262 43 29MB
English Pages 227 [228] Year 1989
Linguistische Arbeiten
207
Herausgegeben von Hans Altmann, Herbert E. Brekle, Hans Jürgen Heringer, Christian Rohrer, Heinz Vater und Otmar Werner
Lilo Moessner
Early Middle English Syntax
Max Niemeyer Verlag Tübingen 1989
CIP-Titelaufnahme der Deutschen Bibliothek Moessner, Lilo : Early Middle English syntax / Lilo Moessner. -Tübingen : Niemeyer, 1989 (Linguistische Arbeiten ; 207) Freiburg (Breisgau), Univ., Habil.-Schr., 1985 NE:GT ISBN 3-484-30207-0
ISSN 0344-6727
© Max Niemeyer Verlag, Tübingen 1989 Das Werk einschließlich aller seiner Teile ist urheberrechtlich geschützt. Jede Verwertung außerhalb der engen Grenzen des Urheberrechtsgesetzes ist ohne Zustimmung des Verlages unzulässig und strafbar. Das gilt insbesondere für Vervielfältigungen, Übersetzungen, Mikroverfilmungen und die Einspeicherung und Verarbeitung in elektronischen Systemen. Printed in Germany. Druck: Weihert-Druck GmbH, Darmstadt
CONTENTS
Preface
vii
Abbreviations
ix
I.
Introduction 1. Research on Middle English Syntax 2. Theoretical Framework 3. Axiomatic Functionalism 4. The Place of Syntax in Axiomatic Functionalism 5. Syntactic Relations 6. Data Base 7. Descriptive Method
1 1 4 6 11 13 20 22
II.
Compulsory Constituents of Early Middle English Clauses 1. The Active Verbal Syntagm 2. The Copulative Syntagm 3. The Passive Syntagm
24 24 42 47
III.
Nominal Syntagms 1. The Substantival Syntagm 2. The Adjectival Syntagm 3. The Pronominal Syntagm
53 53 73 78
IV.
Functional Syntagms 1. The Prepositional Syntagm 2. The Conjunctional Syntagm 3. The Comparative Syntagm 4. The Genitival Syntagm
84 84 86 90 94
V.
The Predicative Syntagm 1. Syntactic Properties of Predicative Syntagms 1.1. Passive vs. Active Verbal vs. Copulative 1 . 2 . Intransitive vs. Non-intransitive 1.3. One Predicative Complement vs. more than one Predicative Complement 1 . 4 . Transitive vs. Non-transitive 2. ICs of Predicative Syntagms 2 . 1 . Nucleus 2 . 2 . Complements 2 . 2 . 1 . Copulative Complement 2 . 2 . 2 . Objects 2 . 2 . 3 . Verbal Complement 3. Types of Predicative Syntagms 3. 1. Type ( 1 ) : Copulative, Intransitive 3. 2. Type ( 2 ) : Active Verbal, Intransitive 3. 3. Type ( 3 ) : Passive, Intransitive 3. 4. Type ( 4 ) : Copulative, Non-intransitive,
98 98 98 100
Complement
119
one Predicative Complement 3. 5. Type ( 5 ) : Copulative, Non-intransitive, more than one Predicative
104 105 108 108 108 108 109 115 116 117 118 118
119
VI
3. 6. Type ( 6 ) : Passive, Non-intransitive, one Predicative Complement 3. 7. Type ( 7 ) : Passive, Non-intransitive, more than one Predicative Complement 3. 8. Type ( 8 ) : Active Verbal, Non-intransitive, one Predicative Complement, Transitive 3. 9. Type ( 9 ) : Active Verbal, Non-intransitive, one Predicative Complement, Nontransitive 3.10. Type ( 1 0 ) : Active Verbal, Non-intransitive, more than one Predicative Complement, Transitive 3.11. Type ( 1 1 ) : Active Verbal, Non-intransitive, more than one Predicative Complement, Non-transitive
121 122 123 125 128
132
VI.
The Clause 1. Nucleus 2. Expansions 2 . 1 . Subject 2 . 2 . Adverbial Complement 2.3. Agent Complement
137 137 137 137 147 150
VII.
Functional Constructions 1. Infinitive Constructions 1.1. Infinitive Constructions without Subject 1.2. Infinitive Construction with Subject 2. Relative Constructions 3. Participial Constructions
155 155 155 157 161 168
VIII.
Conjunctive Constructions
172
IX.
Syntagms with Coordinated ICs
179
X.
Syntagms with Interordinated ICs 1. The Passive Syntagm 2. The Syntactic Type ure iwil us is lod to lete 3 . T h e Syntactic Type pe b e t ... b e b e t 1
Appendix I: Data Base Appendix II:
1. Verbs which only Combine with 'habben' in 'tense' Position 2. Verbs which only Combine with 'ben 1 in 'tense' Position
184 184 184 8 6 188 191 193
Appendix III: Descriptive Models
195
Bibliography
200
Name Index
209
Key Word Index
211
PREFACE
In the early 70s, Professor H. Pilch drew my attention to the lack of a comprehensive description of the syntactic structures of Middle English. After several years of collecting data, at a point when I was about to be overwhelmed by the sheer mass of material, I met, by a very lucky coincidence, Professor Jan Mulder, then chairman of the Linguistics Department of St. Andrews University. He introduced me to his theory of Axiomatic Functionalism. During my repeated visits to St. Andrews, he took a lively interest in the progress of my work and spent many an hour with me on the discussion of individual problems. I owe him special thanks for his unfailing support over the years and also for letting me use a draft version of his latest book. Unfortunately, Foundations of Axiomatic Functionalism appeared only after I had finished the typescript of my own book, so that the most recent developments of this theory could not be incorporated. A German version of my study was accepted as a Habilitationsschrift at the University of Freiburg in 1985. The plan of translating it into English was shaping in my mind, when Professor W. D. Bald offered to recommend my study for publication - in English - to the editors of Linguistische Arbeiten. As it is, a thorough revision of the original has resulted, requiring a reconsideration of already completed descriptions and the reorganisation of the whole subject matter. I gratefully acknowledge the generosity and patience with which Max Niemeyer Verlag put up with more and more delays. The final version of my text profited from the scrutiny of Dr. Janet Harkness and Lorna Walker, who pointed out inconsistencies and made valuable suggestions concerning the finer points of English. The present format would not have been possible without the expertise of Hans-Gunther Borrmann of Freiburg University Computing Centre. I wish to express my gratitude to all those mentioned above,
Vlll
but also to many others on whose professional and personal support I was able to count from the first draft version to the final product. Needless to say, I alone am responsible for all inaccuracies, omissions and mistakes which remain despite all precautions taken. May my shortcomings inspire others to do better and more, thus furthering our knowledge of the structure of Middle English syntax.
Freiburg, July 1989
ABBREVIATIONS
act.verb. act.verb.synt. adj. adj.synt. adv.compl. AF
agent compl. app.suppl. compl.
active verbal active verbal syntagm adjective, adjectival adjectival syntagm adverbial complement Axiomatic Functionalism, axiomatic functionalist agent complement appositive supplement
cop.compl.
compl. comparative complement comparative conjunction conjunction, conjunctional conjunctional complement copula, copulative copulative complement
dir.obj.
direct object
EME emphas. EPBS(s)
Early Middle English emphasizer Extended Predicative-Based Syntagm(s)
GCE
gen.compl. gen.synt.
Grammar of Contemporary English genitival complement genitival syntagm
IC(s) intens. intrans. ( = i n t r . )
Immediate Constituent(s) intensifier intransitive
ME mod.aux.
Middle English modal auxiliary
non-intrans.= non-intr. non-tr.
non-intransitive non-transitive
ob j . OE
object Old English
part.
participle, participial participial complement past participle present participle
comp.compl. comp.conj.
con j .
conj.compl.
cop.
part.compl. part.past ( = p . p . ) part.pres.
pass, pass.cop. PBS(s) PDE pred. pred.nucl. pred.synt. prep, prep.obj. prep.suppl. prep.synt. pron. pron.synt.
passive passive copula Predicative-Based Syntagm(s) Present-Day English predicate, predicative predicative nucleus predicative syntagm preposition, prepositional prepositional object prepositional supplement prepostional syntagm pronoun, pronominal pronominal syntagm
relat. rel.compl. rel.constr. rel.suppl.
relativizer relative complement relative construction relative supplement
subj. subst. subst.synt. suppl.
subject substantive, substantival substantival syntagm supplement
trans. ( = t r . )
transitive
verbal compl.= verb.compl.
verbal complement
I.
INTRODUCTION
T . Research on ME Syntax ME syntax is usually considered a branch of Historical Linguistics. The designation 'historical 1 is ambiguous here. It may refer to a previous stage of a language or to its development in time. In the latter sense it is synonymous with 'diachronic', and this is the prevailing approach of traditional linguistics towards earlier stages of a language. Traditional Historical Linguistics was mainly concerned with the development of the distinctive elements of languages. Therefore many so-called historical grammars are nothing but diachronic phonological studies. Under these conditions it is not surprising that most ME grammars and course-books do not contain a chapter on syntax at all (Bahr, Berndt, Brunner, Fisiak 1964, Fisiak 1968, Jordan, Wardale, Weinstock, Wright). Chapter 14 of M. Roseborough's grammar (1938) is entitled "Accidence:Syntax", and there we find scattered remarks about the obligatory (sic!) status of the subject in ME, the frequency of impersonal constructions, special features of number concord, etc. Mosse apart, we have to wait until the 1970s before more extensive and systematic treatments of ME syntax appear (Dürmüller/Utz 1974, Jones 1972, Moessner/Schaefer 1974; ^ 1 9 8 7 ) . But they , too, are rather specialized, covering primarily syntactic structures attested in the texts which these books include. Individual problems of ME syntax were investigated in detail especially at the beginning of our century (Böhme 1903:Temporal Clauses; v.d.Gaaf 1904:Impersonal Constructions; Gebhardt 1922: Subject Omission). More recent studies deal with wordorder (Harris 1964? McLeish 1969; Swieczkowski 1 9 6 2 ) , pronominal syntax (Rennhard 1962; Suter 1955), relative clauses (Kivimaa 1966), modal verbs (Wallum 1973), and the syntactic functions of the elements
so, al so, as (Nuiranenmaa 1973) and one (Rissanen 1967). Most syntactic descriptions of individual ME texts are also outdated. Only very few texts have been investigated and described syntactically with modern linguistic methods: 'Ancrene Riwle' (Humbert 1 9 4 4 ) , Ormulum' (Palmatier 1969), 'Peterborough Chronicle 1 (Shores 1971), 'Cursor Mundi' (Snouffer 1966). Due to their outstanding literary quality Chaucer's works have also received special attention from linguists (a.o. Fries 1985; Kerkhof 2-|982; Sandved 1985). One book only promises a survey of the whole field of ME syntax: Mustanoja's A Middle English Syntax. But the preface checks any rash expectation, toning down the promise of the title: "It [=the book] deals only with constructions likely to prove difficult to the literary student or of particular interest otherwise." (Mustanoja 1960:5) The Table of Contents shows that Mustanoja's approach is traditional, and rather morphological than syntactic. His subject matter is arranged according to the traditional word-classes: substantive, pronoun, article, adjective, numeral, adverb, preposition, verb. Syntactic constructions are treated under morphological categories, e.g. the group genitive under 'case', verbal syntagms of the type mai cumen under the heading 'mood', etc. This explains why we look in vain for a chapter on clause types or complex sentence constructions. The second volume, which will deal with these, has unfortunately not yet appeared. The most exhaustive- diachronic treatment of English syntax is without doubt Visser's monumental work An Historical Syntax of the English Language. One of the most valuable aspects of the book is the vast amount of data which it provides. Many authors of recent studies on historical English syntax gratefully acknowledge their indebtedness to this inexhaustible data source (Traugott 1972; Faiß 1 9 7 7 ) . Unfortunately it takes quite a bit of effort to find the data relevant to a given topic, because Visser's arrangement of his material is based on the number of verb forms which a construction contains. Thus the first part of the book deals with "syntactical units with one verb", the second part with "syntactical units with two verbs" and the third with "syntactical units with two and more verbs". The author freely admits the "purely
formal and mechanistic" character of this method. It has the disconcerting effect that syntactically similar constructions like he is rich and you are being naughty are not treated together, but in different sections of the book (§241:§1834), whilst syntactically very different constructions like the old woman knew well to babble and do, my dear love, write to me are treated in the same volume (§§1185, 1 4 2 8 ) , because both contain two verb forms (knew, babble; do, write) the second of which is an infinitive. Any diachronic study presupposes a series of synchronic studies of different stages of the language under consideration. Visser describes the ideal working conditions for a diachronic grammarian as follows: "Ideally speaking, the historical grammarian should have at his disposal a complete series of exhaustive synchronic descriptions of the state of the language from one generation to another." (Visser I:v) He is well aware that we are still far from this goal with regard to English. An exhaustive description of the syntactic structure of PDE is provided in A Grammar of Contemporary English and Λ Comprehensive Grammar of Contemporary English. OE at the other end of the scale has also received increased attention. Pilch included a chapter on syntax in his Altenglische Grammatik ( 1 9 7 0 ) , McLaughlin published a booklet with the title Old English Syntax. A Handbook ( 1 9 8 3 ) , and 1985 saw the publication of what will remain the standard reference work on the subject for many years to come, Bruce Mitchell's Old English Syntax. However, apart from Mustanoj a ' s book there is a huge gap in the description of English syntax between OE on the one and PDE on the other hand. As it is a well-known fact that every synchronic description implies a certain amount of diachrony - the more so when a historical state of a language is concerned - it was necessary to decide which time span should be covered. Barbara Strang's remark was my guide-line here: "Some developments will not be clearly established within a century. Doubling this span gives a meaningful stretch of time, but one not too complex to account for in a reasonably unified way." (Strang 1 9 7 4 : 2 1 f . ) Thus I restricted my investigation to texts of the 12th and 13th
centuries, the period usually referred to as Early Middle English (EME).
