0–). But this suggestion does not solve the problem There are similar sentences like 'he thought about the planjabout the child', 'the teacher talked about the student, in which the case-category O/R could be assigned to the NP's underlined A case-category like 'referential has been suggested (but not developed) by some linguists 61 Whatever may be the ultimate selection, the postulation of F appears to be only a partial solution of this problem II. 3 The case-category Iſnstrumental) is assigned to an NP which is the inanimate force or object causally involved in the action or state identified by the verb "62 If the NP (acting as an argument of a verb) is not inanimate, the case-category assigned to it, cannot be an I. Thus, sen- 59 60 61 62 Baron, NS (1971), p 75 Cf, Filmore, C J (1971) he proposes to drop Fand assign Goto the NP's concerned (see f n 48) Baron, NS, (1971), p 77 Stockwell, R P, et al , (1968), vol II, p 957 Fillmore, ), (1968a), p 24 Cf. id , (1969), p 116 'the stimulus or immediate physical cause of an event Cf id (1971), p 251 'the case of immediate cause of an event or the stimulus, the thing reacted to', ·80 SANSKRIT SYNTAX 1 1 With pen 2 * tence 1 has an I, but not sentences 2 and 3 Sentence 2 is unacceptable 63 chátrah halamena lihhati A nom I Student Writes (the student writes with a pen) chatrah balakena likhati A nom 1 student with boy Writes 3. chătrah bālalena lehhayatı 61 PR A student by boy mahes write (the student makes the boy write) 63 Sentence 2 15 acceptable, if interpreted as an instance of Comitative(for details,Kumar, Coordination in Sanskrit (forthcoming), and as an alternative form of 'chatrah balakena-saba-likhati' 'chatraḥ-balakah-ca-lıhhatah' (the student and the boy write) 64 PR (ompter) refers to the case-category which initiates an action state identified by the verb, in addition to an A(gentive) These constructions containing PR and A, may be referred to as causa tive constructions Such constructions being complex, there is no inconsistencyıf PR also is referred to as A/D (as the case may be) However, PR is used to distinguish it from AD of the embedded S, and, secondly, the use of PR indicates the neutralisation of A D in the matrix S(1 e the matrix S would show PR in place of A/D) Causative constructions do not necessarily javolve mar phological distinction of verbal forms Sentence A is a ron causative construction, sentence B is a causative one A balalah pustakani ganayati A O boy books (the boy counts books) pita balakena pustahanı ganayatı PR nom o father by boy books makes count (the father makes the boy count the books) non acc counts B А ACC ·CASC-CATEGORIES 81 In a sentence like 4, the NP “hastinā' specified as I should be interpreted as 'with the body of an elephant' (hastinahśarirena).65 This is supported by the unacceptability of a sentence like 566. duştah - hastinā - dvāram runaddhi A I 0 wicked with elephant - door obstructs (the wicked obstructs the door with an elephant) - sa 5. guruh bāla kena patram likhati A 1 O teacher with boy letter writes The above characteristic of I becomes explicit, if sentences 6 and 7 are compared. Sentence 6 is derived from an underlying structure as represented in figure 1. But, sentence 7 has an underlying structure which is not complex (as in figure II). 6. sevakena bhittim pātayati PR A 0 he by servant wall makes fall (he makes the servant knock the wall over) 7. khanitrena bhittim pātayati A he with spade wall makes fall (he knocks the wall over with a spade) 65. Fillmore, J ,(1968a), p 24 (f. n. 32) : he refers to, ard discusses an example suggested by P. Postal : 'I rapped him on the head with a snahe.' 66. Refer to f. n. 63 above, sentence 5 like sentence 4, is acceptable, if it is interpreted as an instance of Com. 6 sa ·82 SANSKRIT SYNTAX Figure I S M [+pees.] PR [+causative] NP K - 52 ME [t-pres.] pātay NE NP K NP 1 sevaka bhitti Figure II S M I [.+pres.] pātas NP K K NP NP 1. khanitra Sa bhitti ·CASE-CATEGORNS 83 II 3 1. The casc-category I 15 distinguished from A/D/R/O/I.It is distinct from A/D as a subjectivalisationrulc normally sclects A/D for subjectialisation, and not I. If A/D is not selected, and R and I are present in the cascframe, the process of passivisation applies to R, but not to I If I is the only case-category, in addition to AD which has not been subjectivalısed, the process of passii isation without subjectivalisation operates 67 Morcover, A/D/R arc assigned to NP's which are [+animate], but I is not The case-category I is also distinct from F/0, though both are assigned to NP's which are ſ-animate). If a casc-frame contains the case-categorics A, I, O, and A is not selected for agreement with the V, then, it is O rather than I which is selected by the process of passivisation: Thus, a sentence like & would correspond to 9 wherein O, not I, enters into agreement with the V If I is selected for agreement (as in sentencc 10), the sentence is unacceptable 8 mātā churihayā phalam chinatti A nom o acc mother with knife ftust (the mother cuts fruit with a knife) 9 churlaya phalam chidyate I by mother with knife fruit (fruit is cut with a knife by the mother) mātrā churika - phalam - chinattschudyate I nom O accinom by mother - knife fruit cutsjis cut Sentences 8, I have the underlying structure as in figure III cuts matra 1 1 O nom - Is cut 10 * - 1 67 for details sce II II ·84 SANSKRIT SYNTAX Figure III 'S 1 M P [+pres:] v chid K ne NP NP K NP K mātar phala churikā The subjectivalisation-rule applies and selects A for agreement with the V, the result is as in figure IV. Figure IV S M -+pres. +sing. L+third- NP mätar +sing. +-fem. +third chid Ne NP K NP K phalar+sing 7 churikāstsing. -mas. t-fem. -fem. ] ·85 CASE-CATEGORIES The affix-substitution rule substitutes the features with affixes, as in figure V, and the final form is as in figure V. Figure V S M NP -S chid mātar NP K I. 1 NP 1 churikā phala am ya Figure VI S mātā chinatti phalam churikayā Alternatively, if A is not selected for subjectivalisation at the stage of figure III, the process of passivisation 68 applies. The result is as represented in figure VII (from figure I). 68. For details see II. 1. 1. and II. 2. 1. 2.

·G
SIXSXRIT SIST
Fire I'll
S
VE
.
picka
Ing.
--third
NP
Ne
1
chid
phala ---sing:
":tful
-fen.
1
XP
.XP
måtar - remi ]churika[:en]
17.c alix-sulnutution sule applics, and il.c soul is as in
4:170 VIII.
J'sure I'III
1.

·CASŁ-CATEGORIES 87 Figure IX LO phalam - chidyatc mātrā churkaya It may be noted that a sentence like 11, with no A in its caseframe, is suspect churiha phalam chinatti I O knife fruit cuts II 3 2 1 In this sub-section certain syntactic peculiarities of I are considered 11 * nom cuts Consider a sentence like 12 in which I is the only case-category present, 12 churikā (svayameva) chinatti I knife by itself (the knife cuts by itself) In constructions like 12, the I conveys the performance of the action with ease or without effort 69 subjectivalisation rule which has selected I for agreement with the V in 12, is optional If the rule does not apply, sentence 12 is the alternative form The 69 Also see II 2 1 2 (senter ces 10 13) ·88 SANSKRIT SYNTAX i by knife 12'. churikayā (svayameva) chidyate (by itself) is cut (it) is cut by the knife by itself) Sentences 12, 12' have the underlying-structure as in figure X. Figure X S NI [+pres.] v 1 chid NP K PP churika svayameva The subjectivalisatin-sule applies, and the affix-substitution rule replaces the features with affixes, and the result is as represented in figure XI (omitting intervening details). Figure 21 S ΔΙ 1 1 PP chid NP K churtha svayameva The final form is as represented in figure XII.

·89 CAST-CATEGORIES Figure XII S NP PP chinatti churikā svayameva A PP-deletion rule deletes 'svayameva' optionally, and we obtain sentence 12. Alternatively, if the subjectivalisation - tule does not apply, the process of passivisation without subjectivalisation operates, and the result is as in figure XIII (from figure X). Figure XIII S *+present +sing: -+third 1 chid NP PP 1 churikā [+sing. svayameva -+fem. The affix-substitution rule replaces the features with affixes, and the result is as in figure XIV.

·90 SANSARIT SYNTAS Figure XIV S - NI - te chid-ya NP K PP 1 y a stayameva churtha The final form is as in figure XV Iigure XV chidyate churikaya - siayameva The PP-deletion cule deletes 'svayameva' optionally and we obtain sentence 12 II 3 2 2 Consider, next, sets of sentences in which the process of passivisation, in relation to I, operates in various ways As mentioned earlier (III), if there is R/O/F in a case frame, and the subjectivalisation cule does not process passivisation with the subjectis alisation of R/O/F takes place But, if there is an I in the case frame (with or without O), it is, gencrally, not $ubject operate, the ·CAST-CATEGORIES 91 to the process of passivisation with subiectiv alisation (as is cvident from sentences 13', 13"). However, in sentence 14", thc I is optionally selected for subjectivalisation with the process of passivisation; sentence 14' shows the ususal syntactic behaviour of I, 1. C. its non-subjectivalisation with the process of passivisation. In sentence 15', still another peculiarity of I is indicated, wherein the I co-occurs with an O. The process of passat isation operates, but it is I, not O (cf 11.2 1.2) which is subjectivalised optionally (sentences 15', 15) This characteristic of I (sets B, C below), as mentioned above, is due to its occurrence in the case-frame of a verb like "divyatı'. However, if a pre-verb like 'prati' is attached to.divyatı', the process of passivisation does not select the I for subjectivalisation, rather it selects O optionally for subjectivalisation (sentences 16', 16") In set A, the process of passivisation does not select I foc subjectivalisation (sentence 13' is unacceptable), even if a subjectivalisation-cule does not select A for agreement with the V (sentences 13, 13) In sets B,C the verb "divyatı' has the caseframes (A,1],[A,1, 0-] respectively In both these caseframes the process of passivisation selects I (not O) optionally for agreement with the V (sentences 14", 15'), in case the subjectivalisation-cule has not selected A In set D, the verb 'pratıdıvaytı' has the caseframe (A, I, 0-1 The process of passivisation selects O and not I) for agreement with the V (sentence 16 ) in case subjectivalisation-rule does not select A 70 Thus, we need ad hoc features on some verbs (as well as a specification of their case frames), if we are going to predict correctly the possibilities of passivisation 70 ·92 SANSKRIT SYNTAX А 13 13 bālalah gendukena khelatt A nom boy with ball plays (the boy plays with a ball) bālalena gendukena khelyate А by boy with ball is played ((it) is played with ball by the boy) bālalena gendukah Chelyate A nom by boy ball is played 1 1 13" * 1 B 14 dyūtaharah 21 dit yati А nom ahsan 2CC ahsaih 1 gambler dice/with dice gambles (the gambler gambles with dice) 14 dyütaharena ahsaih divyate А I 1 by gambler with dice is gambled 14 dyūtaharena ahşah divyante A by gambler dicc arc gambled both=((it) is gambled with dice by the gambler) 1 I nom - C 1 15 dyūtaharah - 1 ahşān satasya -dityati nom асс O aksaih I gambler - dicc/withdice - of hundred - gamble I morphophonemic rule accounts for inflexional variation (acc/1) in the context of the icb'di yati' A morphophonemic rule accounts for the genitric irflexional rcalisation of o 72 ·CASE-CATEGORIES 93 15 1 (the gambler gambles hundred (pieces of money) with dice dyūtakarena aksah - satasya — divyante А I nom O 8 by gambler - dice - of hundred - are gambled (hundred (pieces of money) is gambled with dice by the gambler) dyūtakarena - aksaih – śatasya divyate А by gambler-with dice-of hundred -18 gambled (hundred (pieces of money) is gambled with dice by the gambler) 15'. 1 16 dyūtaharah - aksaih А I satam”3) -pratıdīvyatı nom 1 acc satasya 16. 1 gambler - with dice - hundredſof hundred -gambles (the gambler gambles hundred pieces of money) with dice) dyutalarena - ahsath - śatam - pratıdīvyatc А I 1 O nom by gambler - with dice - hundred - is gambled (hundred (pieces of money) is gambled with dice by the gambler) djūtakarena - aksaih - śatasya - praudivyate А I 1 8 by gambler-with dice-of hundred-is gambled (hundred (picces of money) is gambled with dice by the gambler) 16" 73 A morphophonemic rule accounts for the optional genitive inflcxion in the context of pre serb + divyatı' ·94 SANSKRIT SYNTAX a S2 - II 3 3 1 Consider, next, set of sentences containing NP's which refer to natural forces and are specified as I 17 vayuna dipam — samayatı A I he with wind — lamp extinguishes (he extinguishes the lamp with wind) 18 vaijñānikah varsābhih sasyam - nāsayatı А I 0 scientist With cains ссор destroys (the scientist destroys the crop with rains) 19 vayuh dipam śamayatı I O wind lamp extinguishes (the wind extinguishes the lamp) 20 varsāh śasyam nāśayanti 1 rains crop destcoy (the rains destroy the crop) 21* dandah chatram samayati I R stick student quictens 22* astram sainiham nāśayatı 1 R weapon soldier Mills 23 dandena chatram śamayati I R teacher with stick Student quictens (the teacher quictens the student with a stich) 24 sainikah astrena satrum nāsayati nom I R soldier with a capon cncmy - destroy's (thc soldier destroys the enemy with a scapon) - - - guruh 1 асс ·CASE-CATEGORIES 95 It is apparent that the NP's Svãyunā' and 'varsābhih' (specified as I in 17, 18) fuaction like A in 19, 20(cf. 'guruh' and 'sainilah' specified as A in 23, 24) It may be noted that 'dandena' and 'astrena' (specified as I in 23, 24) Cannot function like A, as sentences 21, 22 ace unacceptable. It 15, perhaps, proper to distinguish between 'vāyunā' and 'varşābh'h'ın 17, 18 from "väyuh' and 'varsâh' in 19,20. In 17, 18 the I's occur with A The NP 'vāyunā' (with wind) may be esther the servant's breath ('mukhasya vāyunā': with the wind of the mouth) or the wind of a fan ( vyajanasya-vāyunā) Similarly, a scientist may create artificial rain to destroy produce crops, as in sentence 18 Now, the problem : how to specify the NP's "vāyuh' and 'varsäh'ın 19, 20 which are phenomena not subject to anybody's control (e g the 'vāyuh' in 19 could not be interpreted as the breath of someone) It has been suggested that there could be a case like 'Force?.?4 As Fillmore mentions, the proposed case Force can never occur in contrast with A/I 75 One may suggest that the NP's referring to natural forces may be specified as either A or I If the natural forces are specified as A, then, it will have to be a type of A that cannot occur with I. But, if they are I, then, this characteristic need not be specifically mentioned. Secondly, if they are I, theic 'event-causing'?4 characteristic becomes automatically explicit However, One may argue that they do not co-occur with A(as in 19,20 : provided the natural forces specified as I in 17, 18 are considered distinct from those in 19, 20) But, even in such sentences (19, 20) one can conceive of some Agent after all However, the problem of the occurrences of NP's referring 74 75. Huddleston, R, (1970), p 504 Fillmore, C ), (1971), PP 253 254 ·96 SANSKRIT SYNTAX - water to natural forces as A (in sentences like 19, 20) remains to be solved. It has been mentioned earlier (II. 3. 2. 1.) that there are constructions in which I functions like A i.e. the subjectivalisation-rule selects it for agreement with the V. Similarly, in the case of NP's indicating natural forces and specified as I, the subjectivalisation-rule may also apply. II. 3. 3. 2. Now, consider a sentence like 25 in which one could argue that the case-frame is either [I], 1,–] or [A76, I--). 25. agaih tāpena jalam śoşayati I, I, fire with heat - dries (fire dries water with its heat) However, the NP agnih' (in 25), need not be specified as I in 25. Sentence 25 is derived from the structure underlying sentence 26, as the relation between 'tāpena' (heat) and agnih'(fire) is an inalienable one.77 [agneh tāpah] jalam sosagati ΙΟ of fire heat dries (the heat of fire dries water) Thus, there would be no need to specify agnih' (in 25) as I, as case-categories are to be specified in their underlying structures. Fillmore suggests that the NP's indicating natural phenomena, in constructions specified as I, may be derived from superordinate structures. There appears to be no justification for accepting this suggestion which involves treating such NP's at a level different from those NP's which do not refer to natural phenomena, and In case one argues that the NP's indicating natural forces are.. See IV, 3. 2. 1. 1.5. Tillmore, C. J., (1971), p. 254. 26. - water 76. 77. 78.

·CASE-CATEGORIES 97 1 are assigned the case-category I. Fillmore himself has not amplified or csemplified his suggestion. Moreover, the postulation of the case-category I in the case of NP's refering to natural forces, as in sentences 17-26, does not violate the principle of 'one-instance-[of one case-catcgory] bsh.perclause'. II. 3.4 Consider, next, sentences which contain NP's expressing emotive states. 27. sevahah bhayena/bhayāt hampate A 1 ab servant -with fear/from fear trembles (the servant trembles with fear) 28. bāla sohena śokāt rodit: А I 1 ab girl with grieffrom grief weeps (the girl weeps out of grief) 29 matā hrodhena/hrodhāt putram - tādayati A I ab R mother - with anger/from anger -son beats (the mother beats her son out of anger) 30 pitā - snehena / snehát putram - āhvayati А I ab R father - with affection/from affection -son-calls; (the father calls his son with affection) 31. mātā-krodhena/hrodhāt-dandena-putram tādayatı А 1 I ab I mother-with anger/fromanger-withstick-son-beats (the mother beats herson with a stich, out of anger) In sentence 31, there appear to be two occurrences of I Sentences like 31, appear to be a strong argument against postulating the principle of one occurrence of one case category in a simple sentence. Three possibilities suggest themsels es : (1) the first relates to the 1 1 1 7 ·98 SANSKRIT SYNTAX 27'. - sevaham servant postulation of a new case-category for the NP's which ein press emotive states, (11) the second is the acceptance that the NP's expressing emotive states and specified as I may Co-occur with another I, (111) the third is deriving these NP's from superordinate structures Of these, (1) the first would lead to the multiplication of case-categories, thus eventually making their postulation trivial, (11) the second makes the 'oneinstance-per clause-principle' untenable The third possibility suggests itself in view of sentences like 27', 28'. bhayam Lampayati PR А fear makes tremble (fear makes the servant tremble) 28 solah bālām rodayatı PR А grief girl (grief makes the girl weep) However, sentences 29-31' appear to be suspects. 229. hrodhah matra putram tādayati PR А R anger by mother makes beat (anger makes the mother beat her son) 730. snehah pitrā putram ah ayati PR R affection by fathermahes call (affection makes the father call his son) 731'. krodhah - mātra dandena-putram--tādayatı PR I R anger by-mother - with stich - son - mahes beat (anger makes the mother beat her son with a stich) Sentences like 27, 28' support the deration of thc NP's indicating emotic states from superordinate makes weep son -son - ·CASE-CATEGORIES 99 32. 1 structures. But, sentences like 29-31' appear to go against this position. II. 3. 5. In this subsection, NP's with 1-infleuon which may be derived from underlying complex structures are considered. The problem of intentional/non-intentional interpretation of sentences containing I, is also discussed. JI. 3. 5. 1. Consider sentences 32-33 which appear superficially to be simple, and may be derived from the structures underlying 32-32", 33-33" respectively. pathihah aśvena yāti A nom traveller with horse goes (the traveller goes on / with a horse) 32'. pathikah aśvam āruhya79 - yāti A nom traveller horse having ridden goes (riding a horse, the traveller goes) 32", pathikah aśvah yātah80 А А traveller horse and go (the traveller and the horse go) 33 sevahah dipihayā yāti A nom with lamp goes (the servant goes with a lamp) 33 sevahah dipikām ādāya yāti A nom O acc servant lamp having taken (having taken a lamp, the servant goes) 79 The process of nominalisation is discussed in IIT 5 and IV 32. 80 For the details see Kumar, Coordination in Sanskrit (forthcoming) Racc V nom 1 servant - V nom goes ·100 SANSKRIT SYNTAX 33". sevahah [dipikāyāh - prahāśe] - yāti Anom LO g servant in light goes (the servant goes in the light of a lamp) of lamp The NP's 'asvena' in 32 and 'dipihayā' in 33 cannot be specified as I Unlike an I, the NP 'aśvena' is [-tanimate], and can function as an A (as in sentence 34 below) The NP 'dipihayā' can be interpreted either as derived from a complex structuce underlying sentence 33 or as a complex structure as in sentence 33". The verb 'yāti' does not allow the occurrence of dipihayā' as an I (as defined in II 3) Compare sentences 35, 35' which can be interpreted either as 33' or 33" Sentence 35 can allow the occurrence of an I, in addition to the occurrence of 'dipihayá' as 35'. One may argue that in 35 (dipihaya-hfielatı : with lamp-play>) 'dipihayā' may be an I, but, in that case, it would be interpreted as a toy. in sentence . 34. aśvah yāti 35. 1 horse bālalah Anom boy balalah - A nom 35 gocs dipihayā . qhhclatı • play Jilhati : Writes frith lamp Ilhadatı : cats halamena dipthaya li hati I 1 with pen - with lamp halamona - dipihasă khelat1 I with lamp - plays camasena - dipikaya İhidani I 1 with spoon - with lamp - cats Writes 1 with pen 1

·CASE-CATEGORIES
101
In sentences 32', 33' the items specified as Nom V81 and the
item 'prakāśe' in 33' are deleted optionally by later rules,
and the NP's dipikām' and 'dipihāyāh' (in sentences 33',
33" respectively) and aśvam' (in sentence 32') realise
as 'dipihayā' and 'aśvena' respectively (1. e. with
1-inflevon). Sentences 32,33 have the underlying structure
as represented in figures XVI-XVII and XVIII-XIX
respectively.
Figure XVI
Se
[+pres.]
yā
NP
K
pathiha
M
[+perf]
R
aruh
āruh
NP
K NP
|
pathiha
aśva
81.
The process of Nominalisation allons deletion of 1 Nom by
1 later rulc

·102
SANSKRIT SYNTAX
Figure XVII
S
NI
[+pres.]
V
ya
saha
ca
NP
NP
NP
1
pathika
aśya
Figure XVIII
S.
M
[+pres.]
ya
S,
sevaka
N
[+perf.]
ada
NP
K NP
dipika
scvaka

·CASE-CATEGORIES 103 Figure XIX M [-+pres.] A ya NP K sevaka 1 I prokāśa NP sa - 1 dipikā Thus, in a sentence like 36 the NP's (asvena', dipikayā' (as m sentences 32, 33) cannot be specified as I. A sentence like 3682 is derived from structures underlying sentence 36' (as represented in figures XX, XXI). 36. aśvena – pathā – dipikayā -- gacchati A he by horse - by road - by lamp - goes (he goes with a lamp on a horse along the road). Patañjalı, Vyakarana Mahabhäsja, ed. M. Pandit Sivadatta D. Kudála (1912), Bombay, Nirnaya Sāgar Press, pp. 259-60 : According to the commentator there are three 'Laranakāraka'-(INSTRUMENTAL)-relations in a sentence like 36. It is not clear how 'patha' can indicate 'karana-hāraha'. Secondly, commentators have not taken into account ambiguous interpretation of 'aśvena' and 'dipikaya (sentence 36'). Thir. dly, if 'harana' is defined as 'the most serviceable in the accomplishment of an act' (Roy, S,p 36 . Pānini's rule I. 4. 42), It is unclear how three actants can be 'most serviceable simultaneouslyın relation to one action Cf. Siogh, J D., (1970),p.29. 82.

·0104 SANSKRIT SYNTAX 36'. sa- - aśvena saha А aśyam āruhya he- with a horse I riding a horse gacchati goes Figure XX S, pathā dipikām ādāya Pa dipikāyāh prakāśe - along taking a lamp road in the light of a lamp M [-+pres.] Pa gacch NP K NP K tat pathin M 1 [+perf.] V - āruh NP K NPK 1 tat asva 1 [+perf.] NP āda NP 1 tat dipika This interpretation is not sepresented in figures IX, XXII. A luce morphophoremic rulc scaliscs it as 'sa'. 83. 84.