2. Theoretical Framework Language data can be interpreted in a variety of ways by different scientific disciplines, e.g. history, literary criticism, psychology, sociology, and last but not least linguistics. We do not expect that such diverse investigations will yield the same results, although they may be based on the same data. Each discipline is interested in different aspects of the same evidence; different solutions are sought to different problems. These solutions have the status of scientific hypotheses, which can be corroborated or falsified by the data. These hypotheses are necessarily different because of the different underlying theories. Linguistic descriptions presuppose a linguistic theory. As there are several competitive linguistic theories, a careful choice must be made, because the quality of a description depends partly upon the quality of the theory behind it. When choosing a theory, two considerations are especially important. Does it meet with the epistemological requirements of consistency, adequacy and simplicity and, is it appropriate for the set of objects which we want to describe? If a theory a'llows descriptive statements of the form "a is identical with b" and "a is not identical with b", it has to be rejected because it is inconsistent. If a theory excludes the description of a subset of our data, it has to be rejected because it is not adequate for our purpose. If two competitive theories rank equally high in consistency and adequacy, the simpler theory is to be preferred. Of course, it is possible to modify theories. The temptation to do so is particularly great when they lack adequacy. Modifications of one part of a theory may, however, entail undesirable consequences for other parts, consequences which may easily be overlooked. In extreme cases such modifications may result in a completely useless theory: "Many, however, if not most are too impatient to acquire the proper experience, or are not self-critical enough to see that
their skill as a descriptivist is still underdeveloped. The latter ones are the most dangerous, as it is they that ultimately may cause the decline and downfall of the theory in question. When they come across a problem which they are unable to solve, or they notice inconsistencies in their descriptions, their inclination is to blame it on the theory, rather than on their own incompetence. Then they start tinkering with the theory - but someone who is not even a good descriptivist can hardly be expected to be a good theory-builder. Even less than those persons are able to oversee all the consequences of their descriptive decisions, are they capable of foreseeing the descriptive consequences of changes in, or additions to, the theory. This is how theories usually die." (Mulder 1988:124 [typescript] ) Mulder touches upon a basic problem here. Very few linguists are trained epistemologists as well. Linguists with a theoretical bias very often lack experience in handling empirical data. Descriptivists on the other hand all too willingly claim a theory's superiority over other theories without attempting to demonstrate this alleged superiority.^ Others simply state that for their descriptive purposes no linguistic theory is available, thus justifying a highly eclectic combination of different theories: "With regard to the present study then, the conclusion is that no descriptive framework currently available could adequately accomodate the different grammatical aspects which appear to be crucial features of the grammar of subjectless predicates." (Elmer 1981:11) Such eclecticism reflects yet another predicament. One theory may be perfectly adequate for the description of one set of linguistic data or for one structural level, but rather clumsy for the description of another set of data or another structural level, whereas the reverse may be true for a second theory. Which theory is then better? These considerations show how dangerous it is at the present stage of development in linguistics to make statements about the superiority of one linguistic theory over another, let alone over all others. This does not mean that descriptivists are right in adopting one theory for one purpose and a different theory for an1
cf. Fowler, R. An Introduction to Transformational Syntax, London 1971, p.vii: "its [=TG's] general framework has become accepted by the majority of western linguists as providing the most reliable and revealing version of linguistic analysis."
other, or even combining theories at will. In this way we will never arrive at a consistent description of all structural levels of one or more languages or of one structural level at different periods of one or more languages. It is quite legitimate that descriptivists should choose a theory which they consider appropriate for their purpose. As long as they stick to one consistent theory, it is likely that they will produce a consistent description. Such descriptions may be challenged by others, on the basis of different theories. If they are more adequate and/or simpler, they are to be preferred. But even then this would not mean that the first theory must be discarded, because it might still prove better at the end of the day when all structural levels of one or more languages have been described. It is therefore with no claim as to the superiority of the theory of Axiomatic Functionalism (AF) that I chose this theory as the framework of my description, although I believe that this theory meets the epistemological requirements mentioned above and allows a consistent, adequate and relatively simple description of EME syntax. If there are shortcomings in the description which must be blamed on the theory, competing descriptions based on other theories will certainly be put forward. Even if these are better, it must still be proved that their underlying theories also allow better descriptions of other sets of linguistic data and/or other structural levels. To help in the evaluation of the following description I shall outline the structure of AF, concentrating on those parts of the theory which are relevant to syntax.
3. Axiomatic Functionalism Axiomatic Functionalism is a linguistic theory which was developed by Jan Mulder, now professor emeritus of the University of St. Andrews, Scotland. It is in the tradition of two well-known European linguistic schools: Prague School Linguistics and Glossematics. Mulder acknowledges his indebtedness to both schools explicitly: "The theory of functionalism has always struck me as extremely sound in principle, scoring higher than any other school of
thought in those points [consistency, adequacy, simplicity]. An exception may, perhaps, be Hjelmlev's theory of Glossematics, which is very well organised, very explicit, reasonably simple, and agreeably consistent." (Mulder 1980:28) It
was mainly through the person of Andre Martinet that Prague School Linguistics influenced Jan Mulder. 2 The function after which Martinet's linguistic approach was named is the communicative function: "La fonction essentielle de cet instrument qu'est une langue est celle de communication." (Martinet 1980:§1.4). Although functionalists are well aware of the existence of other functions of language as well, they consider them to be of secondary importance. A functionalist linguistic description contains only those elements which are relevant for the act of communication. The functional aspect is thus the point of view from which the object of linguistics is investigated. This object itself is defined as "those communication systems that have a vocal character and are in general use among all human communities" (Martinet 1975:12). Hjelmslev's influence on AF can be seen in the terminology: Ceneme, plereme, etc. have been taken over from Glossematics. The Glossematic relations of interdependency, determination and constellation correspond to the AF relations of interordination, subordination and coordination. More important than terminological similarities is the adoption of Glossematic methodology. This methodology makes possible a high degree of stringency in the presentation of AF theory in the form of an algebraic calculus.^ The AF requirement that a theory must be arbitrary and appropriate also goes back to Hjelmslev's Glossematics. 2
"Roman Jakobson und Andre Martinet kann man in der internationalen Sprachwissenschaft als die beiden bedeutendsten Erben der Prager Schule ansehen. Martinet betont ausdrücklich die 'funktionale' Orientierung seiner eigenen Vorstellungen." (Lepschy 1969:85) [In international linguistics Roman Jakobson and Andre Martinet can be called the most important heirs of Prague School Linguistics. Andre Martinet explicitly stresses the 'functional' point of view of his own approach.] 3
cf. Mulder 1980:chapt.4
In his theory of AF Mulder distinguishes three parts or subtheories: a signum-theory (originally called sign-theory), which defines the ontological status of linguistic objects; a systemtheory (or systemology), which covers phonology and grammar; a theory of semantics.4 Phonology is subdivided into phonematics, phonotactics, and paraphonotactics, grammar into plerematics (=morphology), plerotactics (=syntax), and paraplerotactics (=parasyntax). This means that in the present context only the systemtheory is relevant. Before describing the structure of the theory I would like to stress a feature which makes AF unique among the various kinds of present-day linguistics. I am referring to the rigorous separation of theory and description. This may seem trivial, but none of the other current approaches to linguistics seem to bother about such trivialities. Consequently their descriptions cannot be properly evaluated, because the criteria on which an evaluation could be based are not specified. Descriptions which are based on the theory of AF can be evaluated against the theory and against the data, because they are viewed as "the application of a particular linguistic theory to a selected field of linguistic phenomena" (Mulder 1980:17). The theory of AF contains terms and statements. The statements are of three different kinds. Axioms, that kind of statements which gave AF its name, are neither true nor false; by definition they are not subject to refutation. They have to be accepted as starting-points of deductive theories, cf. "The initial justification of these statements is that they seem reasonable and acceptable to others, and their further justification is that they are assumed, in the absence of refutation, to be appropriate." (Mulder 1980:23) Axioms imply theorems, the second kind of statements. The terms which are used in axioms are variables, i.e. terms which have no meanings at all at the outset, but which receive meanings through definitions, the third kind of statements. Then they are no longer variables, but defined terms. While attributing meanings to vari4
This part of the theory was developed by Mulder's collegue Sändor Hervey, cf. Hervey, Sändor. Axiomatic Semantics, Edinburgh 1979.
ables, definitions may introduce new variables into the theory. The process of defining goes on until all variables have become defined terms. Without circularity this is only possible in a theory which also contains primitive terms, i.e. terms which need not be defined, because they are understood intuitively. Critics of AF may indeed find a weak point of the theory here. The inclusion of primitive terms as a third kind of terms (beside defined terms and variables) introduces a certain amount of subjectivity into the theory. Mulder is aware of this problem, and that is why he makes every effort to keep the subjective, and therefore intuitive, component of his theory as small as possible, cf. "By pushing this intuitive interpretation as far back as possible, we are removing much of its harmfulness." (Mulder 1980: 25). Nevertheless he considers a subjective residue unavoidable in any theory which is not circular: "We may still take this state of affairs to be, scientifically speaking, deplorable, but we have to live with it. Our system of definitions will either be circular, or it will contain this intuitive residue of primitive terms, just as a theory that does not contain primitive, i.e. intuitive statements, namely the axioms, will always be a circular theory." (Mulder 1980:25) The structure of AF as sketched above is represented in the following diagram:
terms
statements
primitive terms defined terms axioms theorems variables {. meanings — .definitions Some definitions assign meanings to variables which denote processes, others to variables which denote entities. The latter correspond to descriptive entities which stand in a relation of isomorphism to classes of speech phenomena. The descriptive entities are called models; the corresponding theoretical terms are called theoretical models or meta-models.
10
Like any scientific theory AF lays claim to the attributes consistency, adequacy, and simplicity. Being hypotheses, these requirements cannot be part of the theory itself. They have the status of meta-hypotheses because they are hypotheses about the theory. Descriptions which are based on the theory of AF neither contain axioms nor definitions. The statements they contain have the status of scientific hypotheses. One of their most important properties is that they must be refutable by the speech phenomena they are meant to describe. Descriptive statements are about descriptive models, their constituents, and their relations. They, too, must meet the requirements of consistency, adequacy, and simplicity. Descriptions are inconsistent not only when their hypotheses contradict each other, but also when they are not interrelated. Inconsistency of a description may imply inadequacy, because speech phenomena cannot adequately be described by contradictory hypotheses. Irrespective of other qualities, hypotheses have to be rejected or at least modified if they are refuted by the speech phenomena. Hypotheses about the grammatical structure of the data must also be checked against their meaning, c f . "grammatical statements ... have to be consistent with what we know to be ... the meaning of the construction in question" (Mulder 1980:13). Some statements of descriptions resemble the definitions in the theory. This similarity is only superficial; it does not mean that these descriptive statements are definitions. They introduce useful mnemonic labels (e.g. 'subject', 'transitive', 'prepositional syntagm', etc.) for language-specific constructions. The methodology used in AF can be characterized as hypothetico-deductive. AF descriptions are deductive insofar as their statements are derived from the theory (of AF) on the one, and from the data on the other hand. They are hypothetical because their statements are presented as hypotheses. It is especially the deductive component of the methodology which Mulder considers important. He rejects any kind of speculation and inductivism, speculation being non-scientific and inductivism leading to circularity. As inductivism exclusively relies on the observation of data, its statements - however general and
11
abstract they may be - must remain part of the description. Only by an arbitrary act can they be made statements of the theory. Such a theory is useless because its application to a field of speech phenomena yields exactly the descriptive statements which the theory started with.-* The underlying theory must therefore also be based on a deductive methodology. It follows that it will contain a limited set of primitive terms and statements from which all other statements can be derived by logical operations. It is independent of any specific language, but it provides the framework for the description of any set of language data.
4. The Place of Syntax in Axiomatic Functionalism The structure of the system-theory of AF is based on Martinet's tenet of the double articulation of language. That part which deals with the distinctive elements is called phonology, the other part, which deals with the significative elements is grammar. For Martinet it is this double articulation which accounts for the economy and the efficiency of language. Mulder, however, argues that these qualities also depend on the fact that on both levels the arrangement of the elements may be relevant, i.e. they may stand in ordering relations. The English word sick, for example, can be analysed on the second level of articulation as a sequence of the phonemes /s/, /i/,. /k/. That the arrangement of these phonemes is relevant can be demonstrated by reversing their order. The language sign kiss ( / k i s / ) has a very different meaning from the sign sick ( / s j k / } . On the first level of articulation the relevance of the arrangement of elements can be demonstrated by the English syntagm dog show. The permutation of its constituents yields a syntagm with a completely different meaning, show dog. The possibility of ordering relations leads Mulder to a redefinition of the principle of the double articulation of language :
5
cf. Mulder 1980:9f.