·CASE-CATEGORIES 105 Figure XXI S MI [+pres.] SI K gacch NP 1 tat M [+pres.] R āruh NP K NPK NP K -- tat as va N prakāśa NP K dipikā II. 3. 5. 2. Consider, next, sentences 37, 38, which may be considered to be derived from the same underlying structure.85 37. sā churihayā phalam khandayati A I O she With Knife fruit (she cuts fruit with a hnife) - cuts 85. Lakoff, G, (1968), pp. 4-29 ·106 SANSKRIT SYNTAX - uses 38 sã-phalam-khandayıtum-churikām-prayuNkte A VNom 0 she --fruit to cut hnife (she uses a knife to cut fruit) Lakoff, while advocating an identical underlying structure for sentences, like 37-38, refers to its justification in the form of the postulation of a significant grammatical generalisation on the basis of conditions of selection and cooccurrence, which now require to be stated only once in the description Chomsky86 has considered different examples to show that Lahoff's position is not tenable In the first instance, Lakoff considers that a sentence with an instrumental adverbial phrase and a sentence containing a verb form lihe prayuNlte' (uses) are synonymous But, this claim appears doubtful in Sanskrit as in English Consider sentences 39,40 39. ākhetabah - pašum -hantum-bānam- prayuNkte A R VNom hunter animal--to kill --arrow-uses (the hunter uses an arrow to Lill the animal) 40 ahhetakah paśum bānena -hanti А R I hunter animal with arrow-kills (the hunter Lills the animal with an arrow) Now, at least with one interpretation, the above sentences are not syaonymous In case the tense in sentences 39, 40 is interpreted as simultaneous with the point of reference (I c speaking), sentence 40 refers to the operation of hilling, but sentence 39 refers to the operation prior to hilling Besides, sentences 39, 40 are not synonymous on another count also Sentence 39 can have a continuation like 39, but sentence 40 cannot be continued like 40. - 86 Chomsky N ,(1969), PP 12 14, p 43 (f n 16) Cf hoou, ) G, (1971), pp 72 79 ·CASE-CATEGORIES 107 sa - Sa 39. ahhetahah-pašum hantum-bānam-prayuNkte hunter animal to kill-arrow-uses parantu palāyate but that (animal) escapes (the hunter uses the arrow to kill the animal, but that (animal) escapes) 40'. * akhetahah pašum bānena -hantihunter animal with arrow-bills parantu palāyate but that escapes (the hunter kills the animal with an arrow, but that (animal) escapes) Thus, Lakoff's proposition that sentences 39, 40 have the same underlying structure and are synonymous, does not appear to be convincing Besides, sentences 41 and 42 show that if Lahoff's suggestion is accepted, then, the I in 41, 42 would come from another structure (superordinate or embedded) but the L would not to be no justification, except the paraphrase with 'use', for postulating this differentiation between the I and the L. 41. (1) mātā kalamena patre lkhati A I L mother with pen on paper Writes (the mother writes on paper with a pen) (11) matā - darvy - pâtre - dugdham - nidadhāti А I L 0 mother with ladle in pot milk puts (mother puts milk in a pot with a ladle) 42 (1) matā - kalamam - patre - lehhitum - prayuNkte L VNom mother - pen - on paper - to write - uses (the mother uses the pen for writing on paper) There appears ·108 SANSKRIT SYNTAX -dogdham- (11) matā dartim pātre L mother ladle nidadhatum prayuNhte V Nom uses in pot - mulk - to put (mother uses the ladle for putting milk in the pot) If it is proposed that the L in 41, 42 also comes then we w.ll have sentence 43(1) and 43(11) in place of sentence 42(1) and (11), and sentence 41 and 43 would not have similar interpretation Consequently, Lahoff's position would be weakened further 43 (1) mata - lelhitum - O 0 -paper -and - to write - prayuNkte - са halamum - patram mother - pen uses -- to-put (mother uses a pen and paper for writing) (11) mātā - dacvim - patram cz - dugdham - 0 0 mother ladle pot and - millnidadhātum prayuNite uses (she uses ladle and pot for putting milk) Thus, it appears difficult to accept Lahoff's position in relation to the case category I IT 3 5 3 Now, it has been argued that the instrument of an argumºnt may be used with an intentional or a nonintentional interpretation An argument assigned the casecitegory Instrumental may have ‘purposite' or 'accidental' scntcnc: lihc 44 is ambiguous Tnc purposesense is indicatc.) if sentence 44 us continued like 44 Thc accidental sense is expressed if 44 is continued as 44' 67 Lloff, G, (1968) scnsc 87 ·CASE-CATEGORIES 109 44'. s'88 44 sevahah vetrena casal am abhanak I servant with stick cup broke (the servant broke the cup with a stick) sevahah -anāyāsena - vetrena - casakam - abhanah servant - with effortlessness - with stich-cup-broke (the servant broke the cup easily with a stick) sevahah-anavadhānena-vetrena-casaham-abhanah servant - with carelessness - With stich - pot broke (the servant broke the glass with a stich carelessly) Lahoff gives a number of 'grammatical contexts where the 'accidental interpretation of a sentence is ruled out He, then, proposes that if there are grammatical contests in which one sense but not the other can appear., then it must be assumed that, two different deep-structures are involved '89 However, he does not propose an underlying structure for the accidental interpretation But, the question remains whether the difference between the accidental and purposive interpretation of a sentence libe 44 is indeed a reflection of two underlying structures Consider sentences 45, 46 In sentence 45, the hand is cut by a Life whether it happened accidentally or was done purposely Thus, the instrumentality of knife is not affected by this ambiguity Besides, the ambiguity remains even in sentence 45 wherein no overt I occurs 45 bālalah churihaya hastam alunat 0 boy with knife hand (the boy cut hts hand with a Laufe) Lakoff G , (1968), pp 8 9 he gives eight such contests Ibid 90 Lakoff, G, (1968), p & 'I am interested in this paper only in the purposive instrumental sense rot in the accidental sense" 90 I - cut 88 89 ·110 SANSKRIT SYNTAX 46. bālakah hastam alunāt 0 boy hand cut (the boy cut his hand) Thus, the intentional/non-intentional or purposive/accidental interpretation of I seems irrelevant to the definition of I. II. 4. In this section the case-categories Locative), Source), GO(al), Ext(ent) and Palth with spatio-temporal orientation 91 are discussed. The discussion starts with a reference to the Sanskrit grammarians, and is followed by that of the views of contemporary linguists. II. 4. 1. The Sanskrit grammarians in their discussion of ‘kāraka'-relations, have referred to what can be comprehended under L and SO.92 They have not discussed 'time? (kāla ) as a ‘kāraka'-relations 93 separately. But, they have 91. The case categories So, Go have a wider scope which is discussed in II. 4. 5. and II. 4. 6. respectively. 92. Roy, S, (1920), pp 57, 92: Pānini's rules 'dhruvamapāye - apādānam' (I. 4. 24), ādhārah - adhiharanam' (I. 4. 45). Cf. Vasu, S. C, (1891a), pp 177, 184 : Pānını's rule I. 4.24 : 'a noun whose relation to an action is that of a (relatively) fixed point from which departure takes place is called 'apādāna', 'Panini's rule I. 4. 45. 'that which is related to the action as the site where the action is performed by reason of the agent or the object being in that place is called 'adhikarana' or location.' 93 Cf. Subramania Iyer, K. A., (1969), pp 322 24, the section on "the abode (adhikarana)'. However, in Panini's rule II 3 64 'kſtvorthaprayoge - Lāle-adhikagane' the word 'Lale' (temporal) modifics adhikarane'. But, it is not a rule trcating temporal location in general. One may argue that Pán!ni's rule like 1.4 45 (f. n 91) includes temporal location as well. But the commentators do rot gise any instarce, urder that rule, with temporal location Cf. Roy, S., (1920); Vasu, S. C, (1891a), Bhattojidilfita.

·CAST-CATTGORILS 111 tatr1 definitely calcn notc of spītio-ecmporal distinction, and, in some of the rules, the terms denoting time and space, occue together." JI. 4.2. The casc-categories mentioned abosc liatc bccn referred to as the 'local' functions of the category of casc,95 which include temporal as well as spatial distinctions.95 Thus is thc rcason for their being considered as 'concrctc' m$c-listinctions as opposed to others which arc called grammatical' or 'abstract'.P8 Some linguists have accognised a general distinction of locatic vs. directional as a particular manifestation of a morc gencral distinction between static and dynamic'97 (1. c. in/at vs. to/form). 94, Rog, S. (1920) Vårstuka on Pinini's zulc I. 4. 31 : 'jatah ca - adhiahilanirminam prñcami, tadyultāc adhvanah prathamāsaptamyau, halit - caprami - - Vaczi': the starting point in the measurement of the length of time and space takes 'pancami', the space measured tahes prathami' and 'saptami', and the time tahes 'pañcami' (PP 62-63). Panini's rule II 35 : 'Laadhranoh- atyantasamyoge- dvuya': time and stage of journcy cake dystná' shen pervasion [1 c. continuity or extensionht) sumplied' (PP 35). Pānis rule II 3 6 'apavargc-trtiga' f 'attainment of the end has to be indicated, time and stage of journey tale 'trusă' when there is pervasion' [ic continuity or extension Ls Jas well (pp 38 39). Cf. Subrahmanya Sastri, P S, (1957a), pp 116 117 Vartola on Panini's rule I 4 51 "halabhāvādhvagantavyāh - harmasamini - hyakarmanam' time, duration and the distarce to be travelled take 'karmasamjña' with reference to 'alarmahadhātu" time-'mäsam - ăste / suapiti' { he stars / sleeps for a month box), duration : 'godoham aste/ svapıtı' (he stays / sleeps during the time of milking), 'krosam- aste / sv apiti'.(he stays / sleeps for two miles) Also sec rule II 37. : 'saptami - pañcampau - Lárakamadhye. 95 Lyons, J, (1968), P 298 96 Lyons, J. (1968), pp 295, 299 97. Lyons, J. (1968), 300 CF Bennett, DC, (1970), pp 28ff. .

·112 SANSKRIT SYNTAX This distinction will not be considered independently with reference to the above-mentioned case-categories in this analysis It is indicated (f n 91) that the case-categories SO, GO are not related to the spatio-temporal axes only (see II 4 4 5 and II 4 6) There have also been attempts to interpret the case-categories in terms of a localistic approach 98 Fillmore 99 refers to one case-category locative, and assigns the distinctions like the one between 'in London/to London, to associated verbs This treatment can, at least, be considered sketchy However, in a recent paper100, he has discussed this case category afresh and has postulated in its place the case-categories Place, Time, Source and Goal Following Bennett103, he suggests the case-category Pa(th) also 202 Lehiste has mentioned a case-category like Terminative, 103 which corresponds to Ext(ent)104 proposed here, as the latter term seems more appropriate II 4 3 The distinction between case-categokies Place/ Location and Time is not retained, henceforth There is more than one reason for this In the first instance, if the distinction is to be maintained uniformly, then this would applyto SO, GO and Ext as well, and, thus, the number of case categories would multiply, as in sentences 1-4 the case categories would have to be specified as place SO/GO, 98 99 100 Anderson, J M, (1971) Tillmore, C ), (1968a), pp 25 26 Tillmporc, C J. (1971)' pp 251, 258 61 he mentions 'Place, Time' (p 251) and then changes over to 'Location Time' (P 258) Berncit, D C (1970), p 2, alio in Semctica V 1 (1972),p 59. Tillmore CJ. (1971) p 259 I chiste, 1, (1969), p 334 Bernett also suggests Ext 101 102 103 104 ·CASL-CATEGORIES 113 1 time SO/GO, place Ext, time Ext, without any consequent advantage in description There appears to be no significant syntactic evidence to justify this distinction despite the obligatory occurrence of spatial/temporal Locative or Ext with particular verbs Because, such verbs allow temporal/ spatial Locative also in their case-frames, in addition to obligatory spatial temporal Locative or Ext, as in sentences 5, 6 yatrilah-madhyāhnāt-[sāyam yāvat] - abhramat A SO GO traveller - from midday - evening till - wandered (the traveller wandered from middaytıll evening) 2 mantri - chapra nagarāt - patna nagaram - ayat А SO GO minister - from city of Chapra - to city of Patna went (from Chapra, the minister went to Patna) 3 sohanah dinam apathat A Eit dav-long read (s read day long) 4 Sa hrośam ayat A Eut he two miles Went (he went for two miles) 5 sa sāıkılayānam - aNgane - (sāyan) nyasati A O L he -bicycle in yard - (in ei ching) - Leeps (he keeps the bicycle in the yard (in the ctening)) 6 (nagare) melamāsam sthāsyati L 0 Ext (in town) fair month long (thc fair will last for one month (in the town)) Secondly, Tillmorc's acceptance of this distinction between - S - will stay 8 ·114 SANSKRIT SYNTAX place and time is motivated by his 'one-instance [of one casecategory ) - per- clause principle." 105 In a sentence lihe 7 there could be occurrences of both place and time. 7. chätrah sandhyāyām vidyālaye aste D L(time) L(place) student in evening in school stays (the student stays in the school in the evening) However, this could be considered as an idiosyncratic aspect of the case-categories discussed in this sub-section (II 4). There are some other characteristics of L(place in Fillmore s sense) and Pa(th) which are peculiar to them (see II 4, II 4. 8), but on the basis of these no further specification of case-categories has been proposed Thus, it does not appear unreasonable to ignore this occurrence of both place and time (which is specific to L, SO, GO) in a simple Sentence 106 II 4 4 The case-category L covers both spatial and temporal references. In a sentence like 7 above, thus, there is more than one occurrence of L Another characteristic of I lies in the fact that it may involve multiple self-embedding 107 Thus, in a sentence like 8, the three occurrences of L are to be considered as unitary, one within the other, and the NP 'pādapasya nichah’103 (of tree-below) indicates 105 106 Sillmore, C ), (1971), p 248 It may be argued that as no verb can have both L (time) and L(space) as obligatory parts of its case frames, one or the other may be considered as part of a superordinate structure But, this would involve treatment of complex structures and as such is not 10 corporated in this analysis Stratton, CR, (1971), p 226 IV 32 1.17 This NP is a complex structure insolving inalienable clationship, the schema is NPN (O), the NP being dominated by Lin this case 107 108 ·115 CAST-CATEGORIES the inalienablc rclation between the tree and the direction 'nicash' (below). 8. yuvati - upavane-[padapasya nicaih] - phalale - А L L 8 L tisthati young lady-in garden-of tree bencath-on beach-sits (the young lady sits on a bench under the tree in the garden) Sentence 8 has the underlying structure given in figure I. Fig1c I S M P - [+present] tisth NP K NP K yuvati NP phalaka NP NP NP nicaih upavana NP padapa ·116 SANSKRIT SYNTAX - noise- Now, if the three occurrences of L in sentence 8 above, are not interpreted as one embedded within the other, the interpretation of the sentence would be unacceptable Similarly, in a sentence like 9 all the three occurrences of temporal reference function as one contained within the other. 9 adya - tātrau - dasavādanasamaye - kolāhalah L L L 0 ajāyata to day - at night - at ten hour time :-became produced (there was noise at ten at night today) But, in a sentence like 10, the two occurrences of temporal references must be interpreted as instances of coordina- The occurrence of 'ca' in sentence 10 indicates coordination 109 10 vrddhah - pratyūse - sandhyāyām - c2 - bhramati А L L old in dawn - in evening and - walks (the old walls at the dawn and in the evening) II 4 4 1. Now consider sentences 11(a c) in which I is a complex structure 11 (a) [gehāt-dahsına dahsınāhij samipam) -gauh-cacatı LO ab from house -south nesc grazes (b) (gehasya - daskinena / daksınatah / samipam] LO gauh - carati tion - COW of housc - south -ncer - COWgrazcs 109 humar, B S, Coordination in Sanskrit (forthcoming) ·117 CASE-CATIGORIES - near - nel COW - may have (c) [geham -dahsinena samayā / nihasā] - LO acc to house - south gauh - caratı grazes threc=(the cow grazes ncar/south of the house) In the sentences 11a-c, lihe sentence 8 above, L has a comples structure, and is to be interpreted in accordance with a rule-schema NP> N(O) (f n 108), where the relation between the N and is inalienable. The O is realised differently at the inflexional level (It is not genitive in all the three sentences 11a-c) which is partially a matter of free variation and is partially conditioned by the features on the lexical entry of the item concerned 110 Consider a sentence like 12 in which L an undecly.ng structure following the schema NP+ N(D)11, as in sentence 13 12 [suśılah chátrah] gurau vasati А L well - behaved - Student - in teacher - (a well behaved student lives near/obeys (his) teacher) 13 (susilah - chatrah] - [guroh - samīpam ] - vasati А L D g well-behaved - student - of teacher - near - lives (a well-behaved student lives near (his) teacher) II 4 4 2 Consider, next, sentences 14,14,15,15 which have different surface realisation of the L, and which is not a complex-structure like the one in 12, 13 In sentences 11a C, the lexical item 'dakşına' is distinguished by the clements ‘a/ahı (in 12a), 'ena/tah' (in 11 b), and 'ena' (in 12c), which condition the occurrence of the different inflex- In the case of 'sumipam' 'samaya', 'nikasa', the lexical Items themselves condition the inflexional variation The schema NP + N(D) accounts for 'guroh - samipam" (in 13') (teacher s proximity), wherein the NP is dominated by L Cf IV 3 2 1 1 7 1 lives 110 1ons 111 ·118 SANSKRIT SYNTAX 1 bālā in bed 15 ſupa- vasati ACC 14 bālā sayane tisthati / sete / aste A/D nom L girl in bed sits / sleeps / stays 14. -sayanam - adhitisthati/adhısete / adhyāste A/D nom L acc girl sits, sleeps 1 stays both=(the girl sits / sleeps / stays in bed) 15 tājñi rājabhavane vasati A nom L 1 queen in palace lives rājñi tājabhavanam A nom L anuadhi. queen in palace lives both (the queen lives in the palace) As mentioned above, it may be noted that the difference in inflexional realisation is conditioned both by the leucal items themselves and the features on the lexical entry of the item concerned Thus, sentences like 14' which contains a verb 'asyatı' (places) (which is other than the verbs mentioned in sentence 14), with the pre-verb 'adhi' (as in sentence 14' above), and 15' which has a verb 'vasatı' (as in 15 above) but with the pre-verbs 'ni' and 'pratı' (other than those in sentence 15 above), do not show variation in surface-rcalisation of L Sentences 14", 15" ceptable 14 mātā balalam sayanc adhyasyatı R L (the mother places the boy ia bed) 14 mata balalam śayanam adhrasyatı R L mother boy places arc U1Cacc in bed ·119 CASE-CATEGORIES lives - acc pratis 15". rājai rājabhavane s ni 7- vasati L prati queen in palace (the queen lives in the palace) 15"", rājñi Iājabhavanam ni 7- vasats L queen in palace lives II 4 4 3 It has been mentioned earlier (II 1.1) that the subjectivalisation-rule normally selects A/D for agreement with the V. But, consider a sentence like 16 wherein L, the only case-category in the frame, is selected for subjectivalisation. 16. sthāli (svayameva) L pot (itself) cooks (the pot cooks by itself) Sentence 16 has the underlying structure as in figure II. Figure II S pacatı112 nom M [+pres ] pac PP NP K sthali 112 svayameva The verb 'pacati' has also a case frame like (A, O, 1–) as exemplified by a sentence like 'sudah - sthalyam bhaktam - pacati' A L cook - in pot boiled rice - cooks ·120 SANSKRIT SYNTAL The process applied to sentences 10, 12, 16, 19(in II 2 1 2) operates, and the final form (omitting intervening details) is as represented in figure III Figure III S sthāli svayameva 16 * boiled rice pacati A rule deletes 'svayameva' optionally It may be noted that in constructions like 16 above, if any other case-category co-occurs, L can never be selected for subjectivalsation (16 below is unacceptable). A sentence lihe 16 conveys the sense of ease or effortlessness. sthāli odanam pacati L pot cooks II 4 5 The case category SO(urce) is considered as the 'starting point '113 of an action of process 'The category SO may be assigned to entities which are animate Consider sentences 17-23 17 Sayanagarat - ayat SO ab shc from city of gajā (she went from the city of gaya) 113. Tillmore C ), (1971), p 259 his previous definit on the place to which something is directed c(1969) P (116) cocs rot appear appropriate In the first instarcc, place'makcs at limited in its scope and sccordly '10 which cmeth rg is duscctci' appears to refer to GO rather than to SO CASC- Orinanimate S2 Went ·121 CASL-CATEGORIES nom 1 18 vihagal gulmát dayate A SO ab bird from bush flies (a bird Ales from the bush) 19 nadi himālıyāt pribhavati SO ab Ilver from Himalayas appears (the river rises in the Himalayas) 20 chātrall guroh bhásıhım adhite А SO ab o acc student from teacher - linguistics - Icarns (the student learns linguistics from the teacher) 21 rāmah hausalyāyāh ajāyata D SO ab from h became born (r was born from h) 21 rāmah hausalyāsām ajayata D SO (1 in became born I was born from L) 22 tāmah - daśarath at - ajāyata D SO ab from d became born (r was born from d (1 e father) ) 22 + rāmah daśarathe - ayāyata D SO in d became born [Sentence 22, but not 22, is acceptable The acceptability of Sentence 21 , as opposed to 22, perhaps, relates to the knowledge of the world 1 e mother, not father, bears the child in her womb ] 1 1 - 1 ·122 SANSLRIT SYNTAX 23. asti - adyatah avakāśah SO Es from today holiday is (there is holiday from today). The case-category SO, as indicated earlier (II. 4. 2.) need not refer to temporal and spatial entities only. In sentences 20-22, the case-category SO does not refer to time or place. The case-category SO is distinguished from A/D/Es/R/O/F/I. It is distinct from A/D/E/O as it can never be subjectivalised. Secondly, so presupposes the occurrence of A/D/Es/O in the case-frame as in the sentences above. In certain constructions 0, 1 (sentences 10, 12, 16, 19, II 2 1.2) and L as well (11. 4 3. 3.) can be selected for subjectivalisation, but SO can never be. So is distinguished from R/O/F/GO also in respect of the operation of the process of passivisation, as this process never applies to SO. II 4 5 1. The case-category SO which occurs in sentences 17-23, characterises the entity from which an action process which has been initiated by another actant (1 c. A/D and in certain constructions Es/O), emanates. But, there are const- Iuctions like 24-28 where this definition does not help, and the underlined NP's pose a problem for case-specification. 24 simhāt bibheti SA nom ab 25. hc from lion - fears (hic fears from the lion) snuşă śvašurāt jihreti nom daughter - in - lax - father - in-law-fccis shy (the daughter-in-lax fccls shy of her father-in-lan) ab ·CASE-CATLGORILS 123 26. A nom from cat 27. A nom müsah Otoh niligate ab mouse hides (the mouse hides itself from the cat) mātā putram agneh nisedhati R acc ab mother from fire checks (the mother Leeps her son from fire) hrsalah gám śasyebhyah - vārayatı R acc ab farmer COW from crops - chechs (the farmer 1 the cow from the crops) son 28, - A nom There appears to be three ways for the specification of the NP which has not been assigned a case-category in sentences 24-28 above : (1) SO as the starting point or point of departure for an action process (as in sentences 24-28) is may be considered more abstract and the NP (in 24-28) specified as SO,(11) the NP not specified as a case in sentences 24-28 above, may be explicated in terms of morphological features only (1. e. the ablative inflexion may account (for a case-category or none at all) in the context of the verb concerned); (111) sentences 24-28 may be derived from complex underlying structures The first option implies not only extraction of common semantic eleme of verbs, but also abstraction of these common semantic elements at a level which may become uninteresting ultimatelylld (see II. 4 5 2. also), and as such is not followed in this analysis. The 114, Cf Subrahmanya Sastri, P S, (1957), pp 92 98 . he refers to Patañjali's views assigning the 'apadana - karaha' (compar ableto SO) to the unspecified NP's in the above sentences, which do not appear convincing and reduce the different lexical items to a point where nothing is left of their characteristics as indivre dual lesicaliteris.

·124 SANSKRIT SINTAX third option has been followed by Sanskrit grammarjáns115 in the case of sentences like 25, which could be derived from a structure underlying a complex sentence like 25'. 25'. sausā Śyaśuram drsţvāl16 - R V Nom daughter-in-law father-in-law having seen jihreti Dom a CC feels shy (having seen her father-in-law, the daughterin-law feels shy) But, it is not clear how sentence 24 is different 117 from 25, and thus, justifies a derivation which is different from that of sentence 25(assuming that the derivation of 25 is correct). But, there appears to belack ofsyatactic justification to derivc sentences 24-28 from structures underlying complex sentences. Sentences 27, 28 could be related to sentences 27', 28' wherein the underlined NP is complex in structure, but the problem of case-specification remains as it is. 27'. mātā - putram - agneh - samipat nisedhati A R Og ab mother - Son - of fire - from proximity - checks (the mother hceps her son from the proximity of fire) 28'. Lrşahah - gām - śasyānām - samipāt - vārayati A R 0 S ab farmer cow - of crops - proximity - checks 115 Roy, S, (1920), p. 62 Vārtika on Pinini's rule I 4. 31. Cf. Ros, S, (1920), p. 62 : the Sanskrit grammarlans Tould like to assign obligatorily a pre-scrb to VNom in 25' which would lead to some morphological sariation (t. c. 'eamdya" in the place of 'drfra'), thicbinokes ro changcin the meaning. But, it is difficult to see any sontactic or semantic justification for the position of the Sanskrit grammarians. The Sanskrit grammarians have rof proposed a deritation comparable to that of 25, in the case of scnicnce 24. 116 117.

·CAST-CATEGORIIS 125 (the farmer keeps the cow from the prollmity of crops) The second option mentioned above, amounts to the acceptance of the inadequacy of the case-2otion for certain constructions in other words, the case-category SO is assigned to the underlined NP's in sentences 24-28 on the basis of the ablative inflexion in relation to the verbs Or, alternatively, no case-category is specified, the NP's are explicated with reference to the ablative-inflexion conditioned by the verbs concerned However, consider sentences 29, 30 which appear to come from structures underlying complex-sentences the 29, 30. 29 D nom acc 29 yuyatı - landana tāvarāt - bālaham - pasyati ab R young lady - from London tower - boy - loohs (the young lady looks at the boy from the Londontower) yovati - landanatāvare -sthitva-bālalam-paśyatı D L 1 VNom R young lady - in London tower - staying - boy loohs (the young lady, staying in the London-toner, looks at the boy) guruh - āsanāt śisyản - upadišati А ab teacher - from seat - pupils - preaches (the teacher, from his seat, preaches to the pupils) guruh - asane - sthitvā - śışyān - upadıśatı L 1 VNom R teacher - in seat - having stayed -pupils - preaches (the teacher, staying in his seat, preaches to the pupils) 30 Racc 30 ·126 SANSKRIT SYNTAX There are reasons to derive 29,30 from the structures underlying sentences 29' 30' respectively. In the first instance, the NP's with ablative inflexion (i. e. 'landana-țāvarāt' and Sāsanāt' respectively) limit the locational reference to 'yuvati' and 'guruh’in 29, 30 respectively (cf. sentence 29' wherein the NP "landana- țăvare' may be related to both 'yuvati and, bālakam', and sentence 29" which is unacceptable); and, thus, will be derived from 'landana-țāvare sthitvā' and, 'āsane sthitvā' respectively (in sentence 29', 30'). 29". yuvati landanațāvare bālakam-paśyati D L 1 R young lady - in London-tower - boy - looks (the young lady looks at the boy in the Londontower) acc 29**. yuvati - landanațāvarāt-bälakam- D ab R young lady - from London tower - boy - landanatāyare sthitam paśyati L VNom in London - tower - staying - looks (from the London tower, the young lady looks at the boy staying in the London tower) Secondly, consider sentences like 30, 30". If sentence 30* is interpreted like sentence 30", the NP's "āsane' and 'āsane" (in sentence 30') cannot be co-referential. It indicates that the NP (āsanāt' (in sentences 30"") could only come from a structure underlying 'āsanc-sthitvá' (in sentcnces 30', 30'). 30. guruh isane? - sthitvi - śisyan - äsane - L 1 V Nom R L 1 ·127 CASE-CATEGORIES teacher - in seat - having stayed - pupils -in seatsthitān - upadıśati VNom staying - preaches (staying in his seat the teacher preaches to his pupils staying in the seat) 30". guruh - āsanāt - śisyān - āsane-sthitān- А ab R L 1 VNom teacher - from seat - pupils - in seat - stayingupadıśati preaches (from his seat the teacher preaches to the pupils staying in the seat) The above treatment of sentences 29, 30 raises the problem of the postulation of some form of abstract verb in the place of the VNom in sentences 29:-30', and though there are severe restrictions with regard to items that can occur as VNom in such constructions, the justification for the postulation of one or the other V Nom is not clear at present. However, with this treatment of sentences like 29,30, assigning the case-category SO to the NP's unspecified in sentences 29, 30, is not needed. 118 II. 4. 5.2. Now, consider a sentence like 31, which contains a SO with complex structure (cf. SO in 17-23). 31. [ gehāt / gehasya - antihāt] sevakah - yāti SO O ab ab A nom from house/of house - from near - servant - goes (the servant goes from near the house) The case-categories are specified with refererce to the underlying structures of scnterccs. Assuming that senterces 29-30 arc derned from structures urderlying 25', 30' there is no need to specify case-category SO in 29-30. 118.