12
"As, however, the possibility of ordered combinations, as opposed to the possibility of only simultaneous, i.e. unordered, ones, exhibits the ultimate level of efficiency, I have defined articulation so as to refer to syntax and phonotactics only." (Mulder 1980:147) Consequently, both the phonology and the grammar part of AF contain a component which deals with the respective elements as simultaneous bundles of constituents, and a component which deals with the combination of the respective elements into bigger units. Phonematics treats phonemes as simultaneous bundles of distinctive features, phonotactics deals with their combination into phonotagms (=ordered sets of phonemes). Plerematics is about pleremes as simultaneous bundles of monemes,^ plerotactics is concerned with the combination of pleremes into syntagms.^ The notions 'plerematics1 and 'plerotactics 1 roughly correspond to the traditional terms 'morphology 1 and 'syntax'. As a matter of fact Mulder himself very often uses the traditional terminology, although his definition excludes for example the treatment of compounds in morphology/plerematics. I am going to use the traditional term 'syntax' with the definition given for 'plerotactics'. The relation between the '-atics 1 and the '-tactics' components is that the biggest entities of the first components are at the same time the smallest entities of the second components. Phonemes are on the border-line between phonematics and phonotactics, pleremes on that between plerematics and plerotactics. Phonology and grammar contain a third component, the "para' component. Paraphonotactics accounts for the co-existence of phonotagms and suprasegmental features, paraplerotactics (or parasyntax) for the co-existence of syntagms and suprasegmental features. The relation between the 'tactic' and the 'paratactic' levels is different from that between the 'atic 1 and the 'tactic' levels because the output of the 'tactic' levels, phonotagms and syntagms, furnishes only part of the input of the paratactic levIn functionalist terminology monemes are the smallest nificative elements, cf. Martinet 1960:§1.9. 7
sig-
0ne would expect the combinations of pleremes to be called plerotagms to keep up the terminological parallelism between phonology and grammar. Mulder avoids coining new terms when it is not necessary.
13
els. The place of syntax in the system-theory of AF is represented in the following diagram: system-theory phonology *"
\
phonematics phonotactics paraphonotactics
grammar plerematics plerotactics (syntax) paraplerotactics
5. Syntactic Relations In AF syntagms are analysed into their immediate constituents (=ICs) step by step. The relations which hold between ICs are tactic relations. 8 The syntagm many very old books contains the ICs many, very old and books. They stand in tactic relations to each other. The 1C books can be substituted by elements like houses, hats, dogs, etc. These substitutions do not change the syntactic structure of the syntagm. The substitutable elements stand in the paradigmatic relation of commutation.^ The descriptive models of AF are defined by their ICs and the relations which hold between them. Syntactic relations are therefore very important terms of the syntactic component of the theory. Syntactic relations are a subset of grammatical relations. Some of them hold between elements of the same syntagm (=syntagmatic relations), others between elements of different syntagms (=paradigmatic relations). The only relevant paradigmatic relation is commutation. It holds between ICs which are substitutable for each other in a syntagm. In the syntagm he killed the lion the 1C he commutes with syntagms like my grandfather, Charles, or the extravagant fellow who told everybody he was a vegetarian. Syntagmatic syntactic relations are either tactic or nontactic, either direct or indirect. Tactic relations hold between 8
cf. Mulder 1980:45, def.7c 3
9
cf. Mulder 1980:44, def. 7a2
14
ICs; syntagmatic relations between constituents that are not ICs are non-tactic. Direct relations are those which are not mediated by other constituents.^ The features tactic or non-tactic and direct or indirect are properties of all syntagmatic relations. Therefore syntagmatic relations can be tactic and direct, tactic and indirect, non-tactic and direct, or non-tactic and indirect. The following diagram maps the system of syntactic relations: syntactic relations
^
syntagmatic \
paradigmatic
direct
commutation
/ \
tactic non-tactic
indirect
7\
tactic non-tactic
In the syntagm many very old books direct tactic relations hold between many and books, very old and books, very and old. There is one direct, non-tactic relation, namely between old and books. It is direct because it is not established via another constituent, and it is non-tactic because the constituents old and books are not ICs of the same syntagm; books is an 1C of the syntagm many very old books, old is not, whereas old is an 1C of the syntagm very old to which books does not belong at all. The relation between many and very old is indirect and tactic; many and very old are ICs of the syntagm many very old books, their relation is mediated by the constituent books. Indirect, non-tactic relations hold between many and old, many and very, very and books. In the first two examples the relation is via books, in the last via old. The constituents are not ICs of the same syntagm. It is quite obvious that indirect, non-tactic relations are irrelevant for syntactic analysis. The structure of a syntagm is best revealed by its tactic relations, especially by its direct tactic relations. They are either symmetrical or asymmetrical. Symmetrical relations are called relations of simultaneity, asymmetrical ones ordering relations.^ Subordination and superor10
cf. Mulder 1980:52, def.15
11
cf. Mulder 1980:44, defs. 6a, 6b.
15
dination are asymmetrical, coordination and interordination symmetrical relations. Subordination holds between ICs of unilateral functional dependency, i.e. the function of one 1C is dependent on that of the other, but not vice versa. ^ The functionally dependent 1C is subordinated to the functionally independent 1C. This implies, of course, that the functionally independent 1C is superordinated to the dependent 1C. Usually the relation is stated in terms of subordination. In the syntagm the books the definite article the is subordinated to the substantive books. The function of books is independent of that of the definite article; the function of the definite article, however, depends on the function of books. In the environment ...are expensive the substantive books has the function of a subject whether it is accompanied by the definite article or not, cf. the books are expensive, books are expensive. The function of the can be described as subject modifier. In the environment he bought ... the substantive books has the function of an object whether it is accompanied by the or not, cf. he bought books, he bought the books. The function of the definite article can be described as object modifier in this context. In AF the function of the ICs of a syntagm is described in terms of positions in a descriptive model. Functionally independent ICs are said to stand in nuclear position, they are called 'nuclei'. 1 3 Functionally dependent ICs stand in peripheral positions. If they commute with zero, they are called 'expansions', otherwise they are called 'bound entities' or 'actualisers'.^ jn the syntagm the books the definite article is an expansion, books is the nucleus. In AF the relation subordination is symbolised by an arrow with its head pointing towards the nucleus, e.g. the > books Interordination is a symmetrical relation between ICs which mutually depend on each other for their function.^ In the syntagm 12
cf. Mulder 1980:48, def. 11a
13
cf. Mulder 1980:50, def. 13a
14
cf. Mulder 1980:50, defs. 13b-d
15
cf. Mulder 1980:49, def. 11c
16
the more the better the ICs the more and the better are interordinated. The function of the more depends on the function of the better and vice versa. Syntagrns with ICs in interordination are also known as correlative constructions. The relation of interordination is symbolised by a double-headed arrrow, e.g. the more < > the better As none of the ICs is functionally independent, both occupy a peripheral position. The relation of coordination holds between ICs which are functionally independent of each other.^ In the syntagm old expensive books both adjectives modify the nucleus books. The function of old does not depend on the function of expensive, and vice versa. Therefore old and expensive are coordinated. The relation of coordination is symbolized by a crossed double-headed arrow, e.g. old < > expensive The terms subordination and coordination in particular are wide-spread in modern linguistics. Therefore it is important to stress that in AF these notions are unambiguously defined and consistently applied. Their meanings do not necessarily coincide with the meanings which other linguistic schools give them. A striking example of possible terminological confusion and of inadequate and inconsistent description at the same time is the analysis of syntagms like boys and girls as coordinated constructions. The ICs of the syntagm are said to be boys, and, girls with the relation of coordination existing between the ICs boys and girls. This relation is established by the so-called coordinator and. This analysis is not only inadequate, but also inconsistent with the underlying definition of coordination which requires that the ICs of a coordinated construction have the same distribution as the construction itself. 17 The 1C and has certainly not the 16 17
cf. Mulder 1980:48, d e f . l l b
cf. Bloomfield, L. Language, New York 1933, repr. London 1969, chapt. 12.10: "Endocentric constructions are of two kinds, cp-ordinative ( . . . ) and subordinative ( . . . ) . In the former type the resultant phrase belongs to the same form-class as two or more of the constituents. Thus, the phrase boys and girls belongs to the same form-class as the constituents, boys, girls; these con-
17
same distribution as boys and girls. Two considerations are important for the analysis of the syntagm according to AF principles. Firstly, the relation between boys and girls is not direct, but via and. Secondly, boys has the same distribution as the whole syntagm; the ICs of the syntagm are therefore boys and and girls. 18 This means that the relation between boys and girls is not tactic either. The function of the 1C boys is independent of that of and girls, but not vice versa. The direct, tactic relation between the ICs of the syntagm boys and girls is therefore not coordination, but subordination with boys as nucleus and and girls as expansion. All linguists will probably agree that the ICs the and books of the syntagm the books stand in a relation of subordination. According to AF the same relation holds between the ICs the and book of the syntagm the book. In both syntagms there is unilateral functional dependency with the substantive in nuclear and the definite article in peripheral position. If syntactic relations are defined, however, via paradigmatic classes, 1 ' the analysis of the singular syntagm as a subordination construction is no longer justified. Whereas the elements books and the books belong to the same distributionclass, book and the book do not, cf. the books are expensive, books are expensive; I bought the books, I bought books; the book is expensive, »book is expensive; I bought the book, *I bought book. Bloomfield himself was aware of this probstituents are the members of the co-ordination, and the other constituent is the co-ordinator." Lyons, J. Introduction to Theoretical Linguistics, Cambridge 1971, chapt. 6 . 4 . 2 . : "Endocentric constructions fall into main types: co-ordinating and subordinating. Co-ordinating constructions have thesame distribution as each of their constituents taken separately." Pilch, H. Altenglische Grammatik, München 1970, §39.3: "Gehören a, b und ab alle drei zu einer gegebenen paradigmatischen Klasse, so eißt "Hie Relation zwischen ihnen eine Koordination." [If a, b, and ab belong to the same paradigmatic class, the relation Between tRem is called coordination.] 18
At first sight the grouping boys and and girls seems equally adequate. For arguments against this analysis c f . Mulder 1980:168. 19
cf. Bloomfield 1933:chapt.12.10: "In subordinative endocentric constructions, the resultant phrase belongs to the same form-class as one of the constituents, which we call the head."
18
lern, but he dismissed it as "minor".20 ^s followers do not discuss it at all, but choose their examples with great care so that problems do not arise. In AF problems of this kind cannot occur because syntactic relations are not defined in terms of occurrence dependency, but in terms of functional dependency. Occurrence dependency provides, however, a useful test in syntactic analysis, because there are regular correspondences between certain types of occurrence dependency and certain types of functional dependency. As with functional dependency three types of occurrence dependency can be distinguished: bilateral occurrence independency, bilateral occurrence dependency, and unilateral occurrence dependency.21 Unilateral occurrence dependency is exemplified by the ICs of the syntagm the books; the occurrence of books is independent of the occurrence of the definite article, but not vice versa.22 Bilateral occurrence dependency holds between the ICs of the syntagm the book; neither of them has the same distribution as the whole syntagm.23 The jcs the more and the better of the syntagm the more the better stand also in a relation of mutual occurrence dependency; neither of them can occur in the same environment as the whole syntagm. In the syntagm old expensive books the ICs old and expensive are mutually occurrence-independent. They belong to the same distribution class as the 1C old expensive of the syntagm old expensive books. It follows that there is a one-to-one relation between bilateral occurrence independency and coordination, and a one-to-one relation between uni20
cf. Bloomfield 1933:chapt.12.10: "There may be minor differences of form-class between the resultant phrase and the members." 21
cf. Mulder 1980:49f., defs. 12 a-c
22por Bloomfield unilateral occurrence dependency is the defining criterion for the relation subordination, cf. fn. 19. 2%ilateral occurrence dependency is the defining criterion for Bloomfield's exocentric relation: "The resultant phrase may belong to a form-class other than that of any constituent. For instance, John ran is neither a nominative expression (like John) nor a finite verb expression (like r a n ) . Therefore we say that the English actor-action construction is exocentric; the resultant phrase belongs to the form-class of noimmediate constituent." (Bloomfield 1933:chapt.12.10) That this definition is not consistently applied in descriptions was shown before, cf. p.19f.