·128 SANSKRIT SYTAS ACC m - осс son The comple, structure of SO is to be interpreted in terms of the schema NP + NO) indicating, thereby, the 10- alienable relationship between N and O 119 Consider, next, sentences 32, 33 below and 20 above in which one might argue for the equation of SO with A. 32 mohanah - rāmāt pustalam - hrināti nom SO ab 0 from f. book buys (m buys a book from 1) 33. putrah - pituh patram - grhnāti A SO ab from father - letter tahes (the son tahes a letter from his father) There are, however, reasons for not accepting such an analysis. In the first instance, it has been mentioned (II 4.5 ) that there is need to distinguish SO from 1 because of the reasons given therein. Secondly, such an analysis presupposes a componential treatment of verbs like 'adhite' (leans), Shrinatı' (buys), 'grhnāti' (tałcs) (in sentences 20, 32, 33), and simultaneously tales into consideration 1cobs lhe 'adhyāpayatı'(teaches), 'sikrināti' (sells), 'dadāti' ( ges) respectively:120 But, in this analysis a serbal form is taken as onc complete leucal unit (no componential analysis is taken into account, and case-categories are specified with reference to this verbal form, 1 e without taking into consideration any other serbal form related or othcruise. Thirdly, in broader scnsc, Agentisc), in any case-frame, could be substituted with SO, or SO with Alif SO refers to 119 120 CF IV 3 2 1 1 7 Arder on, JH (1971) ·CASE-CATEGORIES 129 entities which are [+animate], as A 15 the initiator of the action denoted by the verb, though this is possible only when other verbal forms are taken into consideration) But, this substitution (of A with SO or vice-versa) does not appear to be appropriate One distinction between A and SO (in case both refer to entities which are [+animate] as in 20, 32, 33) lies in the fact that A is the initiator of the action! process, but SO is not (with reference to a particular verb), and this is why in such constructions SO presupposes the occurrence of A in the case-frame. If SO in sentences 20, 32,33 is re-specified as A, this significant distinction between A and SO is missed Fourthly, the verbs 'adhyāpayatı' (teaches) and 'dadati'(gives) do not always imply verbs like 'adhite' (sentence 20), and 'grhnāti' (sentence 33) Thus sentences 34, 35 are possible 34. guruh-sisyam-adhyāpayatı-parantu-sa-no-adhitc А R teacher - Student - teaches - but - hc-not learns (theteacher teaches the student but he does not Icarn) 35 rāmah - mohanāya } -pustaham-dadāts-parantu-sa - mohansya o GO - forfof m book gives — but hc no - grhniti not - tahes (c guies a book to m but he does not talc) It is possible that a teacher may not tcach a student, and șct hc my learn from the teacher by oucrhcaring him pect that cien in the case of 'hrināts shrināti, (buz-sell), 10 is not alsays that whenever sells tos, buys it. I I sus- 9 ·130 SANSKRIT SYNTAX m. ini. - sa - II. 4. 6. The case-category GO has been referred to as 'the destination'.121 Fillmore would like to assign this casecategory to an entity without any specification of animacy. Ia sentences 36-40 the case-category GO co-occurs with A. 36. sevakah nagaram / nagarāya - yātı GO acc servant — to town / for town goes (the servant goes to the town) 37. mohanah rāme — pustaham — vihrināti А GOI O book sells (m. sells a book to 1.) 38. - pryāyām patram - presayati А GO 1 he - in beloved — letter sends (he sends a letter to his beloved) 39. pitā — putraya -- pustakam -dadāti A GO d O father book gives (father gives a book to his son) 40. guruh - chātrebhyah/chātresu - pustahāni - vitarati A GO d 10 teacher-for students in students-books-distributes (the teacher distributes books to the students) The case-category GO presupposes the occurrence of A. One characteristic that distinguishes GO from AD (and also 011/L) relates to its non-selection for subjectivalısa- Thus a sentence lshe 41 would be unacceptable. 41.. pitcăpitā putrah-pustakam-dadāti/diyatc AIA nom GO nom o acc by father/ father -- son — book gilcs - to son - tion. 121 Fillmore, C J. (1971), p 259.

·CASE-CATEGORIES 131 O nom 40** gurunā acc The case-category GO is distinct from R/O/F also. If the subjectivalisation-rule does not apply to A/D, and R/O/F along with GO, is present in the case-frame, the process of passivisation follows, and R/OJF, not GO, 18 selected for agreement with the V Thus, 40 and not 40", is acceptable. In 40, O rather than GO has been subjectivalısed. 40 gurunā - chātrebhyah / chātresu - pustahāni- A GO d GO 1 by teacher - to students / in students - books vitiryante are distributed (books are distributed to students by the teacher) - chātrāh - pustahāni - vitiryante А GO nom O by teacher - students - books - are distributed However, a sentence like 36 above is an exception to this characteristic of GO If the subjectivalisation-rule does not apply to Ain sentence 36, the process of passivisation operates, and GO is optionally subjectivalised. But, there is one condition that in such constructions containing verbs like 'Jati' and its equivalents, GO does not co-occur with F/R/O. Again, this exceptionality of GO is valid only when it occurs in the case-frame of verbs which are parallels of 'yāti' (80) Thus, sentence 36 is the alternative form of 36 above (in case subjectivalisation rule has not applied) 36 nagaram GO nom by him is gone (it) is gone to the town bi him) The casc-category GO, like SO, may be assigned to NP's which do not refer to time and place, as in sentences 37-40 tena jajate town - ·132 SANSKRIT SYNTAX servant - Dear GO Dg N({0}). Secondly, the NP specified as GO may be a comples structure as in sentences 42, 43. 42. sevahah - ( gehāt / gehasya antikam] - yặt: GO O ab O 8 - from house / of house goes (the servant goes near the house) 43. putri — [ pituh -antikam) -- yāti А daughter - of father - near goes (the daughter goes near her father) The case-category GO in sentences 42,43 above, is allowed for by the schema NP II. 4 6 1. Consider, next, sentences 44, 44", 44' in which the GO has different surface-realisations which account for some semantic distinction. It is not clear, at present, how to get round this problem. 44 kāmuhah / sa — - dāsyā - dhanam - samprayacchate А nom nom GO 1 O lecher/he - with maid servant - money - gives (a lecher/he gives money to the maid-servant) 444 dāsyai dhanam samprayacchati Anom GO d O he to maid servant — money — gives (he gives money to the maid-servant) 44• * Kamulah — dāsyal dhanam-samprafacchati А GO d 0 lecher - to maid-servant - money - gives In sentence 44, the GO is realised inflesionally as 'i' and in 44' as 'd' along with the difference in the verb-ending (ie it is 'te'in 44 and 'tr' in 44') These surface-distinctions impose meaning-distinctions (sentence 44,44') which be culture-specifi acc sa — асс nom acc may ·CASE-CATEGORIES 133 A nom for son A nom acc son II. 4. 6. 2. In this section, a set of sentences (45-49) which contain NP's (unspecified with reference to any casecategory) ending in 'd-inflexion (cf. GO in sentences 36, 39, 40), and which can be derived from complex structures underlying sentences 45' - 49', are considered. 45. mātā putrāya — śete d mother -sleeps (the mother sleeps for her son) 45'. mātā putram prinayitum — śete. R VNom mother to please - sleeps (the mother sleeps to please her son) 46. kanyā — chātrāya tisthate D nom d girl for student stays (the girl stays for the student) 46'. kanyā chātram prinayitum — tişthati D nom R ACC VNom girl student to please stays (the girl stays to please the student) seyahah phalāya - yati А d for fruit goes (the servant goes for fruit) 47. Stvakah - phalam aptum — yāti A nom 0 VNom fruit goes (the servant goes to get fruit) 48, gurave pustakam ānayati d O hc for tcacher book brings (hic brings a book for his teacher) 47. scrvant асс servant - to get acc ·134 SANSKRIT SYNTAX - acc Svarnan acc cardies nom acc CarciCS 48', sa — gurave dātum pustaham- īnayati A nom GO d VNom he to teacher to gie- booh brings (he brings a book to give it to his teacher) 49. syarnahārah — hundalāya nayati A nom d goldsmith for earring gold (the goldsmith carries gold for earung) 49. svarnahārah-hundalam-nırmātum-svarnan-nayatı A F acc VNom goldsmith - earring to make gold (the goldsmith carries gold to make an earring) One may argue that Go defined as 'the destination'(II 4 5.), may be assigned to the unspecified NP's in 44-48 above But, the definition, as given by Fillmore, needs fucther amplification. The case-category GO is to be assigned to an entity which is an obligatory part of the case-frame of the verb concerned, in addition to its being 'the destination' to which the action is directed. In 44-48, the unspecified NP's are not an obligatory part of the verbs (in 44-48), rather they are obligatory part of an embedded S One might argue that there may be a scmantic interprctive culc to account for the meaning (as indicated in 44-48' above) But, if sentences 44-48 arc deu cd 112 44 -48', the problem of postulating an interpretisc rulc is avoided, and, besides, there is no problem sith icgard to case-specification However, despite the restriction on the items that can occur under VNom in scntcnccs 44-48 (cf sentences 44-48below), there remains tlic difficulty of ICCovering 123 the deleted items (specificd as VNom in sentences 44 -48 ) Chomiks, N. (1965), PP 132, 72({ n 1) 12 ·CASE CATEGORIES 135 45° mātā - putram-* krodhayıtum (to make angry) -sete mother - Son rodayıtum (to make weep) sleeps khădayıtum (to feed) hāsayitum (to mahe laugh) tādayıtum (to beat) pāthayıtum (to teach) svāpayıtum (to make sleep) VNom 46. Lanyā - chātram -* , āhvayıtum (to call) -tisthate girl - student dátum (to give) stays tādayitum (to beat) dhartum (to catch) V Nom 47". sevahah-phalam-* ( hretum (to buy) - yatı servant - fruit khăditum (to eat) goes vihretum (to sell) ganayitum (to count) datum (to give) vodhum (to carry) ānetum (to bring) paktum (to cook) chettum (to cut) VNom 48". $a -gurum- *< darśayıtum(to show -pustakam-anayatı he - teacher - (pāthayıtum(to teach 3-book -brings VNom 49. svarnakārah-Lundalam-* ſvihretum (to sell) goldsmith - earring kretum (to buy) -gold kseptum(to throw) chettum (to cut) -nayati drastum (to see) Vnom As sentences like 45 49 are derived from complex structures wideniging 45-49, the unspecified NP's in 45-49 necd not be assigned a case-category There has been a suggestion 123 to treat-for-phrases' (specified as Benefactives and Delegatives)124 as surface Noore, T. (1971) Moore, T, (1971), p 1 -Svarnam -Carties 123 124 ·136 SANSKRIT SYNTAX he -sa - phenomena and derive them from underlying Datives and Agentives. In Sanskrit, phrases comparable to 'for-phrases' specifying Benefactive ard Delegatire, are differently realised (as in 50-51 below). Sentence 50 cannot have the interpretation of sentence 51. 50. sa — - mitrāya --vāsaḥ — krināti - for friend - dress - buys (he buys a dress for his friend) 51. mitrāt prati - vāsaḥ - krināti ab PP from friend he - dress buys (he buys a dress on behalf of his friend) In so far as sentence 50 is concerned, it will have a derivation similar to that of 48 above. In other words, sentences like 50 will be derived from complex underlying structures and there will be no need to specify a casecategory in sentences like 50. Though the syntactic similarity (in respect of ‘movement around objectives and passivisaticnº)125 between Benefactive (for-phrases) and Dative constructions in English, does not obtain in Sanskrit, yet the treatment of sentences like 45-50 as derived from complex structures falls in line with the suggestion that 'for-phrases' and their cquivalents come from underlying complex structures. As mentioned earlier, complex sentences are not taken up in this analysis, and hence they are not discussed further. II. 4. 7. The case-category Ext(ent) has not been explicitly mentioned by Fillmore. But he hints128 at the possibility of 125. 126. Moore, T. (1971) p. 4. Fillmore, C. J., (1971), p. 260.

·137 CASE-CATEGORIES Eut acc till and necessity for such a case-category, in view of sentences lihe 52, 53. 52. sevahah dinam abhramat А servant day-long - wandered (the servant wandered day long) 53 nadi krosam vahati O Ext fiver - two miles flows (the river flows for two miles) The case-category Est can co occur with SO, GO as in 54. 54 chātrāvāsāt [vidyālayam yāvat] — [ayam - SO ab GO Es from hostel - school this mārgah]-hrosam Ext road two miles - 13 (from the hostel, this road is two miles up to the school) Fillmore argues 127 that Ext cannot occur with SO. But, in sentences 54 (above) and 55 (below) Est occurs with SO, GO, and, in 56, with L 55 adyatah [grismam yāvat] trin māsān SO GO Ert from today -- summer till three months satram -- sthasyati O session will continue (the session will continue, from today till summer, for three months) asti acc 127 Tillmore, C J. (1971), p 260 ·138 SANSKRIT SYNTAX 56 sa - -dine ghantādvayam—vyāharanam-apathat А L Ext he --in day - for two hours -grammar - read (he read grammar in the day for two hours) II. 4.7.1. Consider, now sentences 57, 58 containing the underlined NP's which are to be derived from compler underlying structures in sentences 57', 58'. 57. [ayam vidyālayah]-chātrāvāsāt — Krosahihtose Es SO this school from hostel two miles at two miles nom asti 1S two miles (this school is two miles from the hostel) 57'. [ayam vidyalayah] chātrāvāsat - krošam Es SO Ext this school from hostel atıkramya A-asti VNom L having crossed A-IS (having crossed two miles from the hostel, there is the school (in A)) 58 adyatah māse dipotsavah SO 1 Es from today - 1a one month - festu al of lights - is (the festival of lights is one month ahead) 58 adyatah - māsam - atihramya - dipotsavah - A - asti So Eat VNom Es L from festival of -L.- IS today month crossed lights (haring passed one month from today, there is the festival of lights in A)) asti onc haring ·139 CASE CATEGORIES The NP's underlined in 57, 58, do not mean that the school 15 two miles long or the festival is extending for one month, rather they indicate 'beyond two miles/one month,' and have to be interpreted in terms of sentences 57, 58, respecʻively It may be noted that sentences 57, 58, have a dummy 'L in the underlying structures (as in 57,58,) The NP's underlined in sentences 57, 58, may be left without case indication as they include the L as well (as apparent from 57,58 ), and, besides, case-categories need to be specified in underlying structures only Sentences like 57, 58, differ from sentences 51- 55(above) with regard to the nature of Ext as is evident from sentences 59, 60 Assuming that the mountain and the school are in the same direction, sentence 59 is acceptable, but sentence 60 is not This fact lends support to the position that the underlined NP's in 57, 58 come from the complex structures (in sentences 57, 58) 59 [asmāt sthånāt] - parvatah - hrosam So Es Eit from this place - mountain - two miles long - 1s vidyalayah - ca - hrošah, krośe - asti Es school and - at tao miles - 18 (from this place, the mountain is two miles long and the school is at tuo miles) 60 * (asmāt sthānāt] - parvatah -hrofam SO Ls Lrt from this place - mountain - two milcs - įs Vidyalayah ca- hrošam Es Cut school and two miles II 4 7 2 Considei, next, sentences 61,62 It is not clear hou the scmantic distinction between them, without asti nom • asti ast1 - 15 ·140 SANSKRIT SYNTAX 7 Est 1 acc Est acc refernog to morphological phenomenon, can be accounted for. In 61, not in 62, the interpretation indicates the completion or fruition of the action. 61. chātrah - dinena - upanyásam - apathat A O student-day-long - novel - read (the student read (completely) the norel during the day) 62 chátrah - dinam upanyāsam - apathat А student-day-long - novel read (the student read the novel day-long) The Ext in 61 is realised with an 1-inflexion, and in 62 with an acc It may suggest that a sentence like 61 could be derived from a structure like that underlying sentence 61', but it involves the status of adverbials, which is rot discussed in this analysis 61' chátrah - dinam -apathat saphalam student - day-long-novel - completely - read fruitfulls (the student read the novel day long completely/fruitfulls) A further distinction between 61 and 62 is related to negation 123 Sentence 61' is unacceptable (it may be noted that the adverbial in 61' cannot co-occur with negation), but sentence 62' is ambiguous. 61'* chátrah - dinena upany ásam .no - apathat Lst student - by day - novel not - read - Eu acc upangasam-jauhścsam 128 In this analysis constructions containing regat on harc fot tem discussed.

·141 CASE CATEGORIES nom Ра 1 1 1 62 chátrah - dinam upanyāsam - no - apathat A Ext acc 0 student - day-long - novel not - read (the student did not read the novel (at all) day-long)/ (the student did not read the novel day-long (but for a while)) II 4 8 The case-category Pa(th) has been suggested recently129, and is justified on the basis of constructions like 63, 64 63 handuhah - gavāksena - Tesmani - apatat 0 GO ball - through window - in room - fell (the ball fell in the room through the window) 64 sevahah - ucchistam - mahānasāt - gavāhsena - А o SO ab Pa servant - leftover - from hitchen - through window [prähārasya uparistāt] - mārge -hsipati Ра о GO 1 of fence from above - on road - throws (the servant throws the garbage on the road through the window over the fence) The case-category Pa is distinguished from SO, GO, Ext, in the sense that it can occur more than once in a simple sentence (as in 64 above) It is distinct from L also, as its re-occurrence cannot be interpreted like L in sentence 8) as an instance of multiple sclt cmbedding (c onc occurrence of Pa cannot be conceii cd of as contained within the other) Morcover, constructions lihc 64 cannot be interpreted as conjoined structure (cf sentences 10, 11) Tillmore, J (1971), pp 259 260 Cf Bennett, D C (1972), 129 pp 58 67 ·142 SANSKRIT SYNTAX It is one action which 15 associated with two Pa's in sentence 64 above It has been argued that this Pa case-category can be repeated indefinitelv,130 in a simple sentence. But, perhaps, it is possible theoretically only. In Sanskrit, it is doubtful whether more than two occurrences of Pa are possible without making the construction complex One important characteristic of Pa lies in the fact that it relates to spaceorientation only, and not to temporal orientation. II.4.9. Fillmore is of the opinion131 that a case-category like Ext may combine SO, GO in certain constructions. In sentence 65, SO and GO indicate that is Est in 66. Thus he proposes the possibility of a 'hypercase '132. But, as indicated in 55 (above), Ext can co-occur with SO and GO 65. [rāja-mārgah] - vidyālayāt - Es SO ab main road from school rājendra-mārgam yāvat - asti GO rājendra road (the mun road runs from the school to the rājendra road) nom acc - IS 130. 131. 132 Stratton, CR, (1971), p 226 Tillmore, C ), (1971), pp 260 261 Fillmore, C ), (1971), p 260 'Expressions of duration and distance introduce new orders of problems for a casc-analysis of verbs of mos ements and change because they somchow seem to combine the source and goal notions into a single unit, a 'hypercasc as it ucre But, as shorn carlıer, SO and GO refer to entities at hich are not restricted to scmporal «patial orien Secondly constructions lihc 61, 62 (1 4 7 2 ) COD:ain ing Ext, arc semantically distinguished as well Thirdly, Est an occur with SO ard GO Thus, Est annot be postulated in terms of a 'hypercase' combining SO and GO tition ·144 SANSKRIT SYNTAX both of them may be obligatory parts of a case-frame, but one or the other may be obligatory) to a superordinate structure. Besides, Fillmore is not explicit how this “hypercase' is to be specified in the superordinate structure (1. e. under P or S: it cannot be under M as dominated by 1). Thus, there appears to be no scmantic justification for this suggestion, and so, it is not incorporated in this analysis. cases are not ·III NP-ASTI-NP CONSTRUCTIONS III. 1. However, III. In this chapter, constructions which contain thc verb 'astı' (1s), are discussed and specificd as to the case-categories occurring therein. The discussion begins with a reference to the treatment of such sentences by the Sanskrit grammaflans and by linguists (section III 1.), and is followed by the analysis of the case-categories that can be assigred in such sentences (section III. 2) The Sanskrit grammarians analyse syntactic structures within the theory of 'laraha’-relations they leave out sentences like 1 which are mentioned as examples of the occurrence of 'sasthi-vibhaktı' (sixth-inflexion i e. g) 1. dhanikasya putrah asti 8 of wealthy (man) (wealthy man has a son, there is the son of a wealthy man) Such constructions are excluded from the analysis of 'Kāraka'-relations2 Direct relaticns between the denotations of two nouns are excluded ” Staal refers to such as 'nominal' sentences 3 However, his discussien does not son IS ܝܢ 2 3 'g'refers to genitive inflexion as opposed to Genitive) construction, which are discussed in chapter IV Staal, J F, (1967), p 38 Staal, J F, (1967), pp 52 56 68 10 ·146 SANSKRIT SYNTAX 1. . -15 go beyond the word-order involved therein. Constructicas like 1 (above) and 2 (below) have engaged the minds of linguists, who have classified such sentences as 'copulative' and 'existential' 4 2. ramah - bālakah asti boy is (r. is a boy) The copulative class consists of sentences like 3-7. 3. ramah [ayam chātrah] - asti Es T this student is 4. ramah chātrah - astı : class-membership Es T student (r. is a student) 5. harinah - paśuḥ asti : class-inclusion Es T deer — animal (deer is an animal) 6. phalam — madhuram -- asti : property-assignment Es T fruit sweet (fruit is sweet) 7. bāliha -gthe astı : locational reference Es L girl in house - IS (the girl is in the house) It may be noted that the L in a sentence like 7 can be a complex structure as in 7.3 7 mardiram [uparanasya daksinam] asti Is LO templc of garden south is (the temple is in the south of the garden) Lions, ) , (1968), pp 388-90 Sce II, 1 3 1 is 16 5 ·NP-ASTI-VP CONSTRUCTIONS 147 -- IS - 1S vane - asti The above sentences are distinguished from an existential sentence like 8 8. svarah - asti god (there is a God) It is apparent that an existential sentence like 8 has one argument only , unlike sentences 3-7 But, as an eustential sentence 8 is an exception Because, a sentence like 9 is unacceptable unless it is in reply to a query or it has a locational argument as in sentence 10. 9. simhah - asti lion 10 simhah L Es in forest - lion - 15 (there is a lion in the forest) However, a sentence like 11 with a verb 'bhavati' (becomes) and no locational phrase, is possible Though, I suspect that sentence 12 as opposed to sentence 13, is not acceptable. 11. simhah-bhavati - parantu-elastNgah - no - bhavatı lion - eusts - but (lion exists but unicorn does not exist) 12 * simhah-bhat ati lion 13. Simhah - Jane bhayatı lion - in forest - exists (lion eusts in the forest) For the analysis of case relations, it appears difficult to justify the retention of this distinctcn between ccpulative (including locational) and evstential, specially kccping If the distinction needs in view sentences lihe 7, 7 and 10 -unicorn - not . exists exists 6 See II 1 1 (f n 4) ·148 SANSKRIT SYNTAX occur in to be muntained, it should be between sentences 36 ard 7,7, 10 ie copulative vs. existential ard/or locational Bach refers to such constructions as containing either on NP (in sentences 3-5) or an adjective (in sentence) or locational phrase (in sentences 7, 7' and 10)? Fillmore calls such constructions (3-7' above) 'nominal predicates' or 'N be N type's With slight modification, I will refer to such types, henceforth, as NP - 1sti - NP constructions III 1 1 The NP asti - NP constructions (37' 10), though considered distinct from each other as indicated by the characteristics mentioned on the right hand side above), are taken together for specification of the casecategories The case categories that may NP - astı - NP constructions are discussed below The case-category Es(sive) is assigned to the argumints underlined in sentences 3-7, 10, the other casc categories being T(ranslative) and Locative) In sentences 7,7,10, E, is specified for the NPs bāliha' (girl), ‘simhah' (1107) andʻmındıram' (temple) and L for the NP's "grhc' (in houss), 'upavanasya dahsınam' (south of temple) and 'vane' (in forest) In sentences like 7,7', 10, the L can never be selected for subjectis alisation Thus, a sentence lihe 14 is unacceptable ie there is no equivalent to the sentcncc 'thic house has a girl in it 14 * grham - bālikā house — girl In sentence 6, 'phalam’ (fruit) is Es In any construction 11hc 6, the adjccine (1 e 'madhuram' (sh ect)) cannot be Is asti -1S 7 8 9 Bach, C, (1967a), p 462 Fillmore C.), (1968a), p 84 L(ocative) 13 discussed in II 4 4 ·149 NP-ASTI-NP CONSTRUCTIONS -1S - asti 1S - on tree as the selectional features of 'madhuram' derive from the Es (“phalam") Compare sentence 6 6'. sitä madhurā — asti Es T sugar sweet (sugar is sweet) The adjective 'madhuram'ın sentence 6 cannot be considered a modifier of the Es 1 e 'phalam' Because, in that case there would have been another case-category in the case. frame (such as L in sentence 15) 15 [madhuram phalam) - pādape Es L sweet fruit (there is sweet fruit on the tree) If a sentence does not contain another case category (i e other than Es and the adjective), the adjective cannot be a modifier of the Es, whatever may be the word order of the Sentence In other words, the adjective will realise a distinct case-category and not be an expansion of the case category Es as in sentence 15 above In sentences 16 and 16 the adjective 'tarunah' (young) cannot be a modifier 10 16 tarunah asvah young - horse 16 asrah tarunah horse — young both (the horse is young) However, in 1 sentence lihe 17 the adjective dayaluh' may be interpreted as a distinct case category (1 c T) or as a It may be argued that irtonational pattern may account for predicative ard attributive interpretations of the same adjective But, as mentioned in the beginning, the pronatioral pattern is ignored in this analysis asti -1S asti -- IS 10 ·150 SANSKRIT SYNTAS asti IS modifier of ‘īśvarah'. But, this ambiguity arises due to the exceptionality (which appears to be non-linguistic) of a sentence like 8 above. 17. iśvarah -- dayaluh Es T god merciful (God is merciful / there exists merciful God) In sentences 4, 5 above, frāmah' and 'harinah' are Es, 'chātrah' and 'paśuh' in these sentences cannot be Es, because they take on the features of Es 1 e “rāmal' and ‘harinah'(cf sentences 24-30) Even if it is argucd thì there is a difference between 'chātrah' / 'paśuh' (sentences 4, 5) and ‘madhuram' (sentence 6) tahing on the features of Es, this does not affect the postulation of the same category T for all these 1. e 'chātrah, pasuh and madhuram' Besides, I suspect that 'chātrah' and 'paśuh' cannot modify Es 1. c. ‘rāmah' and 'harinah' respectively (ʻmadhuram'ın sentence 6 does modify Es but in that case it is not a case-category distinct from Es) Thus sentences 18 and 1911 are suspect. 18. ? chätrah rāmah - pāthaśālāyām - asti student 1. - in school 19.) paśuh harinah - mārge - asti animal deer - on road -15 Besides, Es in sentence 4 above, must be [+definite]. Thus a sentence lihe 20 is unacceptable, *hough 21 is acceptable. 20 * bälahah - chitrah - asti boy - student - 15 -15 11. It may bc asked whether 'chitrah rimah' ard 'paſuh harinahi (in 18, 19) are possible as appositionals I doubt the correctness of this position in face of the untenability of postulating rcduced relative clauses in Sanskrit (scc scntcrccs 53 54') ·151 XP-ASTI-NP CONSTRUCTIONS this boy 1s - are 21. ayam bālakaḥ - chátraḥ - asti - student - is It may be argued that sentence 20 is acceptable. But, it could be so only in a context which makes 'bālakah' [+definitc). In the case of sentences like 22, 'satyam' and 'saundaryam' both have the feature [+definite]. It may be noted that sentence 23 is unacceptable. 22. satyam - saundaryam - asti Es T truth - beauty (truth is beauty) 23.* satyāni - saundaryam / saundaryāni - santi truth - beauty / beauties In sentence 3 above the Es, 'rāmah' and 'ayam chātrah' are both[+definite). In such equative sentences either of the NP's may claim to be Es. There are two possibilities: either both NP's may be specified as Es or one of the NP's may be Es and a rule selects this Es for subjectivalisation and the other is T. The choice between these two options, I leave as an open question. Compared with sentence 3 above, sentence 3' contains "ayam chātrah' specified as Es. 3. (ayam chátrah] - rāmah - asti Es Τ' this student (this student is 1.) The case-category Es as mentioned above (pp. 146- 150), is assignd to NP's 'rāmah' (sentences 3, 4), 'harinah' (sentence 5), 'phalam’ (sentence 6), bālıkā' (sentence 7), ‘mandıram' (sentence 7') and ‘simhah' (sentence 10). This widening of the scope of Es is unlike that of Lehiste.12. It 1S 12. Lehiste, I, (1969), pp 332 33, Her Essive includes the underlined NP'sin sentences like (1) NN is our ambassader in London' (11) 'NN is the best student in the class' (111) 'NN is one of the best students in the class', only.

·152 SANSKRIT SYNTAX is difficult to say what Fillmore vould mean by Es, because he has mentioned nothing (except the term) in this connection.13 III. 1 2 Now, consider the case-categories that may occur in addition to Es in the above sentences. T(ranslative) is a case -category that may co-occur with Es. This case-category cin be assigned to NP's "ayam chátrah' (sentence 3), 'chātrah (sentence 4), 'pasuh' (sentence 5), 'madhuram' (sentence 6) This specification of the case-category T does not take into consideration the distinctions between these NP's mentioned on f. 146 There are two reasons for this. In the first instance, the features of Es determine the features of T. Compare sentences 24-37 with sentences 3-6 (above) respectively 24 sītā - [1yam chátrā] - asti Es T this student (s is this student) sītā : [+feminine], [+singular) iyam chātrā. [+feminine], [t-singular) 25 [te balahāh] - chātrāh - santi Es T those boys - students - are (those boys acc students) te bälahāh : (+-masculine], [+plural] chātrāh : [+masculinc], [+plural] 26 harināh - pašavah - santi T decr(s) animals (dcec arc animals) S. 1$ Es -are 13 Tillmore, C J. (1968a), p 84.

·153 NP-ASTI-NP CONSTRUCTIONS harınāh : [+masculine], [+plural] paśavah : [+masculine], [+plural] 27 phalani - madhurānı - santi Es T fruits (fruits are sweet) phalāni [+plural] L-fem madhurāni : C-mas 7. [+plural] Sweet are -mas. zas. ] However, there are exceptions as well. But, such NP's are lexically determinate to the fcatures of gender] and [number] as in sentences 28-30 28. sā - [mama mitram] - asti Es T she my friend (she is my friend) să : [ +-feminine], [+singular] ſ-mas 7'[+singularl] 1S mitram L-fem][= 1S my wives 30 sā - [mama dārāh] - asti Es T she (she is my wife) sā : [-+feminine], [+singular] darah [+masculine], [+plural] asti [+singular] Secondly, the NP's specified as T, follow Es with regard to subjectivalisation If Es is not selected for subjectiyalisation (as the subjectivalisation-lule applies optionally), T also cannot be selected for subjectiv alisation Sentences 4 and 4' (below) have an underlying structurc as in figure I ·154 SANSKRIT SYNTAX Figure 1 S — M [--pres.] V Es T - as NP K NP | rāma chātra Es is not selected for subjectivalisation Consequently, as T is never subjectivalised, no case-category enters into agreement with the V, and the process of passivisation without subjectivalisation (II. 1, II 2) operates The result is as in figure II. Figure II S NI + pres. + sing + third ES 14 bhù-ya- NP K NP K rāmasing chātra sting Lt-mas +mas. A morphophonemic sulc will substitute 'us' under Vouth 'bha' [imas] 14 ·155 NP-ASTI-NP CONSTRUCTIONS The affir-substitution rule replaces features by affixes. The final form is as in figure IV via figure III. Figure III M P te Es bhū-ya- NP K K NP | räma ina ina chātra Figure IV S Es - bhūyate rāmena chătrena Thus, the alternative form of sentence 4 is realised as 4'. 4'. [āmena - châtrena - bhūyate Es T by r -- by student - is become (it) is become student by 1.) Similarly, sentences 3, 5, 6 have alternative forms as 3", 5, 6'. 3. cāmena - [anena chátrena] - bhūyate Es T by r. by this (by) student - is become (it) is become this student by c.)

·156 SANSLRIT SYNTAX 5'. harinā - pasunā bhūyate Es T by deer - by animal - 1s become ((it) is become animal by deer) 5*. phalena - madhurena - bhūyate Es T by fruit - by sweet - is become ((1t) is become sweet by fruit) In all the above sentences (3" - 6") T could not be selected for subjectivalisation even in case of non-subjectivalisation of Es III 1 3 The second case-category that may co-occur with Es is Locative) 15 The case-category L includes temporal as well as spatial distinctions In sentences 7, 7', 10 above, the NP's 'grhe' (in house), 'upavanasya-dahsınam' (south to the garden) and 'vane' (in forest) are L The third case category that may co-occur with Es, is Source)16 Consider sentence 31 31 chātrah - Jhārāt SO student - from Bihara - Is (the student is from Bihar) IIT 1 4 The fourth case-category that may co-occur with Es is GO(ai) 12 In scntcncc 32 GO co-occurs with so as well 32 ſayam mārgah] - chātrālāsät - (udyalayam- Es SO GO this road from hostel • school astt Es 15 16 17 ror details sce II 4 4 for details sec II 4 5 for details scc II 4 6 ·NP-ASTI-NP CONSTRUCTIONS 157 IS - asti 1s evening till yāvat] - asti till (this road runs from the hostel up to the schcol) In sentences 33, 34 GO CO-occurs with Es and L, SO remaining unspecified 33 gha te [mukham yāvat] - jalam - asti L GO Es in pitcher brim full - water - 15 (there is water in the pitcher up to its brim) 34 guruh vidyalaye [sayam yāvat] Es L GO teacher - in school (the teacher is in the school till the evening) III 1 5 Another case category that may cu-occur with Es, is Ext(ent) 18 Consider sentences like 35, 36 35 [ayam mārgah] - krosam Es Ext acc this road two miles - IS (this road is two miles long) 36 godhūmah [daśa Lilogrāmah] Es Ext wheat ten hilograms (wheat is ten hgs in weight) The case category Es may co-occur with an NP whose morphological representation is marked by 'genitive'ls. as in sentences 1 and 37 37 (mama - putrasya pustahanı] - 1541 Es 8 of son book (my son has a book / there is the book of asti - asti nom nom IS nom 1S of my my son) 18 19 For details sce section II 4 7 ror details see chapter IV ·158 SANSKRIT SYNTAX Sentence 37 is ambiguous and may have both interpretations given above With either interpretation, the Es in a sentence like 37 is a complex structure. But, if this complexity of structure is accepted, a sentence like 37 should have a caseframe like [Es), with only one case-category therein. It is unusual to have Es as the only case-category, as a sentence lihe 8 above, has been considered exceptional. Now, it mar be that Es may occur as the only case-category in a caseframe, provided the Es is a complex structure containing an N marked with genitive III 2 It is crident from the above analysis that the casecategories that may be assigned to NP's in NP-asti-NP constructions are different from the case-categories like A/D/R/O/F. At this point it may be relevant to look into the suggestions of Fillmore regarding the NP-asti-NP contsructions Though he mentions20 that the case-categories in the case-frames of NP-asti-NP contructions are different from A/D/R/O/F, he does not discuss them in detail Their inclusion under the sub title 'Problems and Suggestions":1, perhaps, explains lus shetchy treatment. He refers to casecategories like Es and T, but docs not analyse them at all He suggests that a sentence equivalent to 38 may have the case-category A assigned to 'rāmal'. 38 ramah - mūrhhah - asti fool (t is a fool) He argues that, in certain constructions, items likc'mūckhah' (fool) bchase like a V occuçıngına case-frame lilc (1) r 1S 20 rupe Fillmorc c), (1962) p E4 . 'scrtcrces of the N be represcrt a distinct scrierce typefromthosc involving any of the case relations discussed aloic " Tillmore, CJ (1968a) 81 21 ·159 NP-ASTI-NP CONSTRUCTIONS But, this suggestion does not appear to be convincing as it is not clear, at all, why sentence 38 is different from sentences 3-6 above, regarding case-specification Besides, in sentences like 39 which has a non-evaluative predicate, 'mamabhrātā' cannot function like V, and as there is no syntactic difference between them it would not be proper to treat Tamah' in 39 as a case-category different from 'ramah' in sentence 38 39 rāmah - [mama bhrāta] - asti Es T my brother (r. is my brother) It appears that sentences 3-6, 38, 39 should have a similar treatment for case specification Fillmore says that some new case-category may be invented 22 But he does not suggest any himself His reference to Bach's suggestions23 does not appear to be helpful. Bach proposes to derive constructions like 'adjectivenoun' from constructions like 'noun-is-adjective'.24 But this proposal suggests nothing about the case-categories that can be assigned to arguments in such a construction Bach says that in sentences comparable to NPasti-NP constructions, verbs like 'asti' (copula) are 'distinguished syntactically from most true verbs by the fact that they have no selectionl restrictions in themselves "26 The selection reaches across from subject to object or complement 25 Usually the selection restriction is between subject and verb/verb and object/subject-serb-object. In a 22 23 24 25 rillmore, CJ, (1968a), p 84 Bach, L, (1968), pp 98 104 Bach, E, (1968), P 102, f n ? Bach, E, (1967), PP 476 77 ·160 SANSKRIT SYNTAX restriction 1 eats - asti 1. 1S case grammar, selectional restrictions operate between the V and one or more case-categories. In an NP - asti - - NP construction there does not appear to be any co-occurrence between fasti' and an NP. Consider the acceptable sentences 40, 42 along with 41, 43 which are not acceptable (It may be noted that sentence 43 is acceptable if a metaphorical sense is implied) 40. rāmah - phalam - khādati А fruit (r. eats fruit) 41.* rāmah – phalam fruit 42. ramah paśum - tādayati А animal beats (r. beats the animal) 43. rāmah — pasuh animal In the case of verbs, it should be enough to state co-occurrence restrictions in terms of the case-categories that may occur in their respective case-frames. But in the case of Casti' it is not so The cisc-frame (Es, T-] allows a sentence like 4 and also a sentence like 41. In other words Es and T remain at a level different from that of A/R/O (in sentences 40, 42). However, the verb 'asti' does precludc certain cisc-categories from occurring in its casc-franie. r(actituc) is one such case -category that cannot co-occur with the Es R t. asti r. -1S as in 44 44 * tandulah odanam santi Es rice(s) - boiled rice — are ·NP-ASTI-NP CONSTRUCTIONS 161 III. 3. In the preceding discussion, the case-categories that may occur in the case-frames of NP-astı-NP constructions have been specified with one assumption that 'astı' (15) is always present. But, in fact, this is not the case. In a sentence which is not marked for past or future, the 'asti' is optional in surface structure. Thus in sentence 45 (a) the time-element expressed in the sentence is simultaneous with the point of reference in time (1. e. speaking). Optionally, it need not be expressed at all as in sentence 45(b). In sentence 46, 'asīt' (was) refers to a time which is anterior to the point of reference (in time) and in sentence 47, 'bhavisyati' (will be) refers to a time which is posterior to the point of reference (in time). 45(a). [ayam janah] - adhyāpakah asti Es T teacher is (this man a teacher) (b). (ayam janah] - adhyāpakah teacher this man 1S this man [It is not relevant to this discussion whether sentences like 45/45 (b) involve a point of reference in time or none at all). 45' vane simhah in forestlion : () there is a lion in the forest. (11) lions inte in forests. 46. [ayam janah] adhyāpahah - äsit Es T this man teacher Was (this man was a teacher) 11 ·162 SANSKRIT SYNTAX 47 [ayam janah] - adhyapal ah bhavisynti Es T this man teacher will be (this man will be a teacher) The above mentioned characteristic of acui' l as assumed some significance in view of the differing interpretations by grammarians and logicians regarding tlc status of the verb in a sentence in Sansheit The Sanskrit giammariins hold the vien that every sentence('vahya') consists of a verbal form 26 Scme scholars interpret the relevant rule as referring to a verb which ends in an affix of the 'tiN' class only and not to a verbal form in general 27 Thus, a sentence lihe 48 which consists of a verb 'ayat' (went) ending in 'tıN' class affix (1 e is 'tidant) is accepted as a sentence 48 ramah ayat A I went This interpretation would bar 49 as a sentence as it contains 2 verbal form yatah' which ends in a 'sup' class affix (ie is 'subanta' like NP's) rather than a tin' class affix 49 ramah — yatah А 1 Went One distinction between 'tiNanta' and 'subanta' verbal forms, though irrelevant for specifiying case cate gories (and this is crucial) relates to gender The 'tiNanta' verbal form ‘ayat' in 48 is gender neutral, but the 'subanta' 26 Bhartiharı Vak; apadijan. I 1 2 Iyer K AS (1969) p 419 (61 19) Mahabhasya of Patanjalı under rule II 1 1 Subrahmanya Sastr), P S,(1957) vol V, p 204 chatıN vakyam Matilal B K (1966) p 377 Subrahmanya Sastr PS, (1957) vol v p 204 27 ·NP-ASTI-NP CONSTRUCTIONS 163 Verbal form 'yātah' in 49 is not (cf sã yatā : she - went) Even the scholars favouring the interpretation of the rule (f n 26) to include the 't1Nanta' verbal forms only, would treat ftiNanta' and 'subanta' as verbal forms for case-specification, Thus, the interpretation restricting the rule to "t1 Nanta' verbal form only, does not appear to be appropriate, and the relevant rule must be interpreted to refer to verbal form in general This interpretation would be in Leeping with the statement of Patañjalı,28 that there is no sentence which does not contain a verbal form It is significant that he uses the word 'hriya' (verb) rather than the word 'uuNanta' (1 e a form ending in a tiN-affix) The above mentioned characteristics of 'asti' (1 e its optional occurrence in certain constructions, and of a verbal form ending both in 'tjN-class and 'sup'-class of affres) were picked up by the logicians who differed with the grammarians on accepting a verbal form as a necessary constituent of a sentence The optional non-occurrence of astı'ın sentences like 45(b) above, became a strong argument for them They cited examples like 5029 to prove their point 50 trayah - kālāh three (there are three time-stages) They argued that a verbal form lishe 'sanu' (are) cannot be added (10 50), because 'santi' refers to one umestige only (I c present) Again neither 'āsan' (ucrc 1. c past) noc 'bhavisyanti' (sill be i c future) can be added - times 28 'ra hi hris virimuktam sikyam asti' quoted 18 Maulal, BA,(1966), p 378(f 2) JhalalShar, B. (1893) p 068 , R20, V S,(1969), p 13 29 ·164 SANSKRIT SYNTAX Because, either of them refers to one time-stage only To them, and perhaps to the grammatians also, it would be equally ludicrous to have a sentence like 50 50. trayah - halah -santi āsan - bhavisy anti three - times - are - were - will be Thus, they conclude that the inadmissibility of a verbal form in sentence 50 above, shows that a sentence need not consist of a verbal form However, in their arguments, the logicians do not tahe note of either physical reality wherein time may be conceived of as a continuum, 30 or of the fact that the reference to three time stages and the occurrence of a verbal form like 'santi' (are) in sentence 50 need not pin it down to a particular point of reference (in time) which is simultancous with the event of speaking In fact, 50“ is perfectlr grammatical and acceptable 50" trayah - Lalah - santi three (there are three time stages) The logicians, in arriving at their conclusion, were also influenced by ontological considerations in addition to the peculiarities of the Sanskrit language (1 e optional deletion of astı' in certain constructions) To them, substrace (dravya) was the repository of different properties 1 e qualities, actions 31 Consider sentences like 51, 52, which have different interpretations for logicians and grammarians 51 Tamah - vihagam - pasyatı A R bird looks (r looks at a bird) 30 Mandal hh,(1968), pp 156 73 31 Chatterjee S (1967) p 225 drayra or substance is the substratum of qualities and actions times - are 2 ·165 NP-ASTI-NP CONSTRUCTIONS m 52 maitrah - tandulān - pacati A -fice - cooks (m cooks rice) For a logician, the substance-the agent (1. e. ramah maitrah) is central in the sentences In other words, sentences primarily refer to 'rāmah”32 the agent who is 'qualified by effort generating the activity of seeing which has a bird as an object (vihaga-karmaka - darśanānukūlalttiman ramah') and the agent 'maitrah 33 who is qualified by the 'effort favourable to softening subsisting in rice ("tandulanısthaviklittyanukūla - hrtımān maitrah) But, Sanskrit grammarians, or perhaps grammarians in general would interpret the sentences, as if, they refer primarily to an action or operation In other words, to them the action 1s central in the sentence Thus, these sentences refer to an action of seeing which has bird as object and m as doer' ('vihaga - karmala darsanānuh ūla - vyāpāro tāma - kartrhah'), and to an action of producing softening in rice which has m as its doer (standulanısthah villittijanahah maitrābhinnaśrayakah vyāpārah') respectively.34 Thus, Sanskrit grammarians accept the verb as central to the sentence i e as a necessary constituent of it. Logicians, on the other hand, consider that a sentence need not necessarily consist of a verb In a linguistic analysis, the logicians' approach, based partially on ontological consideration and partly on the interpretation of nature and behaviour of 'asti' in a sentence, is not acceptable. III 4 However, one problem remains . how to account for the occurrence of 'asti (is) in 2 sentence with the optionits 32 33 34. Matılal, BK, (1966), pp 388 389 Rao, Y s,(1969), pp 13 19 Rao, P. S, (1969), p 7 f ·166 SANSKRIT SYNTAX ality of its non-occurrence in the present. Chomskyss gives base rules to introduce a copula. Thus, there may be subsequent rules to account for its non-occurrence in certain contests. Consider a sentence like 2 above which has the underlying structure as in figure V. Figure v S M 1 [+pres ] És as NP K NP 1 rāma bālaka The subjectivalisation-rule selects Es for agreemert with the V, T takes on the features of Es, and the result is as repiesented in figure VI. Figure VI S Es Σ 1 +-pres. tsing: +third ] K f rāma as -tsing: +-third +mas. NP } balahart-sing. +mas. balalar 35. Chomsky, N, (1965), p 107.

·NP-ASTI-VP CONSTRUCTIONS 167 The affir-substitution rule replaces features with affixes and the final form is as in figure VIII via figure VII Figure VII Figure VIII S S Es IM! Es tāmah T NP V K 1 rāma ast1 bālakah as NP K balaka Alternatively, at the stage of figure VI, a copula deletion rule optionally deletes the V, as the M shows [+-present] Consequently, the result is as represented in figure IX from figure VI Figure IX S Es NP K rāmi *+sing +third NP +mas balaba [Ignas] The affix substitution cule replaces the features with affixes, and the final form (omitting the intervening details) is as ·168 SANSKRIT SYNTAX represented in figure X Figure X S Es 1 rāmah bālalah Bach3, however, makes the proposal that have' and 'be'need not be introduced in the base-component, but may rather be inserted transformationally It may be mentioned that, in Sanskrit, there is no verb-form comparable to the 'have' of the English, as sentence 37 above indicates. For English, Bach gives additional evidence in favour of his position. His rules which insert 'have/be' transformationally, also account for reduced attributive clauses 37 like 'I saw the man on the roof'-'I saw the man who is on the roof'. But, in Sanskrit, a proposal for transformational insertion of 'asti' based on siniultaneous explication of reduced attributive clause cannot besustained. A sentence like 53 cannot be interpreted as a reduced attributive form of sentence 54 53 D nom R acc balalah dásam mārge apaśyat L 1 boy servant - on road - Saw (the boy saw the servant on the road) 36 37 Bach, E, (1967), pp 462-485 Bach, D, (1967), pp 463-466 ·169 NP ASTI NP CONSTRUCTIONS 54 balalah boy asit tam dāsam - apaśyat-yah — mārge - that servant — Saw #ho- on road was (the boy say that servant who was on the road) In other words, it is difficult to assign a node which would dominate only 'dāsam' and 'mārge' in sentence 53 mārge' in sentence 53 is L dominated by P as is any other case category in the sentence It may be argued that a sentence like 54 is ambiguous unless some word-order is assumed and may be interpreted as 54' 54'. bälakah yah mārge āsit - tam - dāsam - boy who - on road - was-him - servant apaśyat saw (the boy who was on the road, saw that servant) But, with either of interpretations (54/54 ) the indeterminacy of locational reference remains, and the other of the two(1 e. bālakah' and 'dasam') cannot be precluded from sharing the same location I suspect that even if a reduced form of either 54 or 54' is possible, it would not be treated as different from 53 above Though, a priori, neither of Bach's and Chomsky's proposals can be rejected, there seems to be no evidence elswhere within Sanskrit syntax which would justify the transformational insertion of fasti' as suggested by Bach Thus, it appears appropriate to treat ‘astı' like any other verb and introduce it in the base component Secondly, it is not always the case that 'asti' can be optionally deleted in case the node M shows [+present]. Recall sentence 4' and its derivation (through figures I IV), and sentence 2 ·170 SANSKRIT SYNTAX 1 1 and its derivation (through figures V-VIII) It has been mentioned that a rule deletes the V optionally (as in figure IX), as the node M shows [+present] (figure VI) But, this rule does not operate with the configuration in figure II as a sentence like 4" is unacceptable 4" * Tāmena châtrena Es T byr - by student Thus the V-deletion rule is blocked if the subjectivalisationrule has not applied (as in sentence 4 ) Thirdly, in the derivation of some constructions like 46, 47 (as against 45 above), fastı' must be introduced in the underlying structure, and it does not appear to be profitable to have a treatment for V in the derivation of sentences 3-7 above wnich is different from the one proposed for sentences 4,3", 5, 6" and 46, 47 above Even if it is argued that it is possible to specify the environment where 'asti' is obligatorily, as opposed to optionally, inserted, it is not clear how it would help, and would add to the compleuty of rules without any compensatory advantage elsewhere Considering all these arguments, there be no case for not giving a uniform treatment to astı' while there is justifiable reason to treat it as any other verb III 5 In this section, the status of asti in sentences which are subject to the process of mammalisation, is 2.- mined Consider sentences 55-58' It may be noted that sentence 55 does not allow the optional deletion of 'asti' despite the fact that the matrix S (in figurc XV), under thc node 1, shous [+present]. Sentence 55 is unacceptable Sentences 55 58° arc complcı as the process of nominalisation is rclated to embedding of S(IV.3 1) Tough this study, as mcationed in the beginning, docs not account appears to ·NP-ASTI-NP CONSTRUCTIONS 171 asti Es nom 1. 1 asti - acc -1S for complex structures, this departure is considered necessary to examine such sentences with 'asti' occurring therein. 55. ramah gacchan VNom nom going is (r. is going) 55'.* rāmah - gacchan -going 55'. tāmena gacchatā — bhūyate Es VNom 1 by r. - by going is become (it) is become going by 1.) 56. balakah - pustakam - pathitavān Es nom 0 V Nom nom boy book - having read (the boy has read the book) 56', bālakan pustakam pa țhitavān boy book - having read (the boy has read the book) 56". bālakena - pustaham — pațhitavatā — bhūyate Es VNom by boy book – having read - is become ((it) is become having read the book by the boy) 57. balahena pustakam pathitam А Es VNom nom by boy booh read 57', bālakena pustaham - pathitam by boy - book - read both=(the book is having been read by the boy) 57". balakena pustahena pathitena - bhūyate Es VNom 1 by boy by book by read - is become (it) is become having been read by the book by the boy) 1 2CC 1 asti 1 nom is 1 ·172 SANSKRIT SYNTAX nom 58. pustakam - pathanāya - asti Es VNom d book - to read is 58'. pustakam - pathanāya book to read (both=(the book is for reading) 58*. pustakena – pathanāya -- bhūyate Es Y Nom d by book - to read is become (it) is become by the book for reading) Sentences 55, 55' have the underlying structure as represented in figure XI. Figure X S. i M [+pres.] Es as NP rama M [+prog.] gacch rám2 ·173 NP-ASTI-NP CONSTRUCTIONS A nominalisation-rule applies and specifies V as VNom and attaches the feature [+adj] to VNom and the result is as represented in figure XI. Figure XI S. M [+pres.] V Es as к 1 rāma M [+prog.] VNom 1 gacch [+ad;] NP ramı The subjectivalisation-cule selects A (in S.) foc agreement with the VNom. The feature [+prog.) under Vis attached to the VNom and the Mis dcicred. The result is as represented in figurc XII.

·174
SA$RIT SY'TAN
Tigre XII
S,
NI
1
.[+pres.)
Es
as
NP
S,
táma
P
NP
V
Nom
ráma
gacch
[pos]
As the VNom shot's the featurc [+prog], the stom-clement
'at' is attached to‘gacch' and the featurc [+prog ) is deleted
The item under 'A' is deleted under identity The result
is as represented in figure XIII

·175 NP-ASTI.NP CONSTRUCTIONS Figure XIII S M [+pres.] V Es as K NP rāma Si P VNom gacch-at [+ad; ] A node-pruning rule 38 deletes the node S1 and a P-raising Iule attaches the P to P in the superordinate S. The result is as represented in figure XIV. Figure XIV S M [+pres ] Es V As NP Nom gacchat [+adj.] rama 39 Ross, J R, (1967), p 44, (1969), p 289.