19
lateral occurrence dependency and subordination. If two ICs are mutually occurrence-dependent, the relation between them is either subordination or interordination. In the syntagm the book there is mutual occurrence dependency. The statement that the definite article is subordinated to the substantive is based on the hypothesis that the relation between the and book is the same as that between the and books, which is a clear case of subordination. This hypothesis is valid until it is refuted. One nucleus can be expanded by more than one peripheral element. The direct tactic relations in such syntagms are complex. Two complex relations are to be distinguished: disjunctive or diverse determination and conjunctive or parallel determination. The complex relation in a subordinative syntagm with more than one peripheral element is called diverse if the peripheral elements determine the nucleus in significantly different ways.^4 in the syntagm the hunter killed the lion the subject the hunter and the object the lion determine the verb, and they do so in significantly different ways, c f . the hunter killed the lion ^ the lion killed the hunter. The complex relation in this syntagm is therefore diverse determination. If a nucleus c is expanded by the peripheral elements a and b and the complex relation between these ICs is diverse determination, it is symbolized as -> c
If the elements a and b determine a nucleus c and it cannot be ascertained that they do so in significantly different ways, the complex relation is called parallel determination. ·> It is symbolized as a -> c b
In the syntagm many very old books the nucleus books is expanded by many and by very old. It may be - and indeed it is very proba24
c f . Mulder 1980:50, def. 14a
25
cf. Mulder 1980:50f., def. 14b
20
ble - that the relation between many and books is not exactly the same as that between very old and books, but the difference cannot be established on functional grounds. Therefore the hypothesis is that the complex tactic relation in the syntagm many very old books is parallel determination. The expansions many and very old are in different peripheral positions, cf. many very old
-> books
If two elements exclude each other as determiners of the same nucleus, they realize the same position, if not they realize different positions. The plereme many excludes the plereme several as a determiner of books; they realize the same position, which may be called 'quantifier'. On the other hand, many does not exclude elements like his or my uncle's, cf. his many very old books, my_ uncle's many very old books. Therefore his and my uncle's (which exclude each other) realize a different peripheral position from many. The only possibility for two syntagms not to exclude each other and yet to be in the same position is if they are coordinated, e.g. many very old extremely expensive books. The syntagms very old and extremely expensive are in the same position because they are coordinated, cf. many (very
> old)
(extremely
> expensive)
-> books
6. Data Base The title Early Middle English Syntax implies that the descriptive models cover the syntactic structure of all EME texts which have come down to us, even of those which are not yet known, but may be discovered in the future. The material adequacy of this claim may be tested by applying the models to randomly chosen syntagms from randomly chosen texts of the period. The models, however, do not presuppose an exhaustive analysis of all extant texts.
21
They are generative to the same degree as models of a living language are, the only difference being that the amount of available data against which descriptive models can be checked is limited in a dead language, whereas it is unlimited in a living language. Prior analysis of all data is not only unnecessary, but it would also be very time-consuming. In his Manual of the Writings in Middle English Wells lists more than ΘΟ texts of the EME period. I chose a little more than one third of these (33 texts) as the data base for my study. This main body of data is supplemented by references from ME dictionaries and relevant special studies. Heterogeneity was my main concern in the selection of the texts. They vary in literary genre, dialect, size, source and form. The data base includes romances (e.g. King Horn, Havelok) and chronicles (e.g. Peterborough Chronicle), sermons (e.g. Old Kentish Sermons) and debates (e.g. The Owl and the Nightingale). The shortest text consists of 8 lines only (Will and Wit), the longest runs to over 30000 lines (Layamon's Brut). Some texts are are translations or adaptations of foreign originals which they may or may not acknowledge. Layamon, for example, indicates his sources at the very beginning of his text: "he nom pa Englisca boc pa makede Seint Beda an ober he nom on Latin pe makede Seinte Albin & the feire Austin be fulluht broute hider in boc he nom be pridde leide per amidden pa makede a Frenches clerk Wace wes ihoten" (LB 3 1 f f . ) The books he mentions can be identified as the Latin original and the OE translation of Bede's Historia Ecclesiastica Gentis Anglorum as well as Wace's Brut. The ME Bestiary is an English version of the well-known Latin text by Theobaldus. Other texts are original works for which no sources are known (e.g. The Proverbs of Alfred, Poema Morale). Reflecting the proportions in which verse and prose texts have come down to us, the data base contains fewer examples of the latter (e.g. The Peterborough Chronicle, Old Kentish Sermons, The Bodley Homilies, The Shires and Hundreds of England) . Regional variation is more amply illustrated on the eastwest axis than on the north-south axis; most texts belong to the Midlands and the South. For all texts I have used editions of the earliest MS available.
22
7. Descriptive Method The models of a linguistic description form a system with complex relations between its members. It is difficult to present the models in a convenient arrangement because the description of the realization possibilities of each model presupposes some knowledge about the structure of other models. The description of the model of the substantival syntagm, for example, requires some understanding of the prepositional syntagm and vice versa, because a substantival syntagm like those very old books on the green shelf contains the prepositional syntagm on the green shelf as one of its ICs, and the prepositional syntagm on the green shelf contains the substantival syntagm the green shelf as one of its ICs. Although references to later chapters are unavoidable, the arrangement chosen here restricts these to a necessary minimum. The presentation begins with the models of those syntagms which realize compulsory constituents of EME clauses. The chapter which follows deals with nominal syntagms, i.e. syntagms with substantival, adjectival, and pronominal nuclei. The common feature of the syntagms described in Chapter IV. ('Functional Syntagms') is that their nuclei belong to closed classes. The models in Chapter V. ('The Predicative Syntagm 1 ) are a necessary prerequisite for the description of EME clause types which follows. The heading 'Functional Constructions' for Chapter VII. is motivated by the fact that the nuclei of these syntagms resemble those of 'functional syntagms 1 , but the ICs in the peripheral positions are either realized by or contain predicative syntagms. In contrast to earlier chapters, Chapters VIII. and IX. introduce no new models and instead present syntagms characterized either by special syntactic properties of their ICs ('Conjunctive Constructions') or by special syntactic relations between their ICs ('Syntagms with Coordinated I C s ' ) . The last chapter with the title 'Syntagms with Interordinated ICs' takes up the passive syntagm again and treats two constructions corresponding to the PDE types he is easy to please and the more - the better. The descriptions of all the models presented are accompanied by several examples which illustrate some of the realization possibilities. Examples from verse texts are referenced by line num-
23
her, those from prose texts by page and line number. All examples are translated into PDE (preceded by " = " ) . Where helpful to understand the syntactic structure of the EME examples, a word-for-word translation was added, sometimes accompanied by morphological information (in italics). The syntactic structure of the relevant part of each example is analysed in terms of the appropriate model.
II. COMPULSORY CONSTITUENTS OF EARLY MIDDLE ENGLISH CLAUSES
1. The Active Verbal Syntagm In its simplest form the active verbal syntagm consists of an inflected verb form. Examples; beorkeb his hundes (LB21340) bark his hounds =his hounds are barking pu fli st ni tes euer mo (ON238) you fly [at] night ever more =you fly always at night More complex realizations of the syntagm contain a verbal nucleus and one or more peripheral elements. As the latter belong to different paradigmatic classes, several peripheral positions have to be set up in the model of the active verbal syntagm. An action which took place in the past can be expressed by an inflected past verb form (=morphological past) or by a combinaation of an inflected form of habben 'have' and the past participle of a lexical verb (=syntactic past). As in OE there is no complementary distribution between the morphological and the syntactic past. This can be illustrated by the following syntagms:^ Examples; And alle bo pet anie wise po diefle er ikuemde bo bed nu mid him in helle, vordon and vordemde (PM271f.) and all those who in any wise the devil before pleased, they are now with him in hell, ruined and damned =and all those who pleased the devil before in any way are now with him in hell, ruined and damned Blipe mai he panne ben pet god haueö iquemed (PM174) glad may he then be who God has pleased =he who pleased God may then be glad
1
cf. Mitchell 1985:291: "The well-known overlap between the past tense and the periphrasis is attested by their use in parallel or connected sentences."
25
Both passages point out the consequences that man's behaviour in this world will have on his life after death. In the first example the morphological, in the second the syntactic past is used. The constituents haued and iquemed in PM174 are in mutual occurrence dependency. Therefore they are definitely not coordinated. The data show that the number and type of the constituents which are governed by EME active verbal syntagms depend on the lexical verb and not on the tense marker. In our examples the objects po diefle and god are governed by forms of the verb cwemen. The form haued in PM174 has no bearing on the syntactical properties of the active verbal syntagm. This establishes the form of cwemen as nucleus and haued as an actualiser. The peripheral position which is filled by haued will be called 'tense', and the model of the active verbal syntagm can provisionally be represented as tense
> verb.
The peripheral position 'tense' can be realized not only by a form of the auxiliary habben, but also by a form of the auxiliary ben ' b e 1 . Usually the lexical verb determines the choice of one or the other auxiliary.2 The auxiliary habben is used significantly more often than ben. Out of the 163 verbs of my data which are expanded by an element in 'tense 1 position, 116 combine only with habben, 40 only with ben. Bearing in mind the PDE situation, we may infer from this distribution that the substitution of ben in 'tense' position by habben is well under way in EME. This process had started in OE already. According to Mustanoja the distribution of the auxiliaries in OE was such that beon/wesan was used with intransitive, habban with transitive verbs. The latter was "occasionally also found with true intransitive verbs ( . . . ) , which suggests that even in OE it is developing into a kind of general auxiliary of the perfect and pluperfect tenses" (Mustanoja 1960:500). But he misrepresents the EME situation when he maintains that "by early ME practically all transitive verbs use have for the formation of their ^Appendices II.1 and II.2 contain lists of verbs which combine only with one or the other auxiliary.
26
perfect and pluperfect tenses" (Mustanoja 1960:500). The following examples show that in EME intransitive verbs can be combined with habben and transitive verbs with ben. Examples; pa men hafden iuast (LB22310) =the men had fasted him pat hit ischapen is (PA143) he who it created is =he who created it Mustanoja himself admits that the syntactic properties of a verb did not exclusively determine the choice of the auxiliary. He points out that originally both auxiliaries denoted a state, then they came to denote the effect or the result of an action. But only beon/wesan kept the former possibility beside the newly acquired one. Thus the pragmatic/textual meaning can be made to explain the use of ben in contexts where we would expect habben, i.e. in the presence of an object. Another feature which is considered relevant for the choice of ben as auxiliary is the semantic structure of the verb itself. Mustanoja mentions the semantic component 'imitative', which prevents intransitive verbs from being combined with habben. The same semantic component may explain the use of ben with transitive verbs, too. The following syntagm is a case in point. Example; bo he was iknowe pat Rimenhild was his 036 (KH983f.) The verb cnawen which usually means 'to know' is used here as a mutative verb, which expresses the transition from one state to another. The passage should be translated as 'when he [=Horn] learnt that Rimenhild was his own [=had remained faithful to him] ' . After Horn had been forced by an intrigue to leave the country, King Modi had proposed to Rimenhild. As the wooer was supported by Rimenhild's father, the wedding was imminent despite her refusal. In her distress she sent a messenger to Horn. The passage quoted above introduces Horn's reaction. It expresses his transition from the state of ignorance or uncertainty to the state of certain knowledge.
Examples of intransitive verbs combined with habben are more frequent than transitive verbs combined with ben. They foreshadow
27
the later The expansion press the
generalization of have as a tense auxiliary.^ following syntagms illustrate the use of habben as an of intransitive verbs irrespective of whether they exresult of an action or a state (of limited duration):
Examples; Ihc habbe walke wide (KH953) I have walked wide =1 travelled far pe saulen ... pat so hedden isped pat leyen hedden in peouene bed ( V P 2 5 5 f . ) the souls ... that so had behaved that [they] lain had in thieves' bed =those who had behaved in such a way that they had been thieves Hit hap ileie pere Fülle seue 3ere (KH1139) it has lain there full seven years =it lay there seven clear years pos laste on ure habbep i-travailed ( K S 3 4 , 3 ) those last [ones] one hour have worked =those last ones worked for one hour The choice of habben with intransitive verbs may also be motivated by the context: Example; Al to lome ic habbe igelt on worke and on worde (PM11) all too often I have sinned in works and in words =all too often have I sinned in deeds and in words The verbs in the immediate neighbourhood of gelten are transitive (habbe iquede ( 9 ) , habbe idon ( 1 0 ) , habbe ispent ( 1 2 ) , habbe ileid ( 1 2 ) ) . Therefore the use of habben needs no justification here. This series of active verbal syntagms with habben in 'tense 1 position may have had a levelling influence on the construction of gelten. Some verbs of my data combine with habben as well as with ben; bibenchen; To late ic habbe me biboht ( P M 8 ) 4 too late I have me bethought =too late did I think [about it] To bidden5 his milce to late we beod bi-pohte (D055f.) development has to be seen as complementary to another process, namely the generalization of be as a passive auxiliary. 4
cf. (LB16745)
5
cf. (LB2510)
28
breken;
don;
faren:
gan:
leggen;
setten;
to ask his mercy too late we are bethought =too late did we consider to ask for his mercy If du hauest is broken (B137) 6 if you have them broken =if you have broken them [=the laws of the church] bu ert of pyne ibroke (VP10) you are of pain broken-away =you overcame the pain what hauep pes mon ido (PL336) what has this man done =what did this man do mine dagis arren nei done (PA582) my days are nearly done =my days have nearly passed [=come to an end] he ... tolde him ful 3are Hu he hadde ifare (KH465f.) he told him full readily how he had fared =he readily told him how he had fared Julius be kaeisere was from bissen londe ifaren (LB8003f. ) Julius the emperor was from this land travelled =Emperor Julius had left this country Ihc habbe go mani mile (KH1176) I have gone many miles =1 travelled many a mile 333 wass pe33re weeke gan All ut tatt da"3 att efenn (OR4352f.) always was their week gone all out that day at evening =their [=the Jews'] week had always finished in the evening of that day Al to muchel ic habbe ispent, to litel ileid on horde (PM12) all too much I have spent, too little laid on hoard =too much have I spent, too little laid aside ure lord was i-leid him don to slepe (KS32, 1 5 f . ) our Lord was lain him down to sleep =our Lord had lain down to sleep Ice hafe sett her ο pis boc ... manio word (ORd41ff.)