·gacchat (sing NVom 176 SASARIT SY TAS Non, a rulc dclctcs P s hercicr it is dominated by itself (1c P) Thus, the nodc P(encircled) is deleted, and the result in as represented in figure IV. Tigrire XV S M [+prcs ) | 25 gacchat K [+ad; ) NP ram3 Now the subjectivalisation rule may apply optionally and sclect Es for agreement with the V It may be noted that as the VNom shows [+adj], it takes on the features of Es, whether Is is subjectivalised or not Es is selected for subjectivalisation and the result is as represented in figure IVI Figere AVI S 1 Es NI *+pres + sing, NP K +third tama VNom +-sing +third +mas as 39 An informal notation for represerung the specification of the verbal features Some of these will have been introduced by ·177 NP-ASTI-NT CONSTRUCTIONS The affis-substitution rulc replaces the features with affixes and the result is as represented in figure XVII. The final form is as given in figure XVIII. Figure XVII S M - ti VNom NP } rāma as h gacchat Figure XVII S Es tamah VNom - asti gacchan It may be noted that the copula-deletion rule cannot apply at the stage of figure XV (cf. figure IX) as the VNom shows the stem-element 'at'. phrase-structure rule which rewrites a V as CS (eg +1) ard some would have been introduced transformationally (eg the 'at' in the example is a reflex of transformatior ally introduced feature [ + prog. l. 12 ·178 SANSKRIT SYNTAX Alternatively, at the stage of figurc XIV, Es is not selected for subjectivalisation and the process of passivisation without subjcctivalisation (11. 1. 1.) opcrtcs. The final form is as represented in figurc XX (from figure XIV) via figure XIX (omitting intervening details). Figure X7X S NI P te V VNom bhū-y2 -=- NP gacchat K 1 ina rāma Figure 23 S Es V Nom bhūyate tāmena gacchatā Consider, nest, sentences 56, 56', 56'' which have the underlying structure as represented in figure XXI.

·179 NP-ASTI-N CONSTRUCTIONS Figure XXI S. — N [+pres.] Es - as NP NP bälaha S1 M P [+pief.] V path NP K NP bālaka pustaka The nominalisation-rule applies and specifies Vas 'VNom'and attaches [+ad; ] to VNom, the subjectivalisation-rule selects A for agreement with VNom, the M-deletion rule attaches [+perf.] to VNom and deletes the node M As the VNom shows [+perf.], stem-element 'tavat' is added to 'path' and [+perf] is deleted. 'bālakah' under A is deleted under identity (cf. figures XI, XII, XIII) The result is as represented in figure XXII (omitting intervening details) ·180 SANSKRIT SYNTAX Figure XXII S. MI Р [+pres.] Es 2S NP balaka --- VNom - pathitavat (+adj.] NP к pustaka Thc nodc-pruning tule, the P-raising rule and the P-deletion rulc apply (cf. figures XIV, XV), and the result is as represcntcnd in figure XXIII (omitting intervening details). Figure XXIII S M [+pres.) Es VWom 1 25 WT מא 7 pathiravat (+ adj.] pusiala bilala ·NP-ASTI NP CONSTRUCTIONS 181 The subjectivalisation-rule selects Es for agreement with the As mentioned earlier (figure XV), VNom takes on the features of Es, as it shows [+ad;], whether Es is subjectsvalised or not Secondly, it may be noted that even if Es is not selected for agreement with the V, O cannot be subjectivalised (1 e the process of passivisation (II 1 1) cannot operate) Because, in case the subjectivalisation rule has operated (in embedded S) before P-raising, no case category which originated therein can be subject to the subjectivalisation rule or the process of passivisation with subjectivalisation after P-raising The affix-substitution rule replaces features with affixes and the final form (omitting intervening details) 1s is as represented in figure XXIV (cf figures XVI, XVII). Figure XXIV S 1 Es balaken bālakah V VNom V Nom i asti pathitavan pustaham The copula-deletion rule may apply optionally, after the subjectivalisation rule has applied (cf figure VI), and thus sentences 56, 56 are derived Alternatively, at the stage of figure XXIII, subjectivalisation does not apply and the process of passivisation without subjectivalisation operates The affix-substitution rule replaces features with affixes and the result is as represented in figure XXV (omitting intervening details) The final form is as given in figure XXVI from woich sentence 56 ·162 SANSKRIT SINTAX is obtained. [It may be noted that as the subjectivalisationrule has not applied, the copula-deletion cannot apply.} Figure XXV S M - tc Es yNom bhū-ya NP K pathitarat , NP K 1 bālaka ina pustaha an Figure XXI Es VNom bhūyatc bill cha prthitavati 1 pustalam Consider, ncst, «cricnces 57, 57', 57' vleich 121 the underlying structure as represented in figure XX1T.

·183 NP-ASTI-NP CONSTRUCTIONS Figure XXVII S. M [+pres ] V Es 1 as NP K pustaka [+perf] path NP K NPK 1 pustaka balaka The configuration in figure XXVII is similar to that in figure XXI except that Es dominates 'pustaka' and not 'balala' (as in figure XXI) The nominalisation-rule applies and attaches [+ad; ] to VNom in the embedded S, the subjectivalisation-cule selects for agreement with the VNom, the Al-deletion rule attaches [+perf} to VNom in the embedded S and deletes the node M As the VNom shows [+perf], the stem clement 'ta' 15 added to 'path' in the embedded S and [+pecf] is deleted 'pustaka' in the embedded S is deleted under identity The result is as represented in figure XXVIII (omitting intervening details).

·184 SANSKRIT SYNTAX Figure XXVIII S. M P [+-pres.] V Es as NP pustaka VNom pathita [+adj.] NP bālaka The node-pruning rulc, the P-raising rule and the P-deletion rule apply (cf. figures XIV, XV), and the cesult (omitting intervening details) is as represented in figure XXIX. Figure XXIX M [+pres.] Es VNom as NP pashita K [Hadj.] NP pustaka balala ·NP-ASTI-NP CONSTRUCTIONS 185 *The subjectivalisation rule may or may not select Es for agreement with the V. But, as mentioned earlier (figure XXIV), A cannot be subjectivalised. Es is selected for subjectivalisation, the affix-substitution rule replaces features with affixes. The final form is as represented in figure XXX (omitting intervening details). Figure XXX S Es pustakam VNom bālakena asti pathitam The copula deletion may apply optionally after the subjectivalisation-rule has applied, and thus sentences 57, 57" are derived. Alternatively, if at the stage of figure XXX Is is not selected for subjectivalisation, the process passi visation without subjectivalisation operates, and through a derivation similar to that in figures XXV, XXVI, sentence 57' is derived

·186
SANSKRIT SYNTAS
Consider, next, sentences 58, 58', 58' which have
the underlying structure as in figure XXXI.
Figure XXX
S.
M
[+pres.]
ES
-
as
NP
K
i
pustaka
M
[+prog.]
path
NP
pustaka
The nominalisation-rule specifies V as VXom and the
subjectivalisation-rulc, being optional, docs not apply.
Whencicr subjec:ivalisation does not operatc, the M node
is deleted. Thic stem-clement 'ana' is added to 'path' in the
cmbedded S. The result (omitting inter ening details) 18
as represented in figure XXXII.

·187 NP ASTI NP CONSTRUCTIONS Figure XXXII S. [+pres ] Es --- as NP S, pustaka VNom А 1 pathana - adj] NP pustaka *pustaha'in S is deleted under identity, the dummy A is deleted, the node pruning rulc deletes the node Si, the P raising rule and the P deletion rule apply, and the result is as in figure XXXIII (omitting intervening details) Figure XXXIII S. P NI [+pres ] Es VNom 1 25 pathana -adj] K 1 pustaha ·188 SANSKRIT SYNTAX The subjectivalisation-cule selects Esforagreement with theV. It may be noted that the VNom shows (-adj. ] and, hence it cannot take on the features of any case-category. The result is as in figure XXXIV. Figure XXXIV S Es M NP *+pres. +sing: t-third pustaka VNom +sing +third -mas as fem pathana rt-sing. mas. --fem The affix sub,titution rule replaces the featurcs with affixcs. [A morphophonemic rule attaches dative) to 'pathana under VNom] The result is as represented in figure XXXV. Figure XXXV s-IMIH Es t1 K VNom | pustaka am pathana 12 ·NP-ASTI- NP CONSTRUCTIONS 189 The final form is as in figure XXXVI. Figure XXXVI S Es pusta VNom 1 - asti pathanāya The copula-deletion culc delete the V after the subjectivalisation culc has applied. Thus sentences 58, 58' arc dcrived. may Alternatively, if at the stage of figurc XXXIII, Es is not subjectivalısed, the process of passivisation without subjectivalısation operates, and the final form is as represented in figure XXXVII (omitting irreles ant details). I'igure XXXVII S Es VNom bhūyatc pustakona pathanaya Trius, sentence 58' is derived from figurc XXXVII. III 6 Thc rules that operate in the deriration of the abovc scntcnces are as follow : 1. Nomunalisation rulc specifies t'as VNom and attaches [ + adj ) to the itcm under Nom ·190 SANSKRIT SYNTAS 2 A Subjectival sation-rule applies optionally and selects a case-category for agreement with V/VNom However, on the last cycle, the subjectivalisation-rule cannot select any case-category which has originated in an embedded sentence But, VNom may be subjectivalısed in case the subjectivalisation-rule has not applied to the configuration wherein VNom originated. 3 A M-deletion rule operates as follow : (1) if rule 2 has applied, the feature under M is taken on by the VNom and the node M is deleted; (11) if rule 2 does not operate, the feature under M is not taken on by VNom and the node M is deleted. 4 A Copula deletion tule applies on the condition that subjectivalisation has taken place, and deletes optionally, (1) the copula under V if M shows [+pres ] and if there is a VNom also, it does not show the stemelement 'at'; (11) the copula under VNom if it shows [+prog.). 5. A P-raising rule attaches P in the embedded S, to the P in the superordinate S 40 6 A P-deletion rule deletes the node P if it is dominated by itself (1 e P) 10 7 An Affix-substitution rule replaces the features with affixes. 40 These rules mas be dispersed with, in case thc «cope of the rode pruning ruleis caterded to delctethe case-catcgory com 11- ting S ard pas rell (CF Kiparsly, P&J I Stual, (1969), P101 (figure 13)) ·IV GENITIVE CONSTRUCTIONS IV 1 This chapter deals with 'Genitive' which refers to a surface-phenomenon rather than to a case-category in the Fillmorian sense As explained later in this chapter, the Genitive is derived from so many diverse sources that it is considered appropriate to discuss them together with reference to their surface forms By 'Genitive' is meant an NP with the marker 'sya' (1 e genitive') and its vagiants Thus, the Genitives i e NP's underlined in sentences 1-4 indicate the main range of different realisations of 'sya' (1 e genitive) 1. balaka-sya — pustaham G book (boy's booh) [the NP'balaka'15[+masculine]and ends in a lonel ] 2 balika gab- sundaratā G 8 - beauty (girl's beauty) [the NP'bālıha's[+feminine) and crds in a rotel] of boy of girl 1 2 Genitive) refers to corstructions with síos tic Irfle- cn The marker'sra is a coser term fo its variant forms, occurrcrcc of which depends on whether the MP has the ſcatuc (-masi 1[+fom '--mas yard whether it crds ir a vokcl or consonant. Thererms on cland consonant atcLsed as sefernire to pronological units and as defired in Sancit texte ·192 SANSKRIT SYNTAX 3. madhu-nah madhuratā G of honey sweetness (honey's sweetness) [the NP 'madhu'istmas. and ends in a vowel.] 4. vidyarthin-ah agamanam of student coming [the NP'vidyārthin' is [-+-masculine] and ends in a consonant.] IV. 2. The later Sanskrit commentators did not assign a case-category to Genitive; though they included it in the section dealing with the 'kāraka's. Panini's governing rule “sasthi ſese' (sixth (case-affix i. e. genitive) in the rest)4, has led to controversy among the Sanskrit commentators and grammatians with regard to its interpretation. In the first instance, they have tried to show that 'saşthi' indicates a ‘kāraka’-relation as in ‘rājõah - puruṣaḥ (king's man)s. Then they try to interpret the term 'seșe's also as a 'kāraka. So Bhattojidiksıta (1700 A. D.). Panini, II, 3. 50. Iyer, K. A. S., (1969) pp. 324-325; : 'the relation of master and servant in the expression ‘sājñaḥ - purusaḥ :thching's man, implics a previous rclation of action ard accessory or rather of donor and recipient between the two ( It is difficult to accept this assignment of donor-recipient' property to NP's 'raynah'. and purusah' as instrumental in establishing a läraha - relaison. In a sentence lihc 'sājā - puruşam - ādıšatı" (king-to manorders), the same relation of master-scrvant (or master-recipies :) obtains, but the käraka-relation is of 'Lastā-karma' (in relation to the verb adıšalıbsk) ard that is how the relation of master and scruant also becomes a kırd of 'häraha' indirectly. The prehous rclation of master and servant is the cause ard the later relation of master and sccipient is the sesult. Then the pics fous 3. 4. 5.

·194 SANSKRIT SYNTAX and interpretation of the Genitive.? In so far as the transformational model is concerned, it may be safely claimed that the Genitive in Sanskrit has not been discussed in any detail. Kiparsky and Staal have discussed the 'Lācaka's and mentioned the Genitive in the list. But they also have left it without any discussion.8 IV.3. The Genitive in Sanskrit (1.e. NP with a mather 'sya', f. n. 1) can be related to four types of constructions as in sentences 5-8', 9, 9'. 7. Vámana-Jāyāditya, vol. II, pp. 209-11 : 'harmādıbhyah -anyah prātipadılarthavyatırıktan svas vāmibhāvādisambardhah - śeşaḥ - tatra - sasthi vibhaktih-bharati'... whatever is cot covered by the rules governing Läraha relations ard the rule II 3 46, is sesa and takes genitive. Cf. Patañjali - (cd ) Kielhora, F., (1892) vol. I. p. 463. Cf. Pānını - (ed) Vasu, SC, (1891) vol I, p 296 Cf. Spesjer, J. S, (1886), P. 82 (f n 1). "Las on I. 1. 49 'babarah - hi - şasthyarthäh - syasrāmyanantaracamīpacamūharıkāsāpayavādsäh' Panini seems to have not sharply defined the genitue's sphere of employment, at least if tre explain his Satra (rule) II. 3 50 'sasthi- Sese' with thc 'hāśikā,' as meaning in all other sostances (namely If none of the other cases taught (II 3. 1-49) be available), onc should use the sixth case'. But then it is strange, Tänını has not said intersely 'ese-şasthi' (cf, his constant use I 47, 1.4. 108, II. 2. 23, III.3 151, Vn. 2. 90). Now Patañjalı gires a somewhat different explication 'harmadicím - arimhşā - Segah • 'she such WC is required of she curgeries aburi and the researcnot to be distinctly expressed'but tacitis Implied Cf. Bhastojidikşita (ed.), Roy, S (1920), p. 70-71 'Sesa" is 'other than that is said', 'what temains to discuss' etc. The läraka's and the sense of a 'prātipadila' hare been discussed. Hence such relations as that of the possessor and the potressed fet remain to speal. of. These relations will be expressed by the the 'sasthi bhakti', Kiparsl 5, P. ard J. r. Staal, (1969) PP 83 117 .. . 8.

·GENITIVE CONSTRUCTIONS 195 5. sa - Racc 5. Sa he = acc In sentences 5-8' the Genitive is a mere alternative surface-realisation of a case-category normally realised by different morphological forms, as compared with sentences 5-8. priyām —- smarati D he beloved remembers - priyayah - smaratı D R of beloved remembers both = (he remembers his beloved) 6 chātrah guroh caranau - bhajate A G g O student - of teacher -- feet worships 6. chātrah - guroh - caranayoh - bhajate А G g student - of teacher - of feet - worships both=(the student worships his teacher's feet) 7. bālahah - phalaih trpyati D I boy with fruits - feels contentment 7. bālakah - phalanām - trpyatı I 8 boy of fruits feels contentment both=(the boy feels contentment of fruits) 8. dyūtaharah - śatam pratidivyati А O acc gambler hundred gambles 8. dyūtakarah śatasya pratidivyatı А 0 gambler of hundred gambles both= (the gambler gambles hundred (pieces of money) 1 ·196 SANSKRIT SYNTAS (11) In sentence 9, the genitive is derived through the process of nominalisation' from an underlying case-category, thus 9 is transformationally related to sentence 9'. 9. lelhasya - racahah - guruh - āsīt F g VNom of article - Wilter - teacher - was (the teacher was the writer of the article) 9'. lekham - aracayat - guruh F acc A article wrote - teacher (the teacher wiote the article) (111-10) In sentences 10–11, the Genitite is related to what Fillmore would call alienable and inalienable possession. Following Fillmore, the Genitis c in sentence 11 may be developed according to the schema NP> N(S)11. But, in sentence 10, the Genitive wculd not be developed in this way. Discussion of sentences like 10, 11 is taken up later (IV. 3 2) 10 tasyah - vadanam - sundaram - asti G & E т her face beautiful - 15 (her face is beautiful) Nominalisation as a process is related to senterce-embedding (excluding relative structures), the embedded S being dominated by a casc-category The process of nominalisation in regard to 'NP-asti NP' sentences has been analysed in II 5 beargued that seriences like I harca claim to beincluded in the samc chapter (1 e III) But, as they contain Gertries as well, they are better taken up in this chapter I have not assigned case-category to 'racalah' (usiter) in sentence 9 posed that Vom necd not be assigned & case category in NP- 1503 NP sentences or clsewhere Tillmorc, C. ), (1568a), pp 61-81 Cf Chomsky, A, (197(a), nom nom 9 It may It is pro- 10 PP 200ff 11 Tillmore, C), (15642), p 66 ·GENITIVE CONSTRUCTIONS 197 11 tasyah - bhavanam - sundaram - asti G g E nom T nom her house beautiful - 15 (her house is beautiful) Constructions of the type exemplified in sentences 5 8 will not be discussed in this section These relate to surface inflexional realisations of the same case category and can be handled by appropriate affix assigning rules The remaining of the above mentioned constructions(1 c 9-11) will be discussed below IV 31 As mentioned above, sentences like I are lelated to the process of nominalisation, and the Genitne in such constructions is derived from an underlying casecategory transformationally. There have been arguments,12 following Chomshy,13 that as the transformational derivation of nominals does not operate systematically in the case of dorned and gerundne nominals, the Gentsc in both constructions cannot beireated in the same way In other words, the Genitie of gerunduve nominals needs to bcdern edina different way from a]] cthers But in Sanskrit (f n 13), the process of ncminalıcation can be transformationally related to underlying structures in a systematic Asentence containing a rcmunal would be derned from 1 deep structure involving specific casc categories and manner 12 13 Stocha ell R P et al (1968), pp 530 536, 712 Chomsky A (1970a) pp 1E4 221 Chomsky malos a distin cuon between 150 types of romiral sation gorurdisc nominal and der cd rominale & J sscfusirgitc oferss Js refusal of the offer) which d fler in respect of (!) productivity of the process in question (11) thc gercral ir of the selation. (111) ard 'theirtar al stilcture of the rom sal sl tase' The dicho inmy is rot of much corsequerce in 10 far as ro raluat on in Sirskrue is concerned ·198 SANSKRIT SYNTAX the Genitive in the former would derire ultimately from a case on the verb in the underlying structure. Thus, a good number of Genitives (in relation to nominals which are morphologically related to verbs in underlying structures) can be traced back to cases. In sentence l'above, the Genitive 'lekhasya' (of article) is the alternative realisation of the case-category F (in sentence 9), conditioned by the VNom fracakah (writer). IV. 3. 1. 1. In this section, the constructions exemplified in sentence 9(or in other words structures containing Genitives and related to the process of nominalisation) are described through their derivation. Sentences 12, 13 are described in detail. Sentences 14, 15 are described with respect to those details which differ from that of 12, 13. Sentences 12-15' ace alternative realisations due to the non-operation of the subjectivalisation-rule which has been referred to earlier, and hence need not be taken up again. 12. anudāradalasya netā harimohanah - yäti G 8 VNom A of Conservative party - leader - h. goes (h., the leader of the Conservative party, goes) 12'. anudāradalasya - netrā harimohanena yāyate G g VNomi A of Conservative party - by leader - by h. - is gone (it) is gonc by h. the leader of the Conservative party) 13. prabhuḥ - sevakasya sevakena - aśvasya - nagaraya/ A nom G i G d of servant/by servant - of horse - for city nagaram - nayanam - adićati GO acc VNom to city - leading - orders nom i 8 GO master acc ·GENITIVE CONSTRUCTIONS 199 1 g GO to city - (the master orders the leading of the horse by the servant to the city) 13'. prabhunā - seyakasya/sevakena - aśvasya - nagarāya/ А G 8 1 G 8 by master - of servant/by secvant of -horse-for aty/ nagaram nayanam ādiśyate GO acc VNom nom leading 1$ ordered (the leading of the horse to the city by the servant is ordered by the master) 14 pathikasya - jalasya - pipāsa avardhata G g VNom nom of traveller - of water - desire to drink - increased (the desire of the traveller to drink water increased) 14' pathikasya - jalasya - pipāsayā - avardhyata G & VNom 1 of traveller - of water - by desire to drink - was increased ((1t) was increased by the desire of the traveller to drink water) 15 balahasya - rogasya / togena rujā Vardhate 8 VNom nom of boy of disease/by disease-torment - increases (the torment of the boy by the disease increases) 15'. bālahasya - rogasya/rogena - rujaya - vardhyate G VNom 1 of boy - - of disease/by disease -by torment - is increased ((it) is increased by the torment of the boy by the disease) 1 1

·€200
SANSKRIT SYNTAX
Sentences 12, 12' have the underlying structure as in figure I.
Figure 1 -
1
is
S.
M
[+pres.)
ya
NP
S,
harimohan
MI
1
[+prog.) |
V
ni
NP KP
K NP
K
harimohana
anudāradala
Thc nominalisation-rulc specifics V as VNom and attaches
the feature [ +-adj.] to VNomin S. A in S, is subjectivalised.
The featurc [+prog.] is attached to VNom and the noded
is deleted. The result is as represented in figure II (omitting
irrelevant details.)

·GINITIVE CONSTRUCTIONS
201
Figure II
1
1
<
So
M
[+pres ]
yā
NP
harimohana
S,
VNom
-
ni
ܠܨ܂
L#prog]
NP
K
NP
K
harımohana
anudāradala
As the VNom in S, shows [+prog] the stem element 'tr'
is added to the item under the VNom and the feature[+prog ]
is deleted. The item under Ain S, is deleted under denuty
with the superordinate NP
The reult is as represcricd
in figure III (omitting irrelevant details)

·202 SANSKRIT SYNTAX Figure III So M — [+pres.] y ya K harimohana VNom nitr [+adj.) NP anudāradala The node-pruning rule, the P-raising rule and the P-deletion rule apply, and the result (omitting intervening details) is as represented in figure IV. Figure IV S NI [+pres.] V VNom ya NP Knct? NP [+adj.] 1 harimohana anudiradala ·GENITIVE CONSTRUCTIONS 203 The subjectivalisation-rule may or may not select 'A' for agreement with the V, but the VNom will take on the features as it shows [+adj ). Secondly, even if A is not selected for subjectivalisation, the process of passivisation cannot operate and cannot be selected for agreement with the V (sce figure XXIV, III. 5). A is selected for subjectiyalısation and the result is as eepresented in figure V. Figure v S M V Nom | netr +pres. +sing NP K +third] v harimohanas+sing: yā +third +-mas NP K +-sing. L+mas. [-sing anudāradala-mas. L-fem The affix-substitution rule replaces the features with affixes and the result is as in figure VI (a morphophonemic rule attaches genitive) to obligatorily). Figure Vi S $.:] P IMla ti NP Nom →KIA harimohana ya netr h NP K anudāradala sya ·204 SANSKRIT SYNTAX The final form is as represented in figure VII. Figure VII S A harimohanah V Nom yati netā anudāradalasya Thus, sentence 12 is derived from figure VII. Alternatively, at the stage of figure IV, if A is not selected for subjectivalisation, the process of passivisation without subjectivalisation operates (see II. 1. 1.) and sentence 12 is derived. Consider, next, sentences 13, 13' which have the underlying structure as in figure VIII. Figure VIII S. M [-+-pres.] adıś NP S, prabhu Σ [+prog.) R GO ni NP K NP K NP K sevaka aśva nagara ·GENITIVI CONSTRUCTIONS 205 Thic Nominalisation-rulc specifics V as Vom and attaches the featurc (-adj ) to Vom (in S.), the subjectiv alısation-culc docs not opcratc, the M deletion rule deletes the node 1, thc stem clement 'ana' is added to the item under Vaom Thc result is as represented in figure IX (omitting intersching details) Figure IX S. M 1 [+pres ] ādis NP prabhu V Nom А R GO - ni ana 1 [mad; ] NP NP K NP K sevaha asia nagara The node pruning rule deletes the nodeS, the P-taising rule and the P-deletion rule apply, and the result is as represented in figure X omitting intervening details) Figure X So M 1 [+pres ] А VNom R 1 K NP adıś NP prabhu nayana K [-adj ) NP sevaka 1 aśya K NPK 1 nagara ·206 SANSKRIT SYYTAS The subjectivalisation rule may select A (of the matris S) for agreement with the V Recall the subjectivalisation cule explicated earlier (figure XXIV, IL 5), which bars the selection of any case category which has originated in the embedded S, er en if A (in the matrix S) is not selected for agreement with the V That statement holds good here as well However, there is one point to be noted The subjectivalisation rule has not applied to S, (see figure IX above) and no case category has been selected for agreement with the VNom (in S.) Under these circumstances, the VNom behaves like an O, and may be selected for the process of passivisation in case A is not selected for subjectivalisation A is selected for subjectivalisation, the affix substitution rule replaces features with affixes and the final form is as represented in figure XI (It may be noted that morphophonemic rules attach g(enitive) obligatorily to R(asva'), optionally to A(*sevaha' (ostrumental) is ched alternatively), and d/acc to GO (ʻnagara')) Figure X7 atta- S prabhuh Vnom - R GO 1 adišati nayanam sevahasyal aśn asya nagaram/ sevahena nagaraja Thus sentence 13 is derived from figure XI Alternatively, if A is not selected for agreement with the V, at the stage of figure X, the process of passasi sation, operates and the VNom ('nayana') is selected for ·GENITIVE CONSTRUCTIONS 207 agreement with thc V. The result is as represented in figurc XII (from figure X). Figure XII S Nom M na yana sing I+third -mas. -fcm. +pres. +sing: -+-third âdıs-ya NP K NP K 1 prabhu sevaka [+sing L+mas. [+sing. +mas. GO R NP K NP K aśva r+sing nagara *+-sing [+mas. -mas. -fem. The affix-substitution rule replaces the features with affixes (morphophonemic rules attach 'g/l' to Aſof the embedded S), 'g' to R, dlacc' to GO The result is as represented in figure XIII.