I have set here on this book ... many words =1 have written many a word into this book bo vre louerd wes isethe to his supere (PL90) when our Lord was sat [down] to his supper =when our Lord had sat down to his supper With some of the verbs the difference 6
c f . (HK1238)
in construction is
29
matched by a difference in meaning (don, faren, gan). The verbs leggen and setten are combined with ben, when they govern a pronominal object which refers back to the subject. This construction is still current in Modern High German, cf. setzen Sie sich 'take a seat', leg' dich hin und sei still 'lie down and be quiet 1 . In PDE, however, the construction 'causative verb + reflexive pronoun object 1 no longer exists. The causative verbs have been replaced in this construction by their corresponding primaries: 'lay + reflexive pronoun' by lie, 'set + reflexive pronoun' by sit. The nucleus of an active verbal syntagm can also be expanded by a class of elements which correspond in form and function to the PDE modal auxiliaries. This class has the following members: ah, 7 can, dear, maei, man, mot, schal, parf, uten, willen 8 . They require the nucleus in its infinitive form. Examples; ich hire wulle teche as ic con (LR8) I her will teach as I can =1 will teach her as well as I can Summe me may per iseon (VP95) some one may there see =one can see some there Bade hi mu3en bi offerd pet sullen hine bihialde (PM286) easily they may be afraid that shall him behold =they may well be afraid who shall see him In PDE modal auxiliaries cannot manifest active verbal syntagms on their own, lexical verbs can. It is therefore correct to say that there is unilateral occurrence dependency between modal auxiliaries and lexical verbs in PDE, although there are environments in which - due to concord restrictions - the form of the lexical verb cannot replace the whole syntagm, cf.
auxiliary ah occurs in two variants: ah and ah to. They are in free variation, c f . ne og ur non oder to sunen (B285) 'not ought of-us none [the] other to s h u n ' = ' n o n e of us should shun the other one 1 ; Her-of we owe1 penche ( B 2 1 7 ) Of-this we ought think ='this we should consider BGCE §3.20 includes also the elements need and used to for PDE; they do not exist in EME. The EME forms man, parf, uten do no longer exist in PDE. Traces of man survive in (mainly northern) dialects of PDE.
30 he
will come *come
to the meeting
In EME - as indeed up to the Early Modern English period - modal "auxiliaries' can manifest nuclei of active verbal syntagms. Example: We solden us bibenche bet ofte and wel ilome Hwet we bied, to hwam we sulle, and of wham we come (PM327f.) we should us bethink better, often and well readily, what we are, to whom/what we shall and of whom/what we come =we should consider more, often and readily, what we are, where we go to, and where we come from Thus it seems that there is mutual occurrence independence between modal verbs and lexical verbs in EME. This assumption is easily refuted by the following considerations. It is true that sulle in PM328 is the nucleus of an active verbal syntagm. It governs a prepositional object (=to hwam). Sullen in PM286, however, cannot be the nucleus of sullen bihialde, because this active verbal syntagm governs the direct object hine, which is a syntactic property of the lexical verb bihialde. The correct analysis, therefore, is that PM328 and PM286 represent different construcions. Occurrence independence, however, holds only between elements of the same construction. As the syntactic properties of a syntagm are determined by its nucleus, it follows that in EME active verbal syntagms of the form 'modal auxiliary + lexical verb' the latter is the nucleus. The peripheral position which may be realized by modal auxiliaries will be called 'modal 1 , and the formal representation of this part of the verbal syntagm is modal
> verb.
Modal auxiliaries and tense markers do not exclude each other in front of the same verb. Example; Ic miote habbe bet idon (PM15) I might have better done =1 might have done better Therefore two peripheral positions have to be set up in the model of the active verbal syntagm. The elements in the two peripheral positions determine the verb in the same way. The complex syntactic relation involved is therefore parallel determination, and the enlarged version of the model is
31
modal
-> verb
tense The above example is to be analysed as modal:
-> verb: idon
tense: habbe As was shown before, some verbs combine only with habben, others only with ben in 'tense' position, some combine with either, but none with both at the same time. Interestingly enough, we find that more than one modal auxiliary can occur before the same verb in the same syntagm. Example; he sei him cunnen silde wel (PM334) he shall him can shield well =he will be able to protect himself well pat mannkinn shollde muohenn wel upp cumenn inntill heoffne ( O R 3 9 4 4 f . ) that mankind should may well up come into heaven =that mankind should have the opportunity of getting up into heaven The first modal auxiliary shows person and number concord with the subject, the second and the lexical verb are in the infinitive. Theoretically the syntagm may represent four different constructions: a) The two modal verbs realize different peripheral positions, each determining the lexical verb, i.e. modal 1 modal 2
modal 1 -> verb
or
modal 2
-> verb
b) The two modal verbs are coordinated in the same position, each determining the lexical verb, i.e. (modal 1 modal 2 )
> verb
c) The two modal verbs realize one peripheral position, modal 1 being subordinated to modal 2, i.e. (modal 1
> modal 2 )
> verb
32
d) The first modal verb determines a complex verbal nucleus, consisting of the second modal verb and the lexical verb, i.e. modal 1
> (modal 2
> verb).
The first two solutions can be ruled out on the grounds that there is no direct relation between the first modal and the nucleus. This statement presupposes that we know exactly what the syntagm means. The nearest equivalent in PDE are constructions of the type must be allowed to come. In German the isomorphic construction er muß kommen können exists. It can be shown that the permutation of the two modals changes the meaning of the syntagm, cf. er muß kommen können £ er kann kommen müssen 'he must be allowed to come1 56 'he may have to come'. This means that the two modal verbs do not determine the nucleus in the same way. Moreoover, the syntagm er muß kommen können is significantly different from er muß kommen ' he must come'. This means that there is no direct relation between muß and kommen. Accordingly we assume that there is no direct relation between shollde and upp cumenn either. This in turn implies that the second modal cannot form a complex nucleus with the lexical verb, because this would establish a direct relation between the first modal verb and the lexical verb. This was ruled out before. The correct analysis of the construction is therefore c) with the two modal verbs in the same peripheral position, the first subordinated to the second. OR3944f. should be analysed as (shollde
> nu^henn)
> upp comenn.
This analysis is borne out by tests with Modern English informants. Asked to answer questions with complex verbal syntagms, my informants included all elements in peripheral positions in their answers, e.g. Could the dog be lying under the table in B-floor sitting-room? - Yes, he could be. Should parents be allowed to forbid their children to go to the cinema? - Yes, they should be allowed.
Visser lists the construction (§1645, §2134), but he does not analyse the syntagmatic relations between the constituents. He mentions that the construction was frequent in ME, but that it died out after that period, surviving only in Scottish dialects.
33
In nearly all his examples the first modal is shall or should. As the construction is not attested for OE at all, but is common in Dutch and German, Visser conjectures that it may have been borrowed from these languages. As this construction shows, the 'modal' position may also be realized by more complex syntactic entities. Different constructions are possible. Fairly frequent is a realization by two modal auxiliaries, which are connected by a conjunction, in particular by and. 9 Examples; pat kunne & wille r 131 us deme (ONI88) who can and will right us judge =who can and will judge us competently se pet mai and nele pider come (PM372) the [man] that may and will-not thither come =he who can, but does not want to go there The active verbal syntagms of the examples are analysed as modal: kunne & wille modal: mai and nele
> verb: deme > verb: come
An active verbal syntagm may contain a form of ben which requires the following nucleus in present participle form. As the construction is isomorphic with the PDE progressive aspect, the position which can be realized by ben will be called 'aspect'. Expansion by an element in 'aspect' position does not, however, mean that the syntagm expresses simultaneity, incompletion, an action of limited duration, etc. 1 ^ Some verbs, e.g. waxen, walden, wunien occur more often in this construction than others. Examples; nys no wrt uexynde a wude ne a velde pat euer mvwe pas feye furp vp-holde ( P A 1 6 8 f f . ) not-is no herb growing in [the] wood nor in [the] field which ever may the doomed further uphold =no plant grows in the wood or in the field which could ever save those who are doomed to die po were in iherusalem ... Men wunyinde of alle kunnes londe (PL665f.) then were in Jerusalem ... men living of all kinds of countries =people of different countries were living in Jerusalem then the description of this construction c f . Chapt.vill. 10
cf. GCE § 3 . 2 9 f f .
34
preo daeies wes pe king wuniende pere & pan feorde daeie to Tintaieol he wende ( L ß 1 9 2 1 6 f f . ) three days was the king living there and [on the] fourth day to Tintaieol he went =the king stayed there for three days, and on the
fourth day he went to Tintaieol
The verbal syntagm of the first example expresses an action of unlimited duration, which corresponds to the simple form in PDE. Out of context PL665f. is ambiguous: it may either express an action of limited duration (=were living) or a completed action (=lived). LBl9216ff. is a clear case of enumeration of actions completed in the past. Mustanoja (1960:585) assumes that the construction which in OE is restricted to prose was associated in ME with vivid narration. Although ben is by far the most frequent 'aspect' expansion, it is not the only possible one. The following syntagm shows that ben commutes with verbs like sitten 'sit* in this function. When these verbs are used as peripheral elements in an active verbal syntagm, their meaning is nearly equivalent to that of ben. Example; he murninge sat (FB36) he mourning sat =he was unhappy An expansion in 'aspect' position does not exclude an element in 'modal' position, but data are lacking for the simultaneous occurrence of 'aspect' and 'tense' expansions. Visser's earliest examples (§2148).date from the 14th century. The lack of co-occurrence of the two expansions may be due to the fact that the 'have + past participle periphrasis' and the 'be + present participle periphrasis' had not yet well-established different functions. Therefore it is not necessary to set up two different positions in the model of the active verbal syntagm. The active verbal syntagms of the examples are analysed as ys were wes
sat
> > >
uexynde wunyinde wuniende
> murninge
The same position may be realized either by an 'aspect' expansion or by a 'tense' expansion. Depending on the type of expan-
35
sion, the nucleus is either a present participle or a past participle of a lexical verb. The formal representation of the model of the active verbal syntagm is modal tense 1 aspect]
t
-> verb
The nucleus of an active verbal syntagm may also be expanded by a negative particle. In EME the most frequent negative particle is ne, and there is unilateral occurrence dependency between the negative particle and the verb. As the verb determines the syntactic properties of the active verbal syntagm, there is also unilateral functional dependency between these two elements. The verb is the nucleus, the negative particle an expansion realizing a peripheral position which will be called 'negation 1 . Examples; he ne hedden per before ine him beliaue (KS30, 4 f . ) they not had there before in him belief =there they did not believe in him before Hwi ne serue we crist (SC1) why not serve we Christ =why do we not serve Christ The negative particle ne, which precedes the nucleus has two variants: ne and the proclitic form n-. The latter is a positional variant which occurs with the verbs habben 'have', witen 'know', bileven 'believe', seon 'see'. 1 2 Examples; he be bet nise3e (ON382) he the better not-see =he does not see the better hy nvten hwat hi dob (PL442) they not-know what they do =they do not know what they do The position 'negation 1 can also be realized by rK^t, although this element occurs much less frequently. Its usual position is 11
Braces indicate alternative realization possibilities of the same position.
^That the variant rv^ is not obligatory in these contexts can be seen from syntagms like we be ne yseyen (BE305) 'we you
e at.? ' — ' T.rA /Ι η /^ Ί^Λ·Κ ΟΑΑ ti/Mi ' not saw' =' we did not see you'. τ^/^Ί"
36
after the verb, but occasionally no3t also precedes the verbal nucleus. Example; Neded de de deuel no3t (B164) urges you the devil not =the devil does not urge you The frequency of no3t must have increased rapidly after the EME period, because according to Mustanoja (1960:340) it had become the ordinary negative particle in the 14th century. A third possibility is the co-occurrence of ne and np_3t in 'negation 1 position. Examples; ich ne singe nopt whan ich teme (ON1470) I not sing not when I breed =1 do not sing when I breed ich bi dai no3t ne flo (ON372) I by day not not fly =1 do not fly during the day As both elements also occur independently of each other, one might be tempted to interpret their co-occurrence as a sign of emphasis. This interpretation is refuted, however, by parallel syntagms where the combination of the particles commutes with one of them. Examples; we[n]st pu pat ich ne cunne singe (ON47) think you that I not can sing =do you think that I cannot sing pu ... telst pat ich ne can no3t singe ( O N 3 0 9 f . ) you say that I not can not sing =you say that I cannot sing Thus ne-no-^t is a (discontinuous) constituent which commutes with the elements ne and no3t in 'negation 1 position. The analysis of the active verbal syntagms of the above examples is 1IC
neiaueii
ne
serue ±3636^ wvten 14 nede
V
ne noßU liti
fl ° 15 singe '^ singe I D
13
As positional variants are irrelevant in syntactic analyis, it is customary (in AF) to represent pleremes in a 'normalized 1 form. 14
cf. fn.13
15
This analysis neglects the elements in 'modal' position.
16
cf. fn.15
37
They realize that part of the model of the active verbal syntagm which may be represented as negation
> verb.
An element in 'negation 1 position does not exclude any of the other expansions. Examples; No mon ne may herre iseon (VP 34) no man not may higher [trees] see =nobody can see higher [trees] Nabbed he no ping vo^ete of pan pet he iseoen (ΡΜ9Θ) not-have they no thing forgotten of this that they saw =they have not forgotten anything of what they saw nys no wrt uexynde a wude (PA168) not-is no herb growing in [the] wood =no herb grows in the wood The examples illustrate the following combinations of expansions: negation + modal, negation + tense, negation + aspect. This constellation requires that we set up a separate 'negation' position, and the model of the active verbal syntagm is now negation modal Γ tense 1 |aspect]
-> verb
The nucleus of an active verbal syntagm may also be expanded by an inflected form of the verb ginnen; the nucleus follows in its infinitive form. Past forms of ginnen are much more frequent than present tense forms, the ratio being about 20:1. The inflected forms of the past are gan/gon, gunne, gunnen. The phonological similarity of the singular forms gan/ gon explains why they are sometimes substituted by the forms can/con of the modal auxiliary.17 Mustanoja 1960:610. Visser even maintains that can and gan are phonological variants of the same verb: "both 3 and c represent allophones of the same phoneme /g/, differing not in their being voiceless or voiced, but in the energy of articulation" (Visser §1477). He supports his argument by quoting cases of reverse spelling.
38
Example: pe see him con ded prowe Vnder hire chambre wove (KH970af.) the sea him did dead throw under her chamber window =the sea threw him dead under her chamber window This syntagm shows that the original meaning 'begin' became weakened during the EME period, and ginnen could be used as an empty periphrastic verb. Mustanoja's statement that this holds only for the past forms of ginnen should be modified so as to include all forms.^® The following two syntagms illustrate the range of possible meanings of the present tense forms of ginnen. Examples: pe blostme ginnep springe & sprede (ON437) =the blossoms begin [to] spring and spread clerkes ginnep songes wirche (ON722) clerics do songs make =clerics compose hymns The first example occurs in the context of a conventional picture of spring: the snow has melted, early flowers begin to blossom, the birds come back. ON437 describes the beginning of a new season. The context of the second example is a simile in which the nightingale likens church services on earth to the state of everlasting bliss in heaven. Earthly and heavenly hymns are one aspect of the comparison. The form ginnep in ON722 has no ingressive meaning. The different shades of meaning are irrelevant, however, for syntactic analysis, because they have no bearing on the syntactic relation between the forms of ginnen and the following infinitive. It is therefore not adequate to set up two different signs ginnan - as Visser does (§1269, § 1 4 7 7 ) . He explicitly rejects earlier descriptions (Funke 1922, Homann 1954, Mustanoja 1960) which set up one verb ginnen with a range of different meanings. When ginnan expresses the beginning of an action, state, etc., Visser calls it a 'verb of aspect 1 ; when this meaning is not present, it is an auxiliary (subclass 'periphrastic v e r b ' ) . He admits, however, that "it is very hard to discriminate between these two kinds of gan" (§1269). A valuable piece of new information which he contributes 18"Through frequent use it loses its ingressive colour more and more and occurs mainly in the intensive-descriptive functions." (Mustanoja 1960:611)
39
is that the ginnen-expansion is restricted to verse texts. This means that among the usages of ginnen enumerated by Funke - "als Mittel zur Perfektivierung, als deskriptive Ausdrucksform, als farbloses Flickwort" (Funke 1922: 15) - the last one is perhaps even more frequent than was assumed before in the analysis of individual syntagms. Another element which may realize a peripheral position in an active verbal syntagm is the verb don. Here, too, the nucleus follows in its infinitive form. The meaning of don in this function is as opaque as that of ginnen. Very often an expansion by don does not change the meaning of the nucleus at all. When don is used in texts with end rhyme, it may safely be assumed that it serves to place the infinitive into end (=rhyme) position. Example; Abulf hit dude write pat Horn ne luuede no3t lite ( K H 9 3 1 f . ) Athulf it did write who Horn not loved not little =Athulf wrote it, who loved Horn very much There are other instances, however, where don clearly serves the purpose of emphasis. Example; se pet e^hte wile hialde wel pe hwile Jpe hi mot wealde, / 3ieue hi for godes loue, panne ded he hi wel ihialde (PM 5 5 f . ) the [man] who [his] possession will hold well as long as he may dispose [of it], give it
[away] for God's
love, then does he it well hold =he who wants to keep well his possession while he may dispose of it, should give it away for God's sake; then does he keep it well The four active verbal syntagms wile ihialde, mot wealde, -^ieue, ded ihialde form two pairs of semantic contrast in a crosswise arrangement. In the pair mot wealde ; ^ieue the antonymy is realized by the semantic structure of the verbs wealden 'keep' : ^euen 1 give away'. In the pair wile ihialde : ded ihialde the verbal nucleus is the same. The semantic contrast is realized by the peripheral elements, wille and ded. The actualiser wille expresses an intention, ded stresses the realization of this intention. The obviously deliberate rhetorical structure of the passage supports the argument that ded ihialde is not a neutral statement, but an
40
early use of emphatic do.^' It thus refutes Mustanoja's hypothesis that "unambiguous cases of emphatic periphrastic do are not recorded until the 15th century" (Mustanoja 1960:606). The verb don as an actualiser in an active verbal syntagm has a similarly wide range of meanings as ginnen. Funke describes them as follows: "Doch begegnen uns einige Beispiele, in denen don entweder bereits an der Grenze zum farblosen Tätigkeitswort steht, vielleicht völlig pleonastisch geworden ist oder aber zum Intensivuni neigt" (Funke 1922:15). Both elements are either mere periphrastic verbs which do not modify the meaning of the following infinitive at all, or they are aspect markers. 2 ^ Whatever their meaning in an individual syntagm, it seems safe to assume with Visser that they exclude each other in the same active verbal syntagm. 2 ^ This means that they occupy the same peripheral position. Expansions by don or ginnen will be called "periphrasis' expansions. Nearly all active verbal syntagms of my data which have a form of ginnen or don in peripheral position are not expanded by any other element. Only one syntagm illustrates the combination of don and an element in 'modal 1 position. Example; pat schall don vs tene 2 2 (KH683) that shall do us harm =that will harm us Visser quotes one syntagm with a combination of a negative particle and don: 19
The EME text is an allusion to a passage in the New Testament where Jesus promises salvation to those who follow him. The Authorized Version preserves the same rhetorical structure: "For whosoeuer will saue his life shall lose it, but whosoeuer shall lose his life for my sake and the Gospels, the same shall saue it." (Mk 8:35) 2
^Dürmüller/Utz ( 1 9 7 4 ) analyse "be + present participle" and "do + infinitive" as alternative expressions of the progressive aspect. Mosse (1952:§134) considers ginnen as an aspect marker: "To express the idea of a 'beginning action (inchoative aspect) ME normally used the verb gin(ne)". 21
Visser goes even one step further, maintaining that don and ginnen are substitutable for each other. He supports this hypothesis by the observation that the same text has don in one MS and ginnen in another. 22
tene is a verb here.
41
Example; we ne doz noujt ore ordre breke (quot. Visser §1438) we not do not our order break =we do not break our order In the EME texts investigated so far the combination of a 'periphrasis' expansion and either an 'aspect 1 or a 'tense' expansion has not been found. Consequently, all three expansions realize the same position and the model of the active verbal syntagm in its most elaborate form is negation modal tense aspect periphrasis
(
-> verb
Applied to the last two examples, it yields the following analyses. Analysis:
negation: 0 modal: schall tense "\ aspect V: don periphrasis J
-> verb: tene
negation: modal : 0 tense aspect periphrasis
-> verb: breke
(
doz
Usually the nuclear position is realized by one lexical verb whose form is determined by the immediately preceding expansion. One more complex realization is the combination of two (or more) lexical verbs which are connected by a conjunction, mostly by and. 23 Examples; hi wolden heore louerdes dom iseon and ihere (PL228) they would their Lord's judgement see and hear =they wanted to see and hear their Lord's judgement pu mi 3 t, Hule, sitte & clinge (ON743) you may, Owl, sit and wither-away =0wl, you may stay there and decay the description of this construction cf. Chapt.VIII.
42 Analysis:
negation: 0 modal: wolden tense \ aspect t : 0 periphrasis]
-> verb: iseon and ihere
negation: modal: mißt tense "j aspect :0 periphrasis)
-> verb: sitte & clinge
I
2. The Copulative Syntagm Like the active verbal syntagm the EME copulative syntagm can be realized by a finite verb form. It may therefore seem superfluous to set up two separate models. The two main reasons for nevertheless doing so are the internal structure of the copulative syntagm on the one hand and the possible types of complementation on the other. 24 Examples; Stille beo bu benne (KA24, 1) still be you then =be quiet now cof he waxeb ( B 1 2 4 ) bold he grows =he grows bold Ho stondeb pustrur pane pe nyht (VP225) they stand darker than the night =they [=the devils] are blacker than the night The first example illustrates the most frequent realization of the nucleus of a copulative syntagm by a form of ben ' b e ' . The verbs which commute with ben in this function have next to no lexical meaning. They serve to establish a syntactic and a semantic relation between the subject and another 1C (which will be called 'copulative complement' - cf. C h a p t . V . 2 . 2 . 1 . ) . 2 5 24
The complementation types are described in Chapt.V.2.2.
25
In my data the following verbs form nuclei of copulative syntagms: ben, bicumen, bilaefen, biliven, biseme, far en, elden, hoven, laeven, liggen, livien, Sitten, standen, pünchen, waxen, würben.
43
The internal structural differences between the active verbal syntagm and the copulative syntagm concern not only the realization possibilities of the nucleus, but also the number of peripheral positions and their realization. A 'negation 1 position has to be set up for the copulative syntagm, too. It is realized by the particle ne or by the discontinuous constituent ne-no^t. The particle npjt alone is not recorded in this peripheral position of the copulative syntagm. As in the active verbal syntagm, the negative particle ne has two positional variants: ne and the proclitic form n-. The latter negates the forms am, art, is, were, weren , resulting in the forms nam, nart, nis, nere, neren . The particle ne precedes the nucleus of the copulative syntagm (alone or as part of the discontinuous constituent), the element noßt follows it. Examples; pat lond nis god (ON999) that land not-is good =that land is not good Ne beop nopt ones alle sunne (ON1395) not are not the-same all sins =all sins are not alike Adopting the terminology used in the Grammar of Contemporary English (GCE), I shall call the nuclear position 'copula'. The model of the copulative syntagm which is required for the analysis of the examples above can be represented as negation - > copula. The copulative syntagms of the examples- are analysed as ne
-> ->
is262 beop '
In the copulative syntagm the 'modal' position cannot be realized by the elements man and uten. A possible explanation for this is that these modal auxiliaries were no longer well integrat-
pleremes are given in their 'normalized' form, cf. fn.13. 27
cf. fn.3
44
ed in the language system of the EME period.^8 Examples; syker he may sitte pe hyne hauep to i-vere ( P A 2 1 7 f . ) safe he may sit who him has as friend =whoever possesses it [=reason] can be safe se man bet wile siker bien (PM41) the man who will safe be =the man who wants to be safe The element ah in 'modal' position of a copulative syntagm has the variants ah to and ah for to. They are free variants.
Examples; mi wit oh to bi more (PM2) my wit ought to be more =1 ought to be wiser Englond auhte forto ben Youres ( H K 2 8 0 0 f . ) =England ought to be yours The 'modal' position can also be realized by more complex syntactic entities, e.g. by two modal auxiliaries which are connected by a conjunction. Example; he one mai and sei bien engles and mannes blisce (PM378) he alone may and shall be angels' and man's bliss =he alone can and will be the bliss of angels and mankind As in the active verbal syntagm, the elements in 'modal' position are actualisers, because they do not commute with zero. The copulative syntagms of the examples are therefore to be analysed as may wille oh to auhte forto
> > > >
sitte bien bi ben
The formal representation of the respective part of the model of the copulative syntagm is modal
> copula.
Elements in 'negation' position and in 'modal' position are not mutually exclusive in the same copulative syntagm.
Examples; ne sei no merode bi swo muchel swo is godes sihte (PM367) 28
This assumption is supported by the observation that also in the 'modal 1 position of the active verbal syntagm man and uten occur very rarely (4 examples for man, 7 for u t e n ) .