·208
SANSKRIT SINTAX
Figure XIII
Vrom NI
te
nayana
am
R
ādisya
NPK NP
K
NP K NP
K
prabhu ina sevaha syaſına asva sya nagara yalam
The final form is as in figure XIV
Figure XIV
S
-
งsort
nayanam
do
GO
1
adiśjate prabhuna serahasyal asvasya nagarājal
sevahena
nagaram
Thus, sentence 13 is dern cd from figure WV
Consider, neat, seniences 14, 14 and 15, 15 which
have the underlying structures as represented in figures \V
and VI respectively

·GLNITIVT. CONSTRUCTIONS 209. Figure XV [+past] vardh M [+prog.) v Pipas NP K NP pathika jala Figure XVI ၄၀ NE 1 [+pres.] vardh NI I [-+prog.] R ruj NP NP K roga balaka 14 ·210 SANSKRIT SYNTAX The nominalisation-rule specifies V(16 the embedded S) as VNom and attaches (-adj ) to VNom , subjectivalisation does not operate, the node M is deleted, the stem element ā'is attached to the items under V Nom of the embedded Sin figures XV and XVI The result is as represented in figures XVII and XVIII respectively (omitting intervening details). Figure XVII S. M [+past] V vardh Nom pipās-ā (m-adj ) NP K NP pathiha jala Figure XVIII So NI [+pres ] vardh 00 fuj- [-adj] NP K NP balara roga ·GENITIVE CONSTRUCTIONS 211 Now, the node-pruning rule deletes the nodes S. (in figures XVII, XVIII), the P-raising rule and the P-deletion rule apply, and the result is as in figures XIX and XX respectively. Figure XIX S N 1 [+past ] Vom vardh pipāsa (-adi ] NP K NP K I 1 pathiha jala Figme XX S M [+pres ] V Nam vardh rujá [-adj] NP K NP K roga bālaha As mentioned above (figure XI, above), no case-category which has originated in the embedded S can be selected for subjectivalisation after P-raising. But, no subjectivalısa tion-rule has applied to the embedded Sand no case category therein has been selected for agreement with the Vom Under these circumstances, the Vom behaves like O and can be subject to subjcctitalisation and the process of pascitisation.

·212 SANSKRIT SYNTAX Vom 15 selected for agreement with the V, the affis-substitution rule replaces the features with affixes (morpho phonemic rules attach 'g' to Dard O in figure XXI, and gli' to I and 'g' to R in figure XXII) The result is as represented in figures XXI, XXII (omitting intervening details) Figure XXI Nom M a ta pipāsā vaidh NP K NP K pathiha sya jala sya Figure XX1 NI Nom 1 te ruja V vardh K NP K roga sy/ina balala sya ·GENITIVE CONSTRUCTIONS 213 The final forms are as represented in figures XXIII, XXIV, from wrich sentences 14 and 15 are derived respectively. Figure XXIII S Nom pipāsā D avardhata pathikasya jalasya Figure XXIV S Nom rujā V R vardhate rogasya/rogena bālakasya Alternatively, if at the stage of figures XIX, XX, the VNom's ('pipāsā' and 'rujā”) are not selected for subjectivalisation, the process of passivisation without subjectivalisation operates and sentences 14 and 15' are ved from figures XXV and XXVI respectively (from figures XIX, XX) (omitting intervening details) Figure XXI S D Νοπι avardhyata pipāsaya pathihasya jalasya ·214 SANSKRIT SYNTAX Figuri XXVI S Vom R 1 of girl vardhyate rujaya rogasya/rogena balahasya IV 32 In this section, constructions lil e 10, 11 above (IV 3) and 16, 19 below in which a Genitive (f n 1) is related to an NP, are discussed Comparable constructions in other languages, for example in Hind114, have been referred to as 'possessive' I would refer to them as Genitives In sentences 10 and 16, the relation between the Genitive and the related NP is finalienable and in sentences 11, 17 it is alienable' The inalienability and alienability of relation is discussed later (IV 3 2 1) 16 balayah - pita yati G father goes (girls's father goes) 17 balayah - pustaham - asti G 8 Es of girl - book (the girl has a bool/there is girl's booh)IS Various scholars have referred to the structural parallelism between locatre and possessive constructions 16 14 hachru Y, (1969) pp 37-45 TI s ambiguous interpretation of sentences like 11 17 13 dis cussed in IV 3 2 4 Lsons ) (1969) pp 391 334 344 345 Constructiors lilc is beck is the book on ile icblc this parallclism appears to be deperdent on the assumption that the urderl ned scemeris arc optional or structurally dispensable constituents of the Cf id (1967) pp 300 6 -15 15 16 SC. toroc ·215 GENITIVE CONSTRUCTIONS But this statement does not take into consideration sentences like 10,16 above. The main motivation for this parallelism appears to be the fact that 'x's' and 'on the table’16 in sentences such as this is x's book : the book is on the table' are not essential constituents. But, this is not true with all Genitive constructions. In sentences like 10, 16, ‘vadanam' (face) being a body-part term and “pitā' (father) being a relational term, imply a Genitive like 'bālāyah' which is not a non-essential constituent. This is true even in sentences like 'vadanam - sundaram - astı' (facebeautiful-Is) and 'pitā - yāti' (father goes), as the interpretation is always in a context. Lyons does not discuss constructions like 10, 16 above.17 In the case of constructions like 11, 17 above Lyons is in favour of treating 'have' as a surface realisation of the underlying possessive, thus deriving a sentence like ox has a book'-'the book is x's', on the basis of characteristics in languages like Slavonic and Celtic which have no verbal form comparable to 'have' in English However, the occur. rence of a verb like 'have' in a language like English, makes explicit the distinction between sentences 10, 16 and 11, 17 above 18 It will be proposed that sentences like 11, 17 have a sentential derivation i e have a derivation involving an embedded S(IV. 3 2 4), but that sentences like 10, 16 do not. nom Consider sentences 18-20 besides 10, 16 above 18. tasyāh - vadanam - sundaram - asti g O her - face - beautiful - 15 (her face is beautiful) 17 18. Lyons, J. (1969), pp 388 398 Fillmore C ), (1968a), p 66 ·216 SANSKRIT SYNTAX 4 nom nom 1 19. bhavanasya - praveśadvāram - apāvſtam - asti G of building - entrance-door - opened - is (the entrance-door of the building is open) 20. bhāratasya - usnahālah - duḥkhah - asti G 8 of India - hot weather, - unpleasant - is (India's hot weather is unpleasant) The relationship between the Genitive and the asscciated NP (in sentences 16, 18, 19, 20) 'is 'inalienable'. The relation refers to kinship (in sentence 16), to part-whole (in sentences 19 : in 18 it is body-part and in 19 it iclates to inanimate entities), to weather or meteorological phercmeron (in sentence 20). Though, it is not certain that the relationship between 'uşnakalah' and 'bhāratasya' in sentence 20, is inalienable semantically, the relationship in 18, 19 is semantically inalienable. In so far as the syntactic justification is concerned, sentences 18', 19', 20' with the occurrence of pārśve' are not acceptable (cf. IV. 3. 2. 1.). 18'. * tasyāḥ pārśve - vadanam - sundaram - asti her - face - beautiful - is 19 * bhavanasya párśve - praveśadväram - apāvſtam-, of building - entrance dcor - opened - asti 20'. * bhāratasya pārśre - usnakālaḥ - duhbbah - asti' of India - hot weather - unpleasant - is It may be noted that even in a locative sease the CCCurier.ce of(pārsvc, in 18', 19' 20' is not acceptable. Thus, the sentences 16, 18-20 are marked for finalienable possession' and sentence 17 for 'alienable possession',1' and henceforth will be referred to as such. 19. Tillmorc, C. J.. (156€ a) p. 66. Cf. Chomsky, N., (197(a), pp. 200-07.

·GENITIVE CONSTRUCTIONS 217 IV 3 2 1 Fillmore proposes two schemata for the derivation of constructions with inalienable and alienable possession Thus a sentence expressing 'alienable possession' lihc 'my house' in English would be derned from I have a house', in which the Genitive and the possessed NP come from an embedded S, the schema for the NP dominating the Genitive being NP >N(S) A Sentence expressing inalienable possession like 'my father' would be derived from an NP dominating a D, the schem? being NP-> N(D) 20 In Sanskrit also, this distinction between sentential and non sentential derivation of a construction with a Genitive is crucial syntactically and semantically, and needs to be re tained However, in the absence of a verb like 'have' there must be some other device to bring out this distinction 21 IV 3 2 1 1 Consider first sentences with Genitives which express 'inalienable possession' for which there is a schema lile NP-> N(D) Thus a sentence lihe 21 is distinguished from sentence 22 with alienable possession wherein the occurrence of the PP pārsve, may be accepted as a dis- 20 21 Fillmore CJ, (1968a) p 66 the schema as proposed by Fillmore, is under the sub title adnominal datives for him an adnominal datiye is a possessise modifier (p 66) of a noun In IV 3 2 1 1 5 a schema lihe NP - N(O) also is proposed as it is not the case that the relationship is alırays to an animate entity Fillmore C ) (1968a) p 66 A distinct method is required for introducing the possessive elemert in the case of inalienable possession a method which reflects the fact that the relationship between the io nouns in finalienable posse sion' is not a sen ten sal relationship In Sarshi t the occurrence of parśte brings out this dist nction beiycen alierable ard inalier able to lationship. In the ca e of a Genuise with the co occurrence of parśve indicating the alienable relationship a sentential deriva tion is proposed (IV 3 2 4 ) ·218 SANSKRIT SYNTAX tinguisher of alienability. It may be noted that a sentence lihe 21 cannot have an expansion like 21'. 21 * mitrasya - bhagini yāti G g A nom of friend - sister goes (friend's sister goes) 22 mitrasya pārśve - yat - pustaham - asti - tat - navam G 8 of friend - which-book that -new- 1S - asti 1S IS she - goes 1 (that book which is of friend, is new) 21 * mitrasya pārśve - ya - bhagını - asti - sā - yāti G g of friend who - sister In sentences 21, 22, 'parsve'is not interpreted as a locational reference It may be noted that a sentence lihe 23 contairing ‘parsve' with a locational reference (1 e near) cannot give a sentence lihe 24, 1 e no optional deletion of 'parsve' (in its locational sense) is allowed 23 (a) mitrasya - pārsve - yat - pustaham - asti - tat .. G of friend - near - which - book is that. (that book which is near the friend. ) (b) mitrasya - parśve - matā - asti G g of friend - near mother - 15 (mother is near the friend) 24 * (a) mitrasya yat pustaham - asti - of friend which book that (b) mitrasya mātā of friend mother tat - - IS asti 1S

·GENITIVE CONSTRUCTIONS
219
If the locational reference is not implied, sentences like
24(a) can be derived from 23(a). Horres er, it may be recalled
that, without a locational reference of 'pārśre' sentence
23(b) is unacceptable and as such no question of derlying
24(b) would arise.
In other words, sentences 24(a), 24(b) with the
optional deletion of pārśve' in locational sense) cannot be
derived from 23(a) and 23(b). Sentences 24(a) and 24(b)
ace, otherwise, acceptable.
IV.3.2.1.1.1. Consider sentence 21 above which bas the
under Structure as in figure XXIV.
Figure XXIV
S
N
[-+-pres.]
ya
NP
K
D
bhagioi
NP
)
mitra
D remains inside A and the subjectivalisation rule selects A
for agreement with the V
The result is as in figure XXV.

·220 SANSKRIT SYNTAX Figure XXV S -- 1 NP r+pres. +-sing: L+third ya *+-sing, +third -+fem. N bhagini NP K mitra [t-sing. --mas. fem. The affia-substitution rule replaces features with affires. (It may be noted that a morphophonemic rule obligatorily substitutes 'sya' (genitive) for features under K domnated by D) The result is as represented in figure XXVI and the final form as in figure XXVII. Figure XXVI S M t1 NP K -- ya h N bhagini NP mitra sya ·GENITIVE CONSTRUCTIONS 221 Figure XXVII S mitrasya-bhagini yāti IV.3.2 1.1.2 Consider, nert, sentences 25-32 which have an attribute element also Only sentences 25-26' will be described completely Sentences 27-32 have a derivation similar to that of 25-26' 25 & Es sevakasya - netram hanam - asti D nom T nom of servant - eye blind -1S (the servant is blind in one eye) 1 25'. sevahasya - netrena - kānena - bhūyate D 8 Es T of servant - by eye - by blind - is become ((lt) is become blind in one eye of the servant) 26 sevakah - netrena - kanah - 2sti D nom Es T nom servant - by eye blind - 15 (the servant is blind in one eye) 26' sevahena nettena - banena bhūyate Es 1 T by servant - by eye - by blind - is become (it) is become blind in one eye by the servant) 1 1 ·222 SANSKRIT SYNTAS sa varnena 1 1S nom IS 1 nom 27 tasya - varnah - gaurah - asti G g Es nom T nom his colour - fair 15 (his colour is fair) 28 gaurah - asti D nom Es T nom he with colour - fair (he is fair in colour) 29 tasyāh - prakrtih - madhurā - asti D g Es nom T her nature - Sweet (her nature is sweet) 30 sā - prakrtyā - madhură - asti Dnom Es T she - with nature - sweet (she is sweet by nature) 31 bālıkāyah -ālrtih - mañjulā - asti D Es nom T nom of girl - look - lovely (girl's looks are good) 32 bālikā - ākrtyā - maõjula - asti D nom Es 1 T girl - by look - lovely (the gic is lovely by looks) Sentences 25 26 have three elements (1) possessor-'ccsal assa/ sevahah' (11) body part - 'net-am/netrena' and (111) attribute ‘kanam/hanena', which will be referred to as 'P', 'B', 'At' respectively for discussion belon Sentences 25 26 have the underlying structurc 3$ in figurc LXVIII -IS 1S - IS ·GENITIVE CONSTRUCTIONS 223 Figure XXVIII S M [+pres 1 Fs T as NP K NP K hana N netra NP K sevaka Now, D may or may not remain inside Es D is not promoted 1e remains inside Es, and the subjectivalisation rule selects Es for agreement with the V The affix substitution rule replaces features with affixes It may be noted that T takes on the features of Es (these two processes are similar to those in figures XXV, XXVI above and hence are omitted) The final form is as represented in figure XXIX Figure XXIX S Es sevahasya netram V asti Kanam ·224 SANSKRIT SYNTAX Thus, sentence 25 is derived from figure XXIX. If, at the stage of figure XXVIII, the subjectivalisation-rule does not operate, the process of passivisation without subjectis alisation applies and sentence 25' is derived.22 Alternatively, at the stage of figure XXVIII, D is promoted23 and the result is as in figure XXX (from figure XXVIII). Figme XXX S M 1 [+pres.] Es - as NP K NP K NP netra sevaka Läna 22. 23 This process of passivisation without subjcctiralisation operates in the case of sentence 26' as well, and may operate in the urderlying structures of senter ces 27-32 also As it has been explained more than once before, it will not be referred to again unless necessary. It may be recalled that in the case of sentence 21 (figure XXIV), 110 possibility of this optional promotion of D is indicated This optionality of promotion is possible only when the Gentic and the related NP 15 in body.put(or part-It holc) relation, along with the co-occurrence of Tinits casc frame. In such constructions, only Es can dominate the Din a simple sentence. The CASC-category L (in body-part or part-u holc relation) also may dominate D(sce IV. 3 2 1. 1. 4), but there is so Tin the case frame. In fact this case-catcgory Taleo is constrained by its attributise feature.

·GENITIVE CONSTRUCTIONS 225 Dis selected for subjectivalisation and T takes on the features of D.. The result is as represented in figure XXXI. Figure XXXI- S ΜΣ 1 NP K +-pres. +sing: +third - sevaha +-sing: +third +mas. Es T - as NP K NP r+sing 7 Lana [#mas. netra mas, -fem, 24. There arise three points. In the first instance, figure XXX shows a case frame like ( Es, D, T) But, in the chapter on 'NP-asti-NP constructions', it has not been mentioned that the casecategory Es can occur with D In the second place, what casecategory, if any, may be assigned to 'sevaka' (1.e. D) after its promotion (as in figure XXX) - as it is evident that no caseframe containing Dhas been suggested 10 the chapter III, Then, would it be necessary to relabel D as Es, after its promotion ? The answer is in the negative as cases are specified in underlying structures. A way out of these problems suggests itself in the fact that in such sentences D is always dominated by Es, and as such it need not beindicated in the case frame which is specified as [Es, T --- ). Hon ever, a rule may specify that in case D is promoted, the subjectivalisation rule selects D rather than Es (Es is usually selected in the case of an 'NP-asti-NP construction). 15 ·226 SANSKRIT SINTAX The affix-substitution rule applies and replaces features with affixes. It may be recalled that the features under K dominated by Es, are obligatorily replaced by '1'- affix, in case Es is not subjectivalised. The result is as represented in figure XXXII. Figure XXXII S M - ti 1 NP K Es sevaka h as NP K NP ina kana netra The final form is as in figure XXXIII. Figure XXXIII S sevahah T Es 1 netrena asti Lắnah Now, the T talcs on the features of this D (which has been dominated by Es and aow stands promoted), rather than that of Es (in an 'NP-asti-NP' construction T tales on the features of Es). With these processes, there appears to be no aced to relable D after its promotion, which would, others ise, add to the complexity in the form of tro occurrences of the same casccategory (if Dis sciabclled as Es) in a simple sentence ·GENITIVE CONSTRUCTIONS 227 refers to Thus, sentence 26 is derived from figure XXXIII. As mentioned earlier (f. n. 22) sentence 26' can be derived in case the process of passivisation without subjectivalisation operates at the stage of figure XXX. It may be noted that all sentences (25-32) have the elements P(ossessor), B(ody-part)25 and At(tribute). To account for such constructions (25, 27, 29, 31) Fillmore26 a schema like (i) [P, B] - asti-At. An alternative schema like (ii) Pnom asti (B. At) is proposed for constructions like 26, 28, 30, 32. As shown above, the case-categories D, Es, T occurring in a case-frame, dominate NP's which indicate Possessor), B(odypart) and At (tribute) respectively. The schemata (i) and (ii) above, show the surface-realisations of sentences like 25, 27, 29, 31 and 26, 28, 30, 32 respectively. IV. 3. 2. 1. 1.3. Consider, next, sentences like 33-35, in which the At(tribute) is replaced by a nominal specified as T. 33. bālakah - pustakaib chātrah Es nom T nom boy with books - student - 15 (from his books, the boy appears to be a student) asti 25. One may argue that there is an extension of the notion of 'body part' (in case of sentences 27-32). The term 'body-part may be appropriately replaced by 'part-whole' because the possessed NP, even if not a body-part, is conceived as organically related with P(ossessor) The term 'body-part' is retained to have the minimum modification in the rule referred to by Fillmore. Fillmore, C. J., (1968a), p. 64 : Fillmore is not cxplicit as to what is meant by a symbol like '+. It may be interpreted as P(with genitive inflexion) related to B. The schema as mentioned by him is [Pgen → B] bc A. 26.

·228 > SANSKRIT SYNTAX asti 34. sa Es nom 1 asti 1 kejaib tāpasah T nom he with hair - meadıcant - 15 (from his hair, he is a mendicant) 35 yuvati Sindurega sabhartrkā Es nom T nom young lady - with vermilion - with husband is (the young lady appears married (with husband) on account of vermilion) The underlined NP's - pustahaih' 'Leśaih', 'sındürena' - in sentences 33 35, which have been referred to as B(ody part), may perhaps justifiably, be called 'part-whole'. Whether the appropriate term is body part or 'part-whole', the relation between the Possessor) specified as Es(33-35 above) and the possessed NP's (underlined and unspecified as to their case-category) via T(in above sentences (33 35)), is inalienably conceived. In sentences 33 35 the Possessor) is identified with the nominal specified as T The NP's- 'pustahain', 'kesaih' 'sindūrena' - are considered inalienably celated to T('chātrah', 'tāpasah', 'sabhartrkā) It may be noted that through this relation between the NP's (bodypart part-thole) and NP's specified as T, the Plossessor) (specified as Es) becomes inalienably related to the NP's (body-part/part-whole) This is why the relation between 'balahah - pustahan' (boy with books) in sentence 33, unlike that between bālakasya-pustalam' (of boy-book) in sentence 1 (IV. 1), is not an alienablc one Similarly, the relation between 'yuvati' and 'sindūrena' in sentence 35, is inalienable This position is further supported by the fact that sentences 33, 35 arc unacceptable 33. * bālalih - mohanasya - pustahaih chátrah-asti boy with books - Studert - 15 (thc boy is a student on account of m's books) of m.

·GENITIVE CONSTRUCTIONS 229 35' * yuvati - pitahasya - sindūrena - sabhartıkā - asti young lady-of cashet-with vermilion-married-1 (the young lady is married on account of the vermilion of the cashet) Sentences 33-35 show that the inalienability is probably a category of grammar. The set of elements over whuch this category ranges is language -specific, in that it is dependent on the culturally bound definitions of individual lexical items Compared with sentences 26, 28, 30, 32, sentences 33-35 differ in some respects. In 33-35, T is not an attribute and it is identified with the Possessor) specified as Es and the B(ody part) (1 e 'pustahaih', “hesaih', 'sındūrena') constitutes the defining criteria for Es to be identified with T. These distinctions account for the fact that sentences 33" 35" comparable to sentences 25, 27, 29, 31 are not acceptable. 33 * balakasya - pustakāni - chatrāh g T nom books - students - are tasya - heśah - tapasāh g nom T his hair mendicants - are 35 yuvatyāh - sindūram sabhartrkam - asti T of young lady - vermilion - married santi nom of boy 34 * santi nom nom nom 1S There appears to be no need to describe sentences 33-35 through all stages of their derivation which would be similar to that of any 'NP - asti - NP' sentence (chap III) as is evident from the underlying structure of sentence 33 as represented in figure XXXIV

·230
SANSKRIT SYNTAX )
Figure XXXIV.
2
M
[+pres.]
Es
-
as
NP
K
NP
bālaha
N
chātra
NP
pustaka
The configuration in figure XXXIV shows that in sentences
33-35, T is a complex structure which can be accounted for
by a schema like NP N(O)
IV. 3. 2. 1. 1. 4. The sentences discussed aborc (21, 25-
35) show kinship and body-part (1. c. part-1 hole) relations.
Consider, next, a sentence lihc 36 which also expresses a
body-part rclation.
36. khalah - bälahasya - hastc ähanti
A nom D & L
wicked - of boy
on hand - Strikcs
(a wicked man strikes thc bog on his hand)

·GENITIVE CONSTRUCTIONS 031 The above sentence has a case frame like[A,L-] wherein L is a body-part and dominates D The underlying structure of sentence 36 is as represented in figure XXXV. Figure XXXV S P M 1 [+pres ] -ähan NP K NP khala D hasta NP K bilaka At this stage (of figure XXXV) there are two options (1) D may remain inside Lor(11)D is promoted Subsequently, a subjectivalisation rule applies optionally Consequently, there are five possibilities as follow (a) If D remains inside L, (1) A may be selected for subjectivalisation as in sentence 26 above ·232 SANSKRIT SYNTAX (11) or I may be selected for subjectivalisation as in sentence 36(a) below. (111) the subjectivalisation-rule may not operate as in sentence 36(b) below. Dg 1 g L 1 A 1 (b) If D is promoted, (1v) A is selected for subjectivalisation as in sentence 36(c) below. (v) Dis selected for subjectivalisation as in sentence 36(d) below. 36(a) bālahasya - hastah - khalena ähanyate L nom A of boy - hand - by wiched - is struck (the boy's hand is struck by a wicked man) 36(b) bälahasya - haste - Khalena āhanyate D of boy - on hand - by wicked - is struck ((it) is struch on the boy's hand by a wiched man) 36(c) Thalah - bālakam - haste āhanti A nom 1 wicked - boy - on hand - stahes (a wicked man strikes the boy on his hand) 368) bälahah - haste - Whalena -ahanyate D nom L 1 А boy - on hand - by wiched - is struck (the boy is struch on his hand by : wichcd man) Nou, the option (a) above is selected (t.e the D promotion rule does not applv) Ais selected for subjectirilisation and the cesult is as in figure XXXVI асс 1

·GENITIVE CONSTRUCTIONS
233
Figure XXXVI
M
|
NP
K
r+pres
+sing:
-+third
|
khala
r+sing:
+third
+mas.
āhan
NP
r+sing.
-+-mas..
hasta
NP
+sing.
+mas.
The affis-substitution rule replaces featurcs with affixes. It
may bc recalled that the afius genitac ('s) a) is obligatorily
added to D in case it is not promoted (cf. figurcs XXVI,
XXIX). The result is as represented in figurc XXXVII
The final form is as in figure XXXVII, from sluch wc
obtain sentcncc 36 aborc.