45
not shall no mirth be so big as God's sight =no is_ so sight of God - - - delight —— j — _ big ^ as the _ ^ bu ne mißt mid us bo blibe (ON418) you not can with us be merry =you cannot be merry with us The representation of the model of the copulative syntagm which allows the analysis of the examples above is negation
-> copula
modal Analysis:
negation: ne -> copula: bi modal: sei negation: ne -> copula: bo modal: mißt
The 'tense' position in the copulative syntagm is realized by forms of habben 'have' or ben ' b e ' . The two auxiliaries stand in complementary distribution such that the nucleus ben is expanded by forms of habben, the nuclei waxen and würben are expanded by forms of ben.^9 The combination of ben in 'tense' position with the nuclei waxen and würben may be due to the fact that these nuclei express the transition from one state to another.30 The occurrence of have in 'tense' position needs no justification. It seems quite natural that an element "on its way to being generalized in a particular peripheral position will also combine readily with the most frequent nucleus of the construction. Examples; Wei longe ic habbe child ibien on worde and on dede (PM3) well long I have child been in words and in deeds =a very long time have I been a child in my words and in my deeds nuclei are not recorded with an element in 'tense' position in my data. 30
The semantic component 'mutative' is also considered responsible lor the combination of particular verbal nuclei with ben in 'tense' position, cf. chapt.II.1.
46
Godess enngless waerenn pa Well swipe glade wurpenn Off pat (OR3914ff. } God's angels were then very glad become of that =then God's angels had become very glad about it That part of the model of the copulative syntagm which is realized in these examples can be represented as tense
> copula.
The copulative syntagms of the examples are analysed as habbe waerenn
> ibien > wurpenn.
Apart from elements which can occupy the positions "negation , 'modal' and 'tense', no other expansions occur in the copulative syntagm. Constituents in the peripheral positions 'modal* and 'tense' determine the form of the following element. An element in 'modal' position requires the following constituent to have infinitive form, an element in 'tense 1 position requires the following constituent to have past participle form. The very first element of a copulative syntagm shows person and number concord with the subject. An element in 'negation' position cannot be marked for person or number and it has no influence on the form of any of the other constituents of the syntagm. The elements in the peripheral positions do not determine the nucleus in a noticeably different way. The complex tactic relation in the copulative syntagm is, therefore, parellel determination. As there is unlimited combination between the different expansion types, three peripheral positions have to be set up in the model. Its generalized form can be represented as 1
negation modal tense
-> copula.
The following syntagms illustrate 'tense' and 'negation 1 + 'tense'.
combinations of
Examples: hit miote han iben wel his Wille (FW87) it might have been well his will =it may well have been his will nedde lust I-ben of mine moube (FW100) not-had lust been of my mouth =if there had not been my gluttony
'modal'
47
Analysis: negation: 0 modal: mi^te tense: han negation: ne modal: 0 tense: hadde
-> copula: iben
-> copula: I-ben
3. The Passive Syntagm Within the theoretical framework of AF chosen for this study, the passive syntagm needs special justification. In 1968 we find Mulder maintaining that there is no structural difference between a copulative predicative syntagm (e.g. was ill) and the verbal part of a passive construction (e.g. was beaten).3 1 This is the position still taken by him today: "One sees that, unlike most other linguists, I regard the so-called 'passive' as a copulative predicative" (Mulder 1980:158). It is only consistent therefore that he does not set up a model of the passive syntagm for his description of PDE. In EME the passive voice is expressed by a form of ben 'be 1 or würben 'become' plus the past participle of a lexical verb. Both elements are in mutual occurrence dependency. The syntactic relation between them is therefore either subordination or interordination. Both constituents can have other functions as well. Forms of ben can realize the nucleus of a copulative syntagm or the peripheral position 'tense' in an active verbal or in a copulative syntagm. Past participles can be nuclei of active verbal syntagms. Whether a past participle has the function 'constituent of a passive syntagm' depends on the function of the other constituent, and vice versa, since a mutual functional dependency relationship holds between the two constituents, i.e. they are interordinated. It could be argued that the situation was the same "We also see that it is not necessary at all to distinguish between passive and copulative on the higher levels of analysis." (Mulder 1968:86)
48
when ben had the function of a tense marker in an active verbal syntagm. However, this objection is easily refuted, since in an active verbal syntagm of the form 'ben + past participle 1 , an inflected verb commutes with this construction, thus establishing the form of the lexical verb as the nucleus of the syntagm. A second consideration supports the hypothesis of interordination in the passive syntagm. If two isomorphic constructions (e.g. ben + past participle) have two different grammatical meanings (e.g. active, past:passive), it follows that the syntactic relations between their constituents must be different, too. Otherwise the different grammatical meanings would be unaccounted for. The relation in the passive syntagm would also be different from that in the active verbal syntagm, if we assumed - as Mulder doesthat the past participle was subordinate to the form of ben. This would mean that the copula was the identity element of the syntagm, which determines its syntactic properties. This is not the case. The syntactic properties of the passive syntagm are partly determined by one of its constituents, partly by the other. Neither of them can be called the identity element of the construction. The past participle determines whether the passive syntagm can be expanded by an object ( c f . PDE I was given an apple [1 object: an a p p l e ] ) , the copula restricts the number of possible object expansions by one ( c f . PDE he gave me an apple [2 objects: me, an apple]). The model of the EME passive syntagm contains therefore two equivalent positions; they will be called 'passive copula' and 'past participle'. The formal representation of a model which accounts for very simple realizations is passive copula
past participle.
In the discussion earlier of the relation between the constituents 'passive copula' and 'past participle, 1 only the element ben was mentioned. The reason for this was numerical: ben simply occurs much more frequently in 'passive copula' position than wurpen does. The two also involve somewhat different event-notions: ben expresses primarily a state, wurpen a process, although
49
this is of secondary importance in syntactic analysis.32 The model developed so far accounts for passive syntagms like the following. Examples; hanged wurpe he on a hok (HK1102) hanged be he on a hook =may he be hanged on a hook parmide beob men acwalde (ON1370) therewith are men killed =men are killed with it Analysis: passive copula: wurpe < passive copula: beo£>
past participle: hanged > past participle: acwalde
Usually the nuclear position 'past participle' is realized by just one plereme. Occasionally the position is occupied by two past participles which are connected by a conjunction.33 Examples; hw he weren born and fedde (HK2984) =how they were born and fed if bu art iworpe oper ishote (ON1121) =if you are hit or shot dead hit was swo i-seid and be-hote hwilem bi po profetes (KS26, 2 1 f . ) it was so said and promised whilom by the prophets =thus it was once said and promised by the prophets The complex nucleus of the passive syntagm can be expanded by a number of peripheral elements. While the peripheral positions and their realizations resemble those of the active verbal and the copulative syntagm, they are not completely identical with them. The particles which are used for the negation of passive syntagms are ne (with the positional variant n-, c f . C h a p t . I I . 2 . ) , npjt and ne-no^t. Ne is considerably more frequent than the other two. 32-rhe possibility of distinguishing between a static passive and a dynamic passive was lost in English when wurpen dropped out of use as a passive copula, i.e. at the end of the ME period ( c f . Mustanoja 1960:439). It was not until several centuries later that this distinction was reintroduced, when the aspect expansion was extended to the passive syntagm: "The passive was being built by X with overt progressive auxiliary ... did not develop until the end of the eighteenth century" (Traugott 1 9 7 2 : 1 4 4 ) . 33syntagms of this type are called 'conjunctive constructions'; for their description cf. Chapt.VIII.
50
Examples; he nis barof bireued (ON120) he not-is thereof bereft =he is not bereaved of it [=his head] neren hi nouht ihud (PL645) not-were they not hidden =they were not hidden The 'modal' position can be realized by the auxiliaries maei, mot, schal and willen. They require the following constituent to have the infinitive form. Comparing the possible elements in this peripheral position of the active verbal, the copulative and the passive syntagm, we note that their number decreases from syntagm to syntagm. Those with the highest frequencies in the active verbal syntagm are also recorded in the two other syntagms; those with a lower frequency in the active verbal syntagm also occur in the respective position in the copulative syntagm, but not in the passive syntagm; those with the lowest frequency in the active verbal syntagm are restricted to this construction. Examples; ech ... sal panne ben idemed (PM173) each ... shall then be judged =then everybody will be judged til his sone mohte ... king ben maked of Denemark (HK378ff.) until his son might ... king be made of Denmark =until his son could be made king of Denmark The only auxiliary which can occupy the 'tense' position in a passive syntagm is habben 'have'.·^ 4 it requires the past participle of the passive copula. Examples; twien pu hafuest ibeon ouer-cummen (LB8325) twice you have been overcome =you were overcome twice panne he hauede ben ofte swngen (HK226) when he had been often beaten =when he had often been beaten The nucleus of the passive syntagm can be expanded by more than one peripheral element at the same time. As it cannot be demonstrated that the expansions determine the nucleus in differ34
Although it must be admitted that passive syntagms with an element in 'tense 1 position are not very frequent in EME, Mustanoja misrepresents the situation slightly by quoting just two examples (Mustanoja 1 9 6 0 : 4 4 0 ) . The Haveloktext alone provides fourteen.
51
ent ways, the complex tactic relation in the passive syntagm is parallel determination. The combination of a 'negation 1 and a 'modal' expansion is quite frequent. Examples; hit ne moste ifulled beon (VP142) it not must baptized be =it could not be baptized pis selkouth mihte nouht ben hyd (HK1059) this wonder might not be hidden =this wonder could not be concealed Thus the model of the passive syntagm must contain the two peripheral positions 'negation' and 'modal'. An element in 'tense' position, however, excludes any other peripheral element. This striking difference between the EME and the PDE passive syntagm can be explained by the observation that in EME 'tense' expansions of passive syntagms were not yet well integrated in the language system, as is also suggested by their low frequency. A model of the passive syntagm which accounts for this situation must provide two alternatives in its peripheral part; one alternative containing the two positions 'negation' and 'modal 1 , the other only one position, 'tense'. This model can be represented as35 negation
tense
modal
-> (passive copula
past participle)
The examples with elements in peripheral positions are analysed as negation: ne, 'tense: 0 modal: 0
-> (pass.cop.: is
p.p.: bireued)
negation: ne 'tense: 0 modal: 0
-> (pass.cop.: weren
p.p.: ihud)
slash indicates alternative possibilities here, as braces are not available when one alternative contains more than one position.
52
negation: 0
tense: 0
-> (pass.cop.: ben
p . p . : idemed)
negation: 0 / /tense: 0 modal: mohte/
-> (pass.cop.: ben
p . p . : maked)
modal: sal
negation: 0 modal: 0 negation: 0 modal: moste
tense: hafuest
tense: 0
negation: nouht tense: 0
modal: mihte
-> (pass.cop.: ibeon < > p.p.: ouer-cummen)
-> (pass.cop.: beon
(pass.cop.: ben
p . p . : ifulled)
> p . p . : hyd)
Elements in peripheral positions of the passive syntagnt do not restrict the realization possibilities of the nuclear positions. The following syntagm illustrates a combination of two participles connected by a conjunction in the position 'past participle', and an element in the peripheral position 'tense'. Example; unto pis [ilke] day haue ich ben fed and fostred ay (HKl433f.) unto this very day have I been fed and fostered ever =to this very day I have always been fed and fostered Analysis: negation: 0 modal: 0
tense: haue
-> (pass.cop.: ben
subst.: flocc 4 (wel > blipe) > subst.: mon (ful > god) > subst.: weder (swa swide ungemetlice > mycel) > subst.: wind
3
For the description of the adjectival syntagm cf.
III.2. 4
The plereme patt is not included in this analysis.
Chapt.
55 adj.synt.: (leide adj.synt.: ((sube
fair)
subst.: seftes (and hende)) > subst.: mon
Elements in the 'adjectival syntagm' position co-occur with a number of other peripheral constituents. These include possessive pronouns, demonstrative pronouns, articles, numerals, and indefinite pronouns. Examples; possessive pronouns;·* h re wode wise (ON1029) =their crazy habits leuene mine fadre (LB3549) dear my father =my beloved father his lemman hende (FB820) his darling pretty =his pretty love demonstrative pronouns; pis halij godspel ( B H 4 , 1 ) =this holy gospel pat maide hende (FB478) =that pretty girl articles; 6 a gud mascun (FB326) =a competent mason be gret parel (VP12) =the great peril numerals; twa gildene scrinen (PC1070,31) =two gold shrines fiftene mycele roden (PC1070,32) =15 big crosses
5
Possessive pronouns usually precede the substantival nucleus, but occasionally follow it, e.g. wordes his (FB757) 'his words'. 6
Inflected forms of the definite and the indefinite article are quite frequent, e.g. in per sae (ΡΜΘ3) 'in the sea', aenne broder (LB556) 'a brother 1 .
56
indefinite pronouns: 7 non ... sturne mon (FB701) =no stern man sum hali^ biscop (BH16,6) =a certain holy bishop This evidence suggests that the constituents with an adjectival nucleus and the other peripheral elements either realize at least two different peripheral positions, or they are coordinated in one position. Coordination implies, however, that the coordinated elements belong to the same distribution class. This is not the case for adjectival syntagms and all the other constituents, and consequently at least two peripheral positions have to be distinguished in the model of the substantival syntagm. The complex tactic relation is parallel determination, because the expansions do not determine the nucleus in noticeably different ways. If none of the expansions illustrated above excluded any of the others, the number of peripheral positions would be identical with the number of expansions. This is certainly not the case. As the structure of the PDE substantival syntagm differs from that of the corresponding OE construction, we expect that the EME substantival syntagm will share some properties with the OE construction, some with the PDE. OE possessive and demonstrative pronouns do not belong to the same paradigmatic class; they can expand the same substantival nucleus. 8 In PDE they exclude each other as expansions of a substantival syntagm. In my EME data elements of these two classes of pronouns do not occur together, i.e. they realize the same peripheral position. The same position can also be realized by a definite or indefinite article. As these elements serve to pick out a closed set of the objects denoted by the substantive, I shall call this peripheral position 'deicticum'. The generalized form of the model of the substantival syntagm developed so far is ^Generally indefinite pronouns precede the nucleus; the negation particle n o ( n ) sometimes follows it, e.g. ioie none (FB106) 'no j o y 1 . 8
cf. Pilch 1970:219
9
cf. GCE §4.27
57
deicticum
-> substantive.
adj.synt. Its application to the examples yields the following analyses: deicticum: höre adj.synt.: wode deicticum: mine
-> subst.: wise
-> subst. : f adre
adj.synt.: leuene deicticum: his adj.synt.: hende deicticum: bis adj.synt.: hali
-> subst.: lemman
-> subst.: godspel
deicticum: pat adj.synt.: hende deicticum: a adj.synt.: gud
-> subst.: maide
-> subst. : mascun
deicticum: pe adj.synt.: gret
-> subst.: parel
As is the case in other periods, in EME the semantic relation 'possessive' can be expressed not only by a possessive pronoun, but also by a substantive in genitive case. This substantive may or may not be expanded. Constructions of this type are called here 'genitival syntagm 1 .^ This means that a genitival syntagm can also realize the position 'deicticum 1 . 10'For the description of this syntactic type cf. Chapt.IV.4
58
Examples: Codes rice (BH2,13) =God's realm blauncheflures chaumbre (FB479) =Blancheflure's chamber Analysis: deicticum: Codes -> subst.: rice adj.synt.: 0 deicticum: blauncheflures -> subst.: chaumbre adj.synt.: 0 As a rule, the genitive case is marked by an inflectional ending, although occasionally this ending is lacking (the so-called zero genitive). Such substantival syntagms in 'deicticum 1 position usually precede the substantival nucleus, but they may also follow it. Examples; his fader chaumberlein (FB18) =his father's chamberlain his leue moder luue (PL704) =his dear mother's love blisse dales breo (LL35) bliss three days =three days' bliss Analysis: deicticum: his fader -> subst.: chaumberlein adj.synt.: 0 deicticum: his leue moder -> subst. : luue adj.synt.: 0 deicticum: daies breo -> subst.: bliss adj.synt.: 0 The function of a genitival syntagm may also be indicated by a postponed possessive pronoun. Example; be king his fader (FB797) =the king's father
59
Analysis: deicticum: pe king his
-> subst.: fader
ad j.synt.: 0 The positions occupied by indefinite pronouns in EME differ somewhat from those possible in PDE. In PDE they realize the same peripheral position as possessive pronouns, articles, demonstrative pronouns and genitival syntagms, cf. my the these the old man's some
expensive books
In EME, in contrast, indefinite pronouns co-occur with elements in 'deicticum' position. Examples; sume pa munecas (PC1083,25) some the monks =some monks/some of the monks heore ineward vych del (VP151) their inwards every part =every part of their inwards The first example illustrates a definite article/demonstrative pronoun, the second a genitival syntagm (zero genitive) in 'deicticum 1 position. The nuclei munecas and del are not marked for case. In EME singular case endings had already been dropped with the exception of genitive. Case opposition in the plural was rare, the £s-ending having been extended to all plural forms. This resulted in the confusion of two formerly distinct constructions. In OE the indefinite pronouns could either realize a peripheral position in a substantival syntagm ( e . g . hilderinc sum (Bwf3124) 'some/a warrior 1 ) or the nuclear position of a pronominal syntagm. In this latter construction the pronominal nucleus could be expanded by a genitival syntagm (e.g. gumena sum (Bwf1499) One of the m e n ' ) . OE pronominal syntagms of this type had a partitive meaning. When the case endings were dropped, the two constructions became isomorphic. Only the context allows us to disambiguate constructions like sume his geferan.^"* It may mean 'some of his 11
cf. the title of T. Heltveit's article ( c f . Bibliography)
60
friends', implying that he has more than these, or 'some friends of h i s ' , without further information about the number of this person's friends. As these PDE paraphrases show, one of the meanings was expressed differently after the end of the ME period, when the construction 'some friends of his' became available. In contexts where the partitive meaning was important, EME had the alternative of an unambiguous pronominal syntagm followed by a prepositional syntagm which was introduced by partitive of (e.g. sume of dam cnihten (PC1083,17) 'some of the warriors'). The isomorphism of the two constructions explains why in EME indefinite pronouns could co-occur with elements in 'deicticum' position. Syntagms of the form 'indefinite pronoun + possessive pronoun, article, demonstrative pronoun, or genitival syntagm + ... + substantive' are syntactically ambiguous in EME. They may realize a substantival or a pronominal syntagm. If they realize a substantival syntagm, the indefinite pronouns are in a different peripheral position from 'deicticum 1 . It will be called 'quantif i e r ' , because the elements which realize it denote a certain quantity. After its integration into the model of the substantival syntagm, its generalized form is quantifier deicticum adj.synt.
-> substantive,
The analysis of the examples with indefinite pronouns in a peripheral position is quantifier: non deicticum: 0 adj.synt.: sturne quantifier: sum deicticum: 0 adj.synt.:
-> subst. : mon
-> subst.: biscop
The following syntagms illustrate combinations of indefinite pronouns with elements in the position 'deicticum 1 . Examples; aenig his foregenga ( P C 1 0 8 7 , 6 4 ) any his predecessor =one of his predecessors/a predecessor of his
61
sum his be3enae (BH18,5) some his servants =one of his servants/a servant of his Analysis: quantifier: «nig deicticum: his adj.synt.: 0 quant i f ier: sum deicticum: his adj.synt.: 0
-> subst.: foregenga
-> subst.: pe^en«
The examples with numerals in a peripheral position have not yet been analysed. Whereas numerals do not exclude constituents in the position 'adjectival syntagm' and also co-occur with elements in the position 'deicticum 1 , they exclude indefinite pronouns in front of the same nucleus. This means that cardinal numbers and indefinite pronouns realize the same position, namely 'quantifier 1 . Examples; hire armes tweie (KH301) =her two arms pat on cupe (FB438) =that one flower-pot pa twegen kyngas (PC1070,45) =the two kings The model of the substantival syntagm does not require to be modified to allow the analysis of the examples with cardinal numbers in a peripheral position. Analysis: quantifier: twa deicticum: 0 adj.synt.: gildene quantifier: fiftene deicticum: 0 adj.synt.: mycele quantifier: tweie deicticum: hire adj.synt.: 0
-> subst.: scrinen
-> subst. : roden
-> subst. : armes
62 quantifier: on deicticum: pat adj.synt.: 0
-> subst. : cupe
quant i f ier: twegen deicticum: pa adj.synt.: 0
-> subst.: kyngas
There are also substantival syntagms with elements in three peripheral positions.
all
Example; his gode knißtes two ( K H 4 9 ) =his two brave knights Analysis: quantifier: two deicticum: his adj.synt.: gode
subst.
Since apart from spelling differences and occasional inflected forms which still occur in EME, the forms are the same as in PDE, it is not necessary to enumerate the individual items which belong to the morphological classes demonstrative pronoun, article, and possessive pronoun. The situation is slightly different for the indefinite pronouns. This class comprises the elements no (variant n o n ) , a l ( l e ) , ani, nanij, s u m ( e ) , ech, fela, mani. Unlike in PDE, this last element combines with singular and plural forms of substantives. When it expands singular substantives, it has the variant mani a.^ Examples; moni wummon (ON1393) =many a woman mani wisdom (ON1756) =many a wise judgement song mani eine (ON759) =many a song selcudes manie ( B 4 4 1 ) =many wonders Mustanoja ( 1 9 6 0 : 2 1 6 ) also includes ober in his list of indefinite variants are still current in Modern High German. The EME syntagm moni wummon could be translated as manche Frau or manch eine Frau. In PDE only many a survives as an expansion of singular substantives.
63
pronouns. This may be justified from a morphological point of view, but the syntactic properties of ober differ from those of the elements mentioned before. It cannot realize the peripheral position 'quantifier', because it does not exclude an expansion by another element in this position. Examples; non ober heuene (FB553) =no other heaven eni ober note (ON557) =any other usefulness sumne oderne mon (BH30,27) =some other man fif 7 twenti odre cnihtes ( P C 1 1 2 4 , 1 1 f . ) =25 other warriors As ober also combines with elements in the positions 'deicticum' and 'adjectival syntagm', we must conclude that it realizes a separate peripheral position of its own. Examples; deo oder acennednysse (BH4, 1 5 f . ) =the other birth pat oper ping (ON784) =that other thing oper wißte gente & smale (ON204) =other smart and graceful creatures Adopting the terminology used in GCE, I shall call this position Ordinal". ^ The enlarged model of the substantival syntagm is now quantifier deicticum -> substantive, ordinal adj.synt. This version of the model is powerful enough for the analysis of the examples quoted above.
13
cf. GCE §4.22. As in PDE, this position can also be realized by ordinal numbers, e.g. be pridde day (FB430) 'the third day' = On the third day'.
64
Analysis: quantifier: 0 deicticum: deo -> subst.: acennednysse
ordinal: oder adj.synt.: 0 quantifier: 0 deicticum: pat
-> subst.: ping
ordinal: oper adj.synt.: 0 quantifier: non deicticum: 0
-> subst.: heuene
ordinal: oper adj.synt.: 0 quantifier: eni deicticura: 0
-> subst.: note
ordinal: oper adj.synt.: 0 quantifier: sumne deicticum: 0 ordinal: oderne adj.synt.: 0 quantifier: fif deicticum: 0
7
-> subst. : mon
twenti
ordinal: odre adj.synt.: 0 quantifier: 0 deicticum: 0 ordinal: oper adj.synt.: gente & smale
-> subst.: cnihtes
-> subst.: wijte
It can also be used for the analysis of more complex constructions. Examples; an oper clene stede (ON590) =another clean place
65
his an oder castel ( P C 1 1 2 4 , 6 ) =his one other castle mani an ober direwerpe ston (FB289) many an other precious stone =many other precious stones fela odre godre cnihtes ( P C 1 1 2 4 , 1 1 f . ) =many other valiant warriors Analysis: quantifier: an deicticum: 0 -> subst. : stede ordinal: oper adj.synt.: clene quantifier: an deicticum: his -> subst.: castle ordinal: oder adj.synt.: 0 quantifier: mani an deicticum: 0 -> subst. : ston ordinal: oper adj.synt.: direwerpe quantifier: fela deicticum: 0 ordinal: odre adj.synt.: godre
-> subst.: cnihtes
The elements which realize the peripheral positions of the model of the substantival syntagm developed so far usually precede the nuclus of the construction. They do not exclude expansions which follow the nucleus. Examples: bo brie kinges of hepenesse ( K S 2 6 , 7 ) the three kings of the orient =the three Magi
insißt in eche songe ( O N I 9 4 ) insight into each song =insight into every song one pipe of one smale wode unripe ( O N 3 1 9 f . )
a pipe of a thin wood unripe
=a pipe made of a piece of thin green wood
The relation between the substantival nuclei and the syntagms
66
which follow them is easily established. As there is unilateral occurrence dependency between the two constituents, the tactic relation between them is subordination. The characteristic element of this kind of expansion is a preposition.^ The constituents which realize it need not always follow the nucleus; the metrical structure of verse texts, for example, may require a different arrangement of the ICs. Example; of operefs] songe ... plaiding supe stronge ( O N H f . ) of [the] other's song ... plea very fierce =a very fierce debate about each other's singing Post-nuclear expansions of substantival syntagms can also be realized by other constructions. The following syntagms illustrate some of the possibilities. Examples; pe louerd pat ous hauep wrout (DS112) the Lord that us has wrought =the Lord who made us o dai pat he ferde to wude (KH938f.) a day that he went into [the] wood =a day when he went into the wood time for to gonge (HK1739) =time to go pa adas ... his dohter pam Casere to gifene (PC1109,4f.) the oaths ... his daughter to-the emperor to give =the oaths to give his daughter to the emperor pi wile ... pat pu me kni 3 ti woldest ( K H 6 4 3 f . ) your will ... that you me knight would =your intention to knight me bi-liaue bet he was diadlich pet diath solde suffri for man-ken ( K S 2 7 , 2 7 f . ) belief that he was mortal that death should suffer for mankind =the belief that he was mortal, that he should suffer death for mankind ore uaste picke hegge Imeind mid spire & grene segge (ON17) =