·234
SANSKRIT SYNTAX
Figure XXXVII
S
M
NP
K
khala
ahan
N
hasta
NP
bālaka
sya
Figure XXXVIII
S
khalah
āhanti
balakasya-hastc

·GENITIVE CONSTRUCTIONS 235 (11) Alternatively, at the stage of figure XXXV, Lis selected for subjectivalisation and the result is as represented in figure XXXIX (from figure XXXV) Figure XXXIX S NP M *+pres +sing, L+third *+sing +third. ahan +mas NP hasta NP K khala r+sing tmas bālaha L+mas The affis-substitution rule replaces the features with affixes It may be noted that selection of L for subjectivalisation, usually does not trigger the attachment of 'te' under M and the addition of -ya- to V (cf II 4 4 3) But here, with body part I these necessarily tal e place The final form (omitting the intervening details) is as represented in figure XL from which sentence 36(a) above is obtained. Figure XL S bālahasya -hastah āhanyate Khalena ·236 SANSKRIT SYNTAX (iii) Alternatively, if the subjectivalisation-rule does not operate and no case-category is selected for agreement with the V, the process of passivisation without subjectivalisation applies and the result is as represented in figure XLI (from figure XXXV). Figure XLI S M - *+pres. +-sing -+third ahan NP K NE K khala [+sing. +mas. #sing:] L+mas. hasta NP K bālaka r+sing. .+mas. The affix-substitution rule replaces the features with affixes and the final form (omitting intervening details) is as represented in figure XLII from which sentence 36(b) is derived. Figure XLII سبب āhangate khalena balakasya • hastc ·GENITIVE CONSTRUCTIONS 237 Now, the option (b) is selected i e D - promotion applies. The result is as represented in figure XLIII (from figure Figure XLIII S M [+pres ] āhan NP K NP K NP Khala hasta bālaka Now, at the stage of figure XLIII there are two possibilities : either A is subjectivalised or the process of passivisation operates and selects D for agreement with V(it may be noted that the D after its promotion behaves as R) (1v) A is selected for agreement with the V and the result is as represented in figure XLIV Figure XLIV S M P. | NP +pres +sing L+third khala +sing +third L+mas ahan NP K NP | hasta [: +sing ] balaka [+sing +mas +mas ·238 SANSKRIT SYNTAS The affix-substitution rule applies and replaces features with affixes The final form (omitting loter ening details) is as represented in figure IV from which sentence 36(c) is derived Figure XLV S P Khalah L āhantı bilaka haste bālalam (v) Alternatively, a subjectivalisation rule does not apply and the process of passivisation operates and selects D foc agreement with the V The result is as in figure XLVI Figure XLVI S M NP K Tpres +sing third] balaha balala r+sing #third -+mas ahan NP K NP halas +sing Thastar Asing tmas L+mas ·GENITIVE CONSTRUCTIONS 239 The affix-substitution rule replaces the features with affixes. The final form (omitting intervening details) is as in figure XLVII from which sentence 36(d) is derived Figure XLVII S P bālalah A A nom - comes ahanyate Khalena haste All the sentences discussed above (21, 25 32, 26) are such in which the Genitive and the related NP are in Linshup or bodypart (part whole) relations The schema proposed for such constructions is NP -> N (D) Fillmore would like to have this schema for a sentence like 37 as well 37 [nidesahasya — sacıvah] - āyatı AD of director secretary Fillmore is not explicit whether the relation between the D ‘nidesahasya' and 'sacıvah' is alienable or inalienable, though by implication (as he refers to the schema mentioned above) he rules out a sentential derivation of such construc It appears that the occurrence of NP's like 'sacivah' (secretary), ‘lipikah' (copyist), 'taNkahah' (typist) which in dicate profession, allow the occurrence of 'parśve' in a non locational sense, as in sentence 38 38 tasya pārsve - lipihah/taNhahah/sacıs aḥ g his copyist/typist / secretary - is (he has a copyist/typist /secretary) tions - asti ·240 SANSKRIT SYNTAX nom It is considered, therefore, appropriate to treat sentences like 37 and 38 with a sentential derivation (which is discussed in IV. 3. 2. 4.). Apparent exception to this position may be sentences like 39 wherein thc Genitive and the rclated NP involve a designation like that in sentence 39. 39. kulapateh - sacivah of vice-chancellor - secretary (vice-chancellor's secretary / secretary to the vicechancellor). It appears-7 that constructions like 39 can better be treated according to the schema NP + N(S). IV. 3. 2. 1. 1. 5. Consider, next, sentences like 40-43, wherein the items related to Genitives refer to inanimate Possessor) in part-whole relation. 40. asandyāḥ - pādah - navah - asti G Es T nom of chair - leg is (the leg of the chair is new) 41. phalakasya - prșham — nirmalam — asti G Es T of board - surface clean - is (the surface of the board is clean) 42. pustakasya — pisthāni pavāni santi G Es T nom of book pages (the pages of the book are new) - dvāram — laghu asti G Es nom T of house -door small (the entrance-door of the house is small) It has been argued that the distinction between 'alier able' ard inalienable is not relevant here as ore can have a sentence like 'pralosthāt-sacivah - āgatah : 'a secretary emerged from the office' (Stockwell, R. P., p 736). — new nom nom nom - new are 43. grhasya nom is 27.

·241 GENITIVE CONSTRUCTIONS - asti -new -1S - asti - 15 - santi (It may be noted that sentences like 40-43 call for a modification in the schoma NP N(D) (cf. IV.3.2.1, f. n. 20) None of the above sentences (40-43) can, according to the earlier criterion, have a sentential derivation for the Genitives It may be recalled that all such Genitives which have alienable relationship with related items (and have a sentential derivation) allow the occurrence of 'pārsve'. But sentences 40-43' arc unacceptable and as such sentences 40-43' cannot harc a sentential derivation 40'.* āsandyān pārśse - pādah - navah of chair - leg 41'.* phalahasya pārśve - prstham - nirmalam of borad - surface clean 42.* pustahasya pārsie - prsthāni - navāni of booh pages gthasya pārśve - dvāram - laghu of house - door - small However, a sentence like 44 wherein the Genitive and the associated item are not inalienably related, is acceptable even without a locational reference (sentences lihe 44 are discussed in IV. 3 24). 44. guroh - parśve28 lehhani G 8 Es nom of teacher pen (the teacher has a pen) IV 3 2 1 1 6. Consider, next, sentences like 45 which contain a directional indicator and exhibit the inalienable relation between the Genitive and the directional indicator. - new - are - asti 43' * -1S - asti 15 28 ‘parsve' may mean ‘near also Here it is not semantically significant i e has no locatioral refererce In 44, it gives the sense which would be captured by 'have' in English ard'he pā a' in Hindi 16 ·242 SANSKRIT SYNTAX mama asti - to left - 15 45. guruh vāmatah Es nom L teacher - my (the teacher is to my left) The relation between 'mama' and 'vāmatali' may be treated as body-part relation. Though NP's indicating direction may occur without any personal references as in 45', a personal reference is always implied. 45' guruh — vāmatahasti teacher — to left (the teacher is on (my) left) Pechaps one may argue that in 45', the reference may be other than the speaker (1 e. 'mama') and the directional NP 'vāmatah' may involve reference to a building or anything else. However, this interpretation is possible only in a context. In fact, the absence of an overt form of personal reference (in 45 ) is due to the fact that they usually refer to position or direction with respect to the speaker or addressee in an utterance, and there are simply many situations in which an adnominal D need not be espressed if it identifies speaker IS or hearer '29 One may suggest that Ross's postulation of a performative analysis30 in the form of a superordinate structure containing the NP “aham' (1) may account for structures libe 45. But, the performative analysis of declaratite sentences does not appear to be strongly motivated in the case of Sanskrit. In a performative analysis, the superordinate structure 'must have first person subject' and 'usually 29 30 Fillmore, C J. (1968), p 80 Ross, J R, (1970), pp 222 272 ·GENITIVE CONSTRUCTIONS 243 occurrence have second person direct or indirect objects in deep structure,"31 Ross considers the of the NP faham' I' in the superordinate performative structure the most significant and gives 'seven arguments suggesting the existence of a higher subject I "32 Ross's other arguments33 relate to the occurrence of the refleuve 'myself' in different structures without the co-occurrence of the pronominal, which necessitates that there is a superordinate structure with T, which would be deleted through a performative deletion-cule, after it has led to the reflexivisation of the co-referential NP In Sanskrit, such exigency does not arise, as would be shown by sentences 46 and 46'. In 46 the reflexive does not occur. 46 patram - rāmena - mayā - (svayam) - ca - alıhhyata O nom A A 1 letter by r - by me - (self). and - was written (the letter was written by r and by myself) 46. mama/mayā -- samah chātrāh my/ with me like (equals) - students (students like myself) Thus, it seems that as Sanskrit has no exact parallel to the English reflexive, and no other motivation is apparent, Ross's arguments do not carry over to Sanskrit Now consider a sentence lihe 47 for which Fillmore3t would propose a schema like NP-->N (L) 47. parvatasya — melhala O 8 of mountain — slope (slope of the mountain) - EN nom 31 32 33 34 Ross, J R, (1970), p 222 Ross, J R, (1970), pp 226, 236 Ross, J R, (1970), pp 225 236 Fillmore, C J. (1968a), p 80 ·244 SANSKRIT SYNTAX But according to him, 'mehhalā (slope) would be a 'locational noun' and 'parvatasya' (of mountain) also would be an 'adnomunal, L. The NP 'mchhala' indicates a partwhole relation and may be assigned a case-category which is not necessarily L. But, there appcars to be no reason to assign the case-category L to 'parvatasya'. An inanimate Genitive need not be L always (cf. sentences 40-43), and it is more appropriate to specify it as O like the Genities in 40-43 above. Thus the schema proposed by Fillmore may be modified as NP-→N(0) In construction 47 above, the relation between the Genitive and the associated NP is a part-whole and thus inalienable. A construction like 47' is unacceptable It may be noted that even in the sense of locational reference the occurrence of ‘pārsve' would not make 47' acceptable. 47'. paravatasya păcśve - mehhalā of mountain - slope IV. 3 2 1. 1. 7. Now, take the problem of some nouns like 'pārśve' (proximity), 'purastāt' (in front), 'parastāt' (beyond) which indicate a location or direction 'with reference to the associated object but not considered as a part of it. "35 Such nouns36 are semantically significant37. Such nouns involve some reference-point and the relationship between them and the objects with reference to which they are associated, is akin to the inalienable relationship between the possessor and the possessed item Such nouns convey spatial/temporal orientation and can be conceived only in relation to the associated object. * 35 36 Fillmore, C J. (1968a), p 81. They are referred to as P(ost) Position), though no restriction to sequential ordering is assumed Robinson, J J. (1969), pp 58 59 CF Kachry, Y, (1969), p 39 37 ·GENITIVE CONSTRUCTIONS 245 The occurrence of 'pārśve as in sentences 48-50 is not permissible. Sentences 48-50' are unacceptable. 48 vidyālayasya — purastāt 8 of school in front (in front of the school) 49. bhavanasya - paścāt g of house behind (behind the house) 50. mandirasya — pārśve near of temple (near the temple) 48'.* vidyālayasya -- pārśve - purastāt of school front of 48".* purastāt - vidyalayaḥ - vidyate in front — school is 49,* bhavanasya - pārśve paścāt of house behind 49".* paścāt bayanam asti behind house is 50'.* mandırasya - pārśve — pārsve of temple 50".* pārśve mandiram asti temple is Sentences like 48", 49' 50' are acceptable in discourse, but they would be interpreted in reference to some object other than that in 48°, 49', 50'. All such constructions 48-50 are accounted for by a schema like NP- N(O), the NP being dominated by L. Thus the underlying structuce of 43 would be as represented in figure XLVIII. near near ·246 SANSKRIT SYNTAX Figure XLVIII L K N pucastāt NP vidyālaya Constructions 49-50 also have an underlying structure similar to that in figure XLVIII. Amorphophonemic rule, later in the derivation, attaches the genitive-affix to O dominated by L. IV. 3. 2. 2. In the foregoing sections (IV. 3. 2. 1. 1.- IV.3. 2. 1. 1. 7), two schemata are proposed : (i) NP N(D) and (ii) NP N(O), to account for the inalienably related Genitives and associated/possessed items. The schema(i) relates to kinship and body-pact relations (sentences 10, 21, 25-32, 36) and the schema (11) relates to part-whole relations inanimate entities (sentences 40-43, 48-50)). In sentences 48-50 the case-category L and in sentences 40-43, 47 the case-category O dominate the NP which is expanded according to the schema NPR N(O). In sentences 44, 45 the caste-category L dominates the NP which is expanded following the schema NP— ND). For all the constructions containing Genitives which express inalienable relation, a schema NP-N may be proposed. IV. 3. 2. 3. Before taking up for analysis constructions which are characterised by an alienable relationship between (8) ·CENITIVI CONSTRUCTIONS 247 the possessor and the possessed/associated item, one point needs to be considered in the context of the decvation of N({B}) the embedded D/O. A schema like NP - N somewhat modifies the notion of case, which has been developed in relation to verbs. Postulation of a schema like the above implies an extension of the notion of case in relation to nouns as well. Now, just as selectional features of verbs, determine a case-frame, 39 nouns like svasā, bastaḥ, dāram páráram, (sister, band, entrance, side) may be considered to have aspects of meaning appropriate to selecting a case. But, it may be difficult to accept that nouns like 'ast āsandi' (house/chair) may have aspects of meaning which induce the selection of a particular case.39 This difficulty may be overcome in the case of a sentential derivation of Genitives which indicate an alienable relation between the possessor and the possessed/associated item (IV. 3 2 4). IV. 3. 2. 4 For constructions which indicate alienable possession, Fillmore, proposes a sentential derivaticn 1, e. NP - N(S) It has been mentioned earlier that such constructions are distinguished from those showing inalienable possession, by the occurrence of 'parśve' in the former set of sentences (1, 11, 17, 44) Consider sentences like 51 which is derived from sentence 51'. 51. chātrasya pustaham G g of student book is asti 'there appears to be 38 Stochu ell, RP, et al (1968), p 721 a very close relationship between aspects of the central essential meaning of a verb and the case frame it selects Stockwell, RP, et al (1968), pp 720 21 39 ·248 SANSKRIT SYNTA (the student has a book / there is the book of the studeat) 51'. chátrasja pārśıc pustakam Asti IS of student book In accordance with the schema, sentences 51, 51' have the underlying structure as in figure XLIX Figure XLIX S. NI ITpres ] Es as K pustaka M 1 [+pres ] D Es as PP NP NP K chatea patste pustaka ·GENITIVE CONSTRUCTIONS 249 Fillmore suggests that the NP under Es in S, is a co-referential copy40 of the N 'pustaka' under Es in S.. This NP can be replaced by a relative-macher to derive a relativcclause from S. Es in S, is selected for subjectivalisation and then there is deletion of the items under Es and V under identity. Consequently, the node M is also deleted. The Iesult (omitting intervening details) is as represented in figure L. Figure L S. < N [+pres.) V Es as NP K . pustaha PP NP K pārsve chătra 40 Fillmore, C ), (1968a), p 49 ·250 SANSKRIT SYNTAX The node-pruning rule deletes the nodes S, and P. The result is as represented in figure LI. Figure LI S NI [+pres.) Es as N pustaka NP K PP chatra párśve The subjectivalisation-cule selects Es for agreement with the V and the result is as in figure LII. Figure LII S 1 Es M +pies. +sing. NP K L+third +sing Luthird N -mas. pustaka -fem. 25 NP K PP chātra [#sing. Tpārsve ·251 GENITIVE CONSTRUCTIONS The affix-substitution rule replaces the features with affixes, a morphophonemic cule obligatotily attaches the genitive ('sya') to K under D due to the co-occurrence of 'pārsve'. The result is as represented in figure LIII. Figure LIII S Es M ti as am -Z . pustaka PP NP 1 chātra sya pārśve The final form is as in figure LIV. Figure LIV S Es chātrasya-pärśve-pustham asti A ‘pārśve' - deletion rule optionally deletes 'pārsve'. Thus sentences 51, 51' are derived from figure LIV. It may be added that a sentence like 51 above is ambiguous. In addition to the meaning the student has a book', it may be interpreted in the sense of there is the book of the student'.

·252 SANSKRIT SYNTAX citram Es nom 1s But, in this latter sense the Es 'pustaka' is interpreted as (+definite]. Consider, lastly, a sentence like 52 which is fourway ambiguous. 52 rāmasya asti G of r picture (there is the picture of 1 lr has a picture) Sentence 52 may be interpreted like sentence 51 above and its derivation (similar to that of sentence 51 (figures L-LIV)) would be from a structure underlying sentence 52'. 52'. rāmasya parśte G 8 Es picture Alternatively, sentence 52 may be derived from structures underlying sentences 53, 54 41 53 citram [ramah citram - aracayat ] - asti A F picture [1 picture - composed] - is 54 citram [ [A ramasya citram - aracayat ]-asti Es А D F picture [(someone) - of r - picture - composed] is citram asti nom of 1 1S Es 41 a Of Chomsky, N ,(1970a), pp 200 07 Chomskywould consider construction like 'J's picture ((1) the picture that J has, (11) the picture of J ), to be related to the interpretation (111) the picturethat) punted,'with a generalisation of the notioa 15aliegable possession' to a kind of 'intrinsic connection' (PP 205 206) Cf Kachru, Y, (1969), p 44, (1º66), pp 107 108 Kachru would like to handle the third interpretation separately from the other two

·GENITIVE CONSTRUCTIONS
253
The derivation of sentences 53, 54 which have the underlying
structures as represented in figures LV and LVI respectively,
involves the process of nominalisation which has been
discussed earlier (IV. 3. 1. 1.), and accordingly it will not be
discussed further here.
Figure LV
S.
M
1
[+pres.]
V
Es
as
NP
citra
M
[+pref]
1
fac
NP
-4
K NP
K
|
căma
citra

·254
SANSKRIT SYNTAX
Figure LIT
So
N
1
[+pres.]
Es
as
S
citra
NI
P
[+pref]
V
А
-
тас
NP
N
citra
NP
rāma

·V LEXICON AND RULES V. 1. In this chapter it is proposed to deal with the nature of the lexion in FCG (section V. 1. 1), and specify a set of lexical entries for verbs (V. 1. 2), and a set of rules related to both the base and transformational components (V. 1. 3). V. 1. 1. As mentioned earlier (I 2. 1. 4 f n. 78), FCG is explicated within the conceptual framework of TG grammar. However, the assumptions (I 2 1 3 -1.2 1.3.4.) of FCG lead to some significant modification in the lexicon. V. 1. 1. 1. The base-component of a TG model grammar is considered to consist of a categorial component with unordered context-free rewriting rules called P(hrase) S(tructure) rules, and a lexicon containing a set of unordered lexical entries. Grammatical properties which are not relevant for categorial component (1. e. P. S) rules and those which are highly idiosyncratic could be assigned to the lexicon. As Chomsky favours selection of V's in terms of the prior selection of nouns and considers subject-object as relational terms in the base-component, the leucon assumes a complex form Two sets of rules are proposed : (1) strict sub-categorisation rules analyse a CS(V) in terms of contextual (categorial) features and (11) selectional rules (or selecuonal restrictions/restrictions of co-occurrence) analyse a CS in terms of syntactic features of the frame in which it (CS) occurs Chomsky, N, (1965) sp 64 74 1 1.

·256 SANSKRIT SYNTAX 2 FCG favours the prior selection of V's followed by the selection of other items in terms of V's (I 2.1.3 5). There is more than one reason to commend this position : (1) verbs 'in a very real sense are selectionally dominant', (11) prior and random selection of nouns may lead to a large number of strings which are to be discarded later on, (111) prior selection of verbs has 'side-effects.'3 This position of FCG, along with the consideration that subjects and objects do not play a role in deep structure, leads to a change in the specification of the lexicon. In the lexicon, selectional restaicuons, unlike strict sub-categorisation rules, are not concerned with the domain of one constituent. Selectional restrictions range over a subject noun and a verb as well as an object noun and a verb.4 Accepting that subject and object are surface-phenomena', 1. c they do not originate in the base-component, there appears to be no need to retain this distinction bet'w cen sub-categorisation tules and selectional restrictions. V. 1. 1. 2 FCG suggests that case-relationships may be specificd in the lexicon in terms of framc-features. These frame-features are spelled out in a case-frame which contains the cases with which a verb co-occurs to form a Péropo- 2 3 Stockwell, R. P, et al (1968), p 935 Stoch well, R. P, et al (1968), p 934 : 'side-effects are efects on the other nodes in a tree after an item has been arserted If verbs are selected first, the case-frames for the scrbs indicate the possible casc-category nodes in the troc. Robinson, J ), (1969), p 71 : 'the samples of selectional restrictions exhibit only those imposed by the inherert features of nouns on vcrbs sith which they occur in subject or object relation' Sec 1 2 1 4 (in 84) 4 S ·LEXICON AND RULES 257 a 1 sition) Thus a verb like 'hanti' (kill) has a case-frame like I. 1. hanti [+V] : [ A, (), R -] indicating, thereby, that the verb co-occurs with Agentive, Receptive and Instrumental which may or may not be, present in a construction like 1 dustah (churikaya) pasum hanti A I wicked with knife animal hills (the wicked man hills the animal with a knife) In the above sentence 1, I is an optional case-category. V. 1. 1. 3 Each lexical entry in a case-frame has associated with it a CS which indicates category -features, contextual features and inherent features. A lexical entry like 'hanti' (kill) has a categoryfeature like [-+-V]. Each CS indicates only one category feature which is specified positively. The case-frame like I given above, shows contextyal features i e. the array of cases with which a leucal entry like 'hantı can co-occur FCG proposes that the optional occurrence of an item in the case-frame may be indicated by parentheses, as in the case-frame I given above Again, FCG suggests linked-parentheses to in dicate that one of the linked items is obligatory. Thus, for sentences like 2 and 3, the case-frame would be shown lihe II. II. pacati - : [+V] (cooks) [(A, F) (L) -] 2. sudah sthālyām bhahtam pacati L F cook in pot TICC cools (the cook cools rice in a pot) 17 ·258 { Ext, (GO).(SO) }) } SANSKRIT SYNTAX 3. sthāli pacati L pot cooks FCG proposes to show inherent features of lerical entries by a schema like N [ + animate ]4,D/ [X-Y] It may be noted that the nouns like 'südali'(cool), "dustah' (wiched) are not inherently A (or any other casecategory for that matter) The only inherent feature that is relevant to them is [ + animate ]. The schema is syntactic and says that a construction would be unacceptable if the feature [ --animate ) is specified for an item which is positively specified as A/D V. 1.2 In this section, a list of verbs with their caseframes, is given. The verbs are specified in their [ + singular ], [ + third ), [+ person ) forms. Polysemous forms are treated with only one scading. 1. arbati : [+] (descrics) [D, O, (SO), S -] : [+V] (is) Es (T) 3 sccbatı : [ + V) (desires) [D, S 4 hathayatı [ +V] (savs) [-1, (R), O-) 5 karoti [+V) [4, O, .] (docs)) 6 hridatt [ + V) (plays) (1), (0))) asti . .

·LEXICON AND RULES 259 [D, R) 7. krudhyatı : [ +V] (gets angry) 8. kośyatı : [+] (becomes feeble/slim) 9. ksipati : [+1) (throws) [D, (I) [4.{}60,(Pa),C),(S04) [ A, O, (), (L) ] [ A, (O), (1), (L) - 8 [A, (I), (L.) - ] [A, L-) 10. khanati : [+V] (digs) 11, khắdati : [ +- V) (eats) 12. garjati : [ +V) (roars) 13 gāhate : [+V (dives) 14 gūhati : [ + V) (conceals) 15 gåyatı : [+] (sings) 16 grasate : [+V] (swallows) 17. grhnáti : [ + V) (tahes) 18 Fighratı : [+1) (smells) 19. calati : [ +V) (moics) [4,0, (I), (L), (SO)--) [AO] [ A, O, (L) (-1,0,50,(1),(L) [ 4, 0, 0) - ] [SA, I, (L) -] 20. c2r3t1 : [ + V) (grazes) { A, O, (L) - ) ·260 SANSKRIT SYNTAX [A,O,(SO), (I),(L)-] [A, O, (L), ا [A, L -] [ A, O, (1), (L) —-] [A, S - 16 21. cinoti : [+V] (gathers) 22. corayati : [+] (steals) 23. cumbati : [+] (kisses) 24. cürnayati : [+1] (crushes) 25. Cestate : [+V) (tries) 26. chinatti : [ +V) (cuts) 27. jayati : [+V] (wins) 28. jivati : [+V] (is alive) -29. jānāti : [+y] (knows) 30. įvalati : [+] (burns) 31. tanoti : [+V) (stretches) [ (1), A, O, (L)) ] [ A, (SO), (0) - ] [D, (Ext), (L) -] [DO) [0,1] [ A, O, L, (I), (Est), (SO), (GO) ] 32. tarjati : [+V) (threatens) [AR —] 6. No thcoretical position is implied by this abbreviatory device. As mentioned earlier (I. 2. 1.3. 5.), sentences are embedded under a case (cf. P. S. rules VI. 1. 3. 1.). The occurrence of S in the case-frames of the underlined verbs, indicates that these verbs do not have simple underlying structures.

·LEXICON AND RULES 261 [A, R, (1), (L) - ) [D] [ALL] 33. tādayati : [+V) (beats) 34. trpyati : [+] (becomes satisfied) 35. tarati : [+V] (swims/crosses) 36 tvarate : [ + V) (hurries) 37 damšati : [ +V] (bites) 38. dahati : [+] (burns) [D, S ] [A, R, (1) ] [ A, C, 11, (O) -] : [ +V] [A, O, GO, (I) —] : [ +V] [ A, I, (O) : [+V) 39. dadāti (gives) 40 divyatı (gambles) 41. dišati (orders) 42. dogdhi (milks) 43. paśyatı (sees) [AR- : [+V] [ A, (O)(R), (L) - : [+V] [D, 07.(1)--) : [+V] [A, (Pa), (L) - ] . [+] 44, dhái ati (runs) 45. dhunott (shakes) 46. dhamti (holds) [(1)(I), O-) : [+V) [A, 9, 1), (L)-] ·262 SANSKRIT SINTAY .. [1, R, (I) بنا ISD? 10 5 47. namati :. [+V) (salutes) 48. naśyatı : [+] (peashes) 49. nindati : [ +V) (spcaks ill of): 50. nityati : [ +V] (dances) : [+V) (cooks) [AR] [-1, (L), --] 51. pacati [1A, (O)(F),(L.), ) 18 (40) ആ : [+V] [ A, (O), (L)--] 52. pathati (reads) 53. patati (falls) : [ + V) {{@}so), SA (SO),(GO), (1)-) 10 {A, O, I-] . { (A)(I), R — -] [O-] .. 54. pibati : [+V] (drinks) 55. pidayatı [+] (hurts) 56. puspyati : [+V) (blooms) 57. pūjayatı [+V) (worships) 58. pūrayati : [+] (fills) 59. prcchati [ +V] (asks) 60. plavate : [+1) (floats) [4,(O)I) - ] [(A), O, L } .. [A, R, O-) L, (1) ܝܙ ·LEXICON AND RULES 263 : [ +V] 61. badhnāti (binds) : [+] 62. bhanalti (breahs) [A, {6}, (4), () [{A,0,0) ] : [+V) [O, (1) --] : [+V] 63. bhāti (shines) 64. bhiksate (begs) 65. bhinatti (cuts) [ A, O, (SO) : [ +V] [(4, 0, {4.00 [+V] 66. bibheti (fears) {D.{} -- [1, (O)(F), (L), (1)--) [ A(L) — ] [4, 5 —] 67. bhrjjati : [+V] (fries) 68. bhramati [ +V] (wanders, walls) 69. manjate [+V) (considers) 70. mantrayati : [ +V) (aduses) 71 mathnāti [ +V] (churns) 72. māts [V+) (measures) [AR - ] A TXO), I) - [1, {R}.) L) -1 73. milatı [-V] (comes togetl cr, joins) [4, R -) ·264 SANSKRIT SYNTAX 74. [O-] 75. mīlati : [+] (closes (as the eyes)) mriyate : [+VI (dies) yāti : [+] (goes) [D, (I) -] 76. [A, (SO), (GO), (Ext) (Pa)] 77. [AS] 78. [OR 1 79. [A, R, (L), (1) -] (ā) rabhate : [+V] (begins) locate : [+] (pleases) runaddhi : [+V) (obstructs) rohati : [+] (grows) likhati : [+V] (writes) lunāti : [+V) (cuts) 80. [O, 1), (L): لا 81. [A, F,(I), (GO)--) IS 82. -] { 4,4), 0} 83. [A, (I), O-] 84. {0, (L)--] 85. vapati : [+V) (shaves) vāti : [+V] (blows) visati : [+V] (enters) vartate [+] (1s, exists, is present) saknoti : [+V] (can, is able) [ A, GO -] 86. [D, L ܢܝ 87. [AS] ·LEXICON AND RULES 265 88. : [+V] [A, S ] 89. : [+] sa Nkate (doubts) śiksate (learns) sete (sleeps) śrnoti (hears) s_" --[, (50), 0_1 90. : [+1] [ 4, (L) --- ) 91. [V+1.3) . [D, {B} 92. ) [ A, O, (L) --] 93. siñcati : [ + V) (sprinkles) snāti : [ +V] (bathes) sprśati : [ +V] (touches) [ A, (T), (L)--- 94. [ A, SO. (1) -] {R} $ 95. hanti : [+V] (Lulls) [ A, R, (1) -] 96. hasati : [ +- V ] (laughs) [A] 97. hvayati : [ +V] (calls) [A, R- V. 1. 3. In this section a list of rules is given. Section V. 1. 3. 1. contains the P(hrase) S(tqucture) rules and section V. 1. 3. 2. specifies a selection of more important Transformational) rules. V. 1.3 1. P. S. pulcs : given #S # 1. S S' C2 ; = 2 2. S M P ·266 SANSKRIT SINTAX 3. 4. P V C, ... Ca G=A/D/R/O/F/I/L/SO/GO/Ext/Pa/Es/T C → SNP+K (S? NP NP ca saha : n > 2 NP 5. 6. N *(!!!) N 7. CS 8. V CS 9. M (Tense/Aspect) 10. Tense past present future modal 11. Aspect Sperfective progressive modal V. 1. 3. 2. In this section some important T(ransformational) rules are specified. T rules : I Nominalisation-rule (optional) [X (M[VY]]Z] S S P PS S SD : 1 2 3 4 5 2 3 4.5 SC : 1 [+Nom. ] 7. Not all the cases may dominate S. However, the restrictions on embeddings are not considered, as they are outside the scope of the present study ·LEXICON AND RULES 267 SD: 1 1 [ [ لا ܢܙ Z SD: . II Modality)-deletion cule [W[ X M (VNom Y) ]Z] S S P P S S 2 3 4 5 6 SC 2 4 5 6 ПІ Subjectivalisation-cule (optional) X M V Y S P P S 1 2 3 4 5 6 SC 1 42 3 5 6 conditions : 1. 4= A/D/O/ES OL 2 1f 4=0/F/I/L which dominates *svayamera' as well then 5* C or 3 1f 4 and indicates a natural phenomenon then 5 A/D IV. Copula-deletion rule (optional) X [ C M [ Cop Y Z S P P S SD 1 2 3 4 5 6 SC 2 3 5 6 Conditions 1 Mis [ + pres ] 2 If Y = Vrom it does not show the stemclcmcnt at Passivisation-rulc (optional) [ V C Z S P P S SD: 2 3 4 5 6 SC:(a) 1 42 3 5 6 ܢܝܙ ] 1 [ M han ܢܝܝ 1 + 2 ·268 SANSKRIT SYNTAX 1 conditions : 1. 4=R/FO (and GO when V dominates a synonym of 'yā' (ti)) 2. 5=A/D (or I indicating a natural phenomenon) elsewhere : SC : (b) SC : (b) Ø 3+2 4 5 6 +ya VI - Modality)-attachment rule [ X MI [ Y ) ) Z S' P P S SD: 1 2 3 4 5 6 2 4+3 5 (this rule follow's Affix-substitution rule which is not given here) ) SC : 1 ·VI CONCLUSION VI This chapter sums up those points of interest in FCG wluch have arisen in the analysis of the foregoing chapters VI 1 As mentioned earlier (I 2 1 3 5, I 2 1 3 ), FCG holds that a case-system could be accepted as a linguistic universal Unless a large number of languages is investigated from this point of view, nothing in this connection can be attested conclusively At the present stage of research there appears to be insufficient evidence for or against this position VI 2 According to FCG cases are syntactically justified semantic primitives? The cases that have been discussed in Chapter I, are distinguished from each other on syntactic as well as semantic criteria However, the distinction between A and Dis motivated by syntactic consideration only (II 1 1) In so far as the semantic justification for cases is concerned, O appears to pose a problem Its definition remains imprecise and its use as a 'waste basket" implies that there is some element of indeterminacy in the specification of this case category 1 Fillmore C ), (1968a), p 5 cases are 'semantically relevant syntactic relationships id (1971) p 254 deep structure cases could be discovered and justified by syntactic criteria pp 246 247 also It appears that Robinson J J (1969) is not correct when she says that in FCG cases are syntactic prim tives (p 62) Fillmore C J. (1971), p 251 2 ·270 SANSKRIT SYNTAX Again, there are certain constructions in which a case-category (e.g. OJI) is not consistent in its syntactic behaviour (e. g. in relation to the process of passivisation : II. 2. 1. 3. 1., II. 2.1.3. 2, II. 3. 2. 2.). This inconsistency may be accounted for by certain features on the lexical items or the lexical items themselves. However, it indicates that the cases by themselves are not adequate. If ad hoc features need to be specified then the major justification for case's seems to disappear. In certain constructions (II. 4. 7. 2.), there is semantic distinction between surface-realisations of the same Case-category. There appears no way to account for this except in terms of their morphological characteristics. VI. 3. The proposal of FCG that 'case-form' or 'surface-cases' are language-specific and may be realised by different syntactic devices like inflexion, pre-/post-positions, word-order, (which is, of course, mentioned by others as well (I. 1. 4., f. n. 28)) appears to be correct. VI. 4. FCG maintains that there is one occurrence of a case in a simple sentence. But there are instances where this principle breaks down. Cases like L and Pa (II. 4. 4, II. 4. 8.) appear to go against this principle of 'one-instance-per-clause'. The case-category I, indicating emotive states also, poses difficulty (II. 3. 4.) as there are simple sentences which allow two occurrences of I. Even the proposal of deriving such I's from superordinate structure does not help. To get round the above problem, it is suggested that this non-application of the principle may be considered specific to L, Pa and I. VI. 5. Thc position of FCG that relationships involving subject and object arc surface-phenomena, appears ·CONCLUSION 271 to be correct and no problem with reagrd to it, has arisen in this analysis VI 6 In addition to the above (VI 1 VI 5), there are a few points that need to be mentioned VI 6 1 In the case of NP asti NP constructions (Chapter III), Es and T appear to be different from other case catego- Lies, with regard to case specification (III 1 1, III 1 2) VI 6 2 Lastly, in so far as the descriptive adequacy of FCG (I 2 1 3 1) is concerned, the analysis in the foregoing chapters, is an insufficient evidence for saying anything cond asive However, FCG appears to be observationally adequate for the description of the Sanskrit language ·BIBLIOGRAPHY 2 Allen. W. S., (1953) Pbonetes ir. Ancient Irdia, Oxford Universits Press. Anderson, J. M., (1971) The Grammar of Cast, Cambridge University Press. Ananthanarayana, H.S., (1970) "The Kāraka Theory and Case Grammar', Indian Linguistics, 31 (1-2), pp. 14-27. Apte, V. S., (1921) A Guide to Sanskrit Composition, Bombay, Gopal Narayona Co. Bach, E., (1967) 'Have and Bc in English Syntax, Language, 43 (2.1), pp. 462-485. (1968) Nouns and Noun Phrases', Untersals in Linguistic Tleory, (cds.) Bach, E.&R. T. Harms, New York, Holt Rinchart ard Winston, pp. 90-122. Bacon, N. S., (1971) 'On Defining 'Cognate Object';' Glossa, (5(1), pp. 71-79. Belvalkar, S. K., (1915) Systems of Sanskrit Grammar, Poona. Bennett, D. C., (1970) 'Some Obserrations concerning the Locative Directional Distinction' (mimcographed), pp.1-35; also Semiotta, (V) (1) (1972), pp. 58-88. Bharthari, Väkyapediya of Barishari, (eds.) Abhyankar, K. V., & V. P. Limasc, (1965), University of Poona, Poona, Bhattoidiksita, Siddhinta Laumudi, (cd.) l'andit Giridhar Sarma, (1964), Varanasi, Motilal Banarsidass. Bhojadcza, (1160 A. D) Seramnail-kerfittercke (cd.) T. R. Chintamani, 193", University of Madras. Blaks, r. R. (1930) 'Semantic Analysis of Case', L«***, Morograph No. 111, pp. 34-49. 18 ·274 SANSKRIT SYNTAX Bloomfield, L, (1935) Largslage, London, George Allen & Unwin. Burron, T., (1955) The Sanskrit Language, London, Faber and Faber. Cassidy, F. C, (1937) 'Case in Modern English', Language 13(3) pp. 240-245. Chatterjee, S, (1967) 'Early Nyāya-Vaiśesika', History of Philosophy · Eastern and Western, Vol. I, (ed ) S. Radhakrishnan, London, George Allen & Unwin, Pp 219 230. Chomsky, N, (1957) Syntactic Strictures, The Hague, Dlouton. (1965) Aspects of the Theory of Syntax, Cambridge, M. I.T. (1969) 'Deep Structure, Surface Structure and Semantic Interpretation' (mimeo graphed), Indiana University Linguistics Club; also Semantics (eds) Steinberg, D.D. & L.A. Jakobovits, (1971), Cambridge University Press, pp. 183-216. (1970a) 'Remarks on Nominalisation, Readings in English Transformational Grammar, (eds) Jacobs, R. A. & P. S. Rosenbaum, Mass , Ginn ard Co., pp. 184-221 (1970) "Some Empirical Issues in the Theory of Transformational Grammar (mimcographed), Indiana Unuersity Linguistics Club, PP. 1-72 Dougherty, R. C, (1970) Roccat Studies on Language Universals', Fourdations of Language, 6(4), Dordrecht, D. Rcidel Publishing Co pp. 505-561. (1970b) 'A Grammar of Coordinatc Conjoined Structures I'Larguage, 46, (4) pp. 850-898.

·BIBLIOGRAPHY 275 .(1971) Edgren, H. (1885) Faddegon, B., (1936) A Grammar of Co-ordinate Conjoined Structures L', Language, 47, (2), pp. 298-339. A Compendious Sanskrit Grammar, London, Trubner & Co. Studies on Panini's Grammar, Amsterdam A Proposal Concerning English Propositions', 19th Monograph on Languages and Lingarstues, (ed. Dinneen, F.P., Washington, D. C., Georgetown University, Press, Fillmore, C. J., (1966a) pp. 19-33. (1966b) (1968a) "Toward a Modern Theory of Case' Project on Linguistic Analysis Report . 13. The Ohio State University Research Foundation, pp. 1-24; also in Modern Studies in Englısh (Readings 11 Transformalional Grammar), (1969), (eds.) Reibel, D. A. & S. A. Schanc, New Jersey, Prentice-Hall Inc, pp. 361-75. The Case for Case' Unrrerrals in Lingmiste Tbeory (eds.) Bach, E. & R T. Harms, New York; Holt, Rinehart ard Winston, pp 1-88. 'Lexical Entries for Verbs', Fourda- 11gns of Language, (4) pp 373-393. *Types of Lexical Information', Studies in Syntax and Semanties, (ed.) Keifer, F., Dordrecht, D. Reidel, pp. 109-137, also (1971) Semantics (eds.) Steinberg, D.D. & L. A. Jakobovits, Cambridge University Press, pp. 370-392. 'The Grammar of Iitting and Bocahing', Rerdings in English Transformatora) Grammar, (eds) Jacobs, (1968b) (1969) (1970a) ·276 SANSKRIT SYNTAX Fillmore, C. J., (1970b) R. A. & P. S. Rosenbaum, Mass., Ginn & Co., pp. 120-133. 'Subjects, Speakers and Roles', Vorking Papers in Linguistics, (No.4), The Ohio State University, Computer Information Research Centre, pp. 32-63. (1971) 'Some Problems for Case-Grammar Working Papers in Linguistics, No. 10; The Ohio State University, pp- 245-265 ; also in Monograph Series on Languages and Linguistics (eds.) O'Brien, R. J., (1971), Georgetown University School of Languages and Linguistics, pp. 35-56. Gleitman, Lila R., (1969) 'Coordinating conjunctions in English', Modern Studies in English: Readings in Transformational Grammar, (eds) Reibel D. A. and S. A. Schane, New Jersey, Prentice-Hall Inc, pp. 80-112; Reprinted from Langnage 41,(1965), pp. 260-293. Goldstucher, T., (1861) Pānını : his place in tle Sanskrit literature, Trübner & Co, London. Gruber, J. S., (1967) 'Look and See', Language, 43, pp. 937-47. Halliday, M. A. K, (1966) Some Notes on Deep Grammar', Journal of Linguistics, pp. 55-67. Huddleston, R. D., (1970) 'Some Remarks on Case-Grammar Linguistic, Inquiry, I(4), pp. 501-511. Jhalkıkar, B, (1893) Nyayakos'a Bombay, Govt. Central Bool Depot Kachru, Y , '(1966) An Introduction to Hirdi Sulax ; Deptt. of Linguistics, University of Illinois, Urbana. (1970) A Note on Possessive Constructions in Hindi-Urdu', Jutral of Linguistics, 6 (1), pp. 37-45.

·BIBLIOGRAPHY 277 Kielhorn, F, (1965) King, R. D, (1969) Kiparsky, P & J F, Staal, (1969) Kooijs J G. (1971) Kumar, BS, (1976) Lahiri, P C,(1933) Lakoff, G & Stanley Peters, (1966) Katyayana and Panın (reprint of 1876 edition), Osnabruck, Otto Zeller, Historical Linguistics and Generative Grammar, New Jersey, Prentice Hall, Inc. 'Syntactic and Semantic Relations in Pânını, Fourdations of Language, 5, PP. 83-117 Ambiguity in Natural Langucge, Amsterdam, North-Holland Publishing Co Is There a Case for the Sambodhana Kārala ?' paper presented in the VI All India Conference of Linguists, Tirupati Studies in the Word Order of Sanskrit Prose, London University unpublished PhD thesis *Phrasal Conjunction ard Symmetric Predicates, in Micdern Studies in Englisb Readings in Transformational Crammar, (eds) Reibel, D. A,& S A Schane, New Jersey, Prentice- Hall Inc ,Pp 113 142 (Reprinted from Matbematical Linguistics and Glomatic Translation), Harvard Computation Laboratory Report No NST- 17, 1966, pp VI-1 to VI-49 Instrumental Adrerbs and the Coacept of Deep Structure', Tourdations of Language, 4, pp 4 29 Abstract Syntax and Latin Complemenfation, Mass, Cambridge M I T. press 'Being ard llaing in Estorian', Tonditions of larguage, 5, PP 324. 31 i Rieder in Norharth Cerity Ilustariced Ind Errefeos Lursouties, Lakoff. G (1968) Lakott, R. T, (1968) Lchiste, 1, (1969) Lehmann, 1', (1967) ·278 SANSKRIT SYNTAX Lehrer, A., (1970) Lyons, J., (1966) (1967) (1968) Macdonell, A. A., (1911) Bloomington, Indiana University Press. 'Verbs and Deletable Objects'. Lingla, 25, pp. 227-253 • "Towards a National Theory of the Parts of Speech', Journal of Linguistics, 2, pp. 209-236. 'A Note on Possessive, Existential and Locative Sentences', Fonrdation of Language, 3, pp. 390-396. Introduction to Theoretical Linguistics Cambridge University Press. A Sanskrit Grammar, London, Longmans Green & Co. A Comparetive study of the concepts of space and Time in Indian Thoughi, Varanasi, Chowkhambha Sanskrit Series. Indian Theorists on the Nature of the Sentence (vāhya)', Foundations of Language, 2, pp. 377-393. 'A Sanskrit Grammar, (ed.)Macdonell, A. A., London, Longmans Green & Co. "The Role of Semantics in a Grammar' Unnersals in Linguistic Theory, (eds.) Bach, L. & R. T. Harms, New Yosh, Holt, Rinchart and Winston, Mandal, K. K ,(1968) Matilal, B K., (1966) Maxmuller, F., (1886) McCawley, J D., (1968) pp. 124-169. (1970) Miller, J. (1971) 'English as a VSO language, 'Langerge, 46, pp. 286-299. 'Some Types of Phrasal Conjunction in Russian', Journal of Linguistics, 7(1), pp. 55-69. Tte Descriptive Todrigue Påmiri, The Hague, Mouton. Misra, V. N., (1966) ·BIBLIOGRAPITY 279 101 Mònice-Williams, M., (1927) A Practical Grammar of the Sexskrit Language, (4th ed.), Oxford, Thc Clarendon Press. Moore, T., (1971) A Case for fener Cases', mimcographed Patañjalı (200 BC) T'fikerara-chilbapa, (cd.)Killhorn, r., Bombay, Gost, Central Hooh Depot, Vol. I, 1892. T'pukeun latinharc,(cd M. l'ardit Sinadatta D. kudala (1912), Bombay, Nirnayasagar press. I'cdersen, I , (1962) 12 Dised of Lory.p, Blooming- Uniscrsity Press, Paperback cdition of Lirguistic Source ir ile Nigenents Contry (tr. J. W Spargo), Cambridge, Mass , 1931. I'erry, C. D, (1885) Sas1 Primit, New York, Columbia Unuscossis l'ress. R20, V S. (1969) The Perlas Plyef Seriener codin Purid, Xet Delhi, Munshiram olarul. Robinson, J. J. (1971) 'Case', Caregory and Configuration, joutral of Liris, 6(1), pp 57-60. Coralrer! - VW vo S4, l'h. D. thesis, 1. T. (1969) I l'rojxxed Rule or Tree powing dins Strawy J. 17h : R 15 irailarisees! Guys.. («c!) Restxi, DIS 1. Schu, cu Jerzet, Pre-sce llall Ire *), Reprises from allshemmed Lures!!! it ad stw IT..!:. Herald Cor Laserpy, Runt.SI 1"15614.1 (199) O Delses'irrel. 71.111* HTNIM I Pent::-.

·280 SANSKRIT SYNTAX Ross, J JR, (1970) 'Gapping and the Order of Constien tuents', Progress in Linguistics, (eds) Bierwisch, M, & K E Heidolph, The Hague, Mouton, pp 249 259. Roy, S (1920) Karaka prakaranam, Calcutta. Sarup, L , (1920) The Nighantu and the Norukta, (the oldest treatise on etymology, philology and semantics), Introduction only, Oxford, University Press, pp 1 80 Saxena B R (1951) Sanskrit Vyakarana pravefikā, Allahabad, Ramnarayanalal Seuren, PAM (1969) Operators and Nucleus, Cambridge University Press Singh, J D, (1970) 'Paniai's Theory of Karahas' (mimcographed) Smith, Carlota S, (1969) 'Ambiguous Sentences with And', Modern Studies in English Readings in Transformational Grammar, New Jersey, Prentice Hall, lac , pp 75 79. Speijer, J S (1886) Sanskrit Syntaa, Leyden, EJ. Brill, also reprinted from 3rd cdition 1968, Kyoto, (Japan) The Roasen- Shoten Bookstore Ltd Staal, J F, (1967) Ford-order 11 Sansierit ard Universal Grammar, Do lrecht, D Reidel publishing Co Stockwell, RP, Integration of Transformational The- Schachter, P and ories of English Syntax, Vol I, Partces, B H (1968) II, Unis ersity of California Stratton, C R (1971) 'The Pathological Case', Working Papers in Linguistics No, 10, Ohio State University, pp 221 230 Subrahmanya Sastri, P S (1957) Lectures on Patañjali's Alabablarga, Vol V, Tiruchirapall: Subramania Iyer, K A (1969) Bbarirtars (A study of the Vilyapadiya in the light of thc Ancient Grammarians), Poona, Deccan College.

·INDEX Ablative - 6 Chomsky, N 1, 8, 12, 15 106, inflexion 125, 126 166, 167, 169, 255 Accent - 22 Cognates 74 Accusative 6, 17 Cognate objects 24, 43, 70, 71, Affix substitution 30, 32, 50, 190 74, 75, 76, 77 Agentive 16, 24, 25, 26, 34, 44, Comitative 18, 19 119, 136 Competence 13, 23 Alienable 196, 214, 216 Componential analysis 128 relation 247 Concatenation system 8,14 relationship 241 Constitutency analysis 8 Context free 255 Bach, E , 159, 168, 169 Contextual features 255, 257 Base component 14, 15, 20, 225, Coordination 116 256 Copula-deletion 167, 182 185, structure : 14 190, 267 Benefactive 18, 40, 135, 136 Copulative 145, 147, 148 Benfey, T., 5 Copying rule 72 Bennett, DC, 112 Co referential 126 Bharthari 4, 193 Counter agent 17 Bhasā 2 Covert categories 14 Bhattojídıksita 4 Body part . 228, 230, 239, 242 Datiye 6, 16, 24, 25, 26, 34, 37, relation 246 Bopp, F. .5 38, 119 inflexion 133, 136 Carey, W 4 Deepstructure : 14, 15, 19 Case 1, 9, 11, 12, 14, 16 21 Delberuch 7 categories 18, 20, 24 Delegative 135, 136 form 5, 21 270 Deletability of object 76 frame 19, 258, 259 Delerable object 76 inflexion 5, 6, 9, 16 Descriptive 12 System 11, 21 adequacy 12, 13, 15, 271 111 Categorial component 8, 20, Directional 255 indicator 241 Category feature Dynamic • 111 Category of case Centrality of syntax 13 Edgren, H 6 handas Emotive State 98 257 5 ·282 SANSKRIT SYNTAX Essive 18, 148, 149, 150, 151, Interpretation, 152, 153, 154, 157, 158, 160 accidental 109, 110 Evaluation procedure 12 purposive 109, 110 Existential 145, 147, 148 Intonational pattern 22 Experiencer 16, 24, 36, 37, 38, Intransitive verb 76 40, 44 Explanatory adequacy 4 Jones, W 4 Extent 18, 24, 111, 136, 137, 140, Karaha 9 142, 157 Kasus 16 Factitive 17 24, 43, 69, 70, 71, Katyayana 3 75, 76, 77, 68, 79, 160 Kinship 230, 239, 246 Faddegon, B 7 Kiparshy, P 7, 9, 194 Force 95 Functions of cases 8 6 Lahiri, P C. Laloff, G 106, 107, 108, 109 Genitive 6, 9, 157, 158, 192, Language classical 1, 2 195, 198, 214, 215, 239, 240, Sanskrit 1, 2, 4, 5 241, 246, 247 specific 14, 22 constructions 9 universal 14 Goal 18, 24, 40, 70, 110, 111, Lexicon 14, 20, 255, 256 130, 131, 132 Location 18, 112 Gonda ) 7 Locational 148 Locative 18, 24, 110, 111, 115, Halliday, M A K 14 116, 148, 156 Harris, J. 5 spatial 111, 113 Hjelmsley 12 temporal Hypercase 142, 143, 144 Local function 111 Locational reference 219 Inalienable 96, 115, 117, 196, I yons, J 14 214, 216, 239 relation 241, 246 Macdonell, A A relationship 128, 244 Modality 15 Inanimate force 17 attachment rule 268 Inflexion 11, 21, 270 deletion rule 190, 267 Inherent features Monier Williams, MI 6 Instrumental 6, 17, 24 79, 83 Multiple self embedding 114 91, 97, 100, 103, 107, 108, 110 adverbial 106 Natural force 95, 96, 97 Intentional involvement 42, 45, phenomena 96 99 108, 110 Node pruning rule 175, 180, 183, Intentionality 42 186 Non 42 Nominal 5 111, 113 4,6 258 ·INDEX 283 Nominal derived 197 Post Position 16, 21, 53, 58, gerundiye 197 60, 67 sentence 145 deletion 90 Nominalisation 77, 170, 177 P raising rule 175, 180, 183, 190 rule 173, 174, 179, 189, 266 Predicate 5 Nominative 6 Predicator 17 Non intentional involvement 41, Pre verb 91, 118 42, 99, 108, 110 Pre/Post Position 61, 21, 270 Non stative verb 34, 42 Proposition 15 Notion of cases 10, 11 Noun 5, 15 Ramcandra 4 phrase 15, 21 Receptive 18, 24, 34, 35, 36, 41, 46 Referential 79 Object 17, 21, 256, 270 Restrictions of co occurrence 255 case 18 Result 17, 43, 69, 70 Objective 17, 24, 43, 45, 46, 47, Resultative 78 77, 78 non 78 Observationally adequate 271 Rewriting rules 255 Ross, J R, 243 Panini 2, 3, 4, 5, 7, 9, 192, 193 Perry, F D. 6 Sanskrit 8 Part whole 228 grammar 6 relation 239, 240, 244, 246 syntax 6, 7, 9 Passivisation 25 39 syntactic structure Process of 26, 31, 36, 46, Sasthi 192 61, 64, 65, 67, 83, 90, Seşa kāraha 193 rule 267 Schema 117, 128, 132, 196, 207, without subjectivalisation 27 227, 239, 240, 241, 243, 258 27, 57, 83 Scrambling 29, 50 Patanjali 3 163 Selectional restrictions 14, 20, Path 18, 110, 111, 114, 142 160, 255, 256 P-deletion 184, 190 Sense Performance 33 accidental 108 Performative analysis 242 purposive 108 Phrase 15 13 structure 19 Sentential derisation 240, 241, structure rulc 255, 265 247 Place 112, 114 Sequence free 15 Possession, Sequential ordering 15 alienable 196, 216, 217 Sentence embedding 20 inalienable 196 217 Source 18 24, 110, 111 120, Bossesstic 214 122 123 125, 127, 128, 129 1 Sentence types ·284 SANSKRIT SYNTAX Space : 111 Time 18, 111, 112 Speijer, J. S. : 6 TiNanta : 162, 173 Staal, J. F. : 7, 8, 9, 194 Transformational generative Stative • 42 grammar : 12, 21 Strict categorisation rules : 256 rules : 266 sub-categorisation rules : 20, Translative : 18, 76, 148, 152, 153, 255 154, 158, 160 Strings of formatives : 13 Subanta : (62 Universal : 12, 22 Subject : 21, 256, 270 Utterance types : 13 Subjectivalisation : 26, 35, 39, 46, 55, 119 Vakya : 162 rule : 26, 27, 28, 29, 35, 36, Vāmana Jayāditya : 4 38, 39, 49, 58, 189, 267 Vedic language : 1 Superordinate constunction : 96, sanskrit : 1,2 98, 143, 144, 242, 270 Verb : 15, 17 Surface case : 17, 21 Vocative : 6 phenomenon : 15, 20, 21 structure : 11, 14, 15, 19 Whitney, W. D. : 5 Syntax : 12 Wilkins, C. : 5 System, Set : 8 Wilson, H. H. : 6 Systematic transcription : 23 Word-order : 8, 9, 270 ) Terminative : 112 Yasha : 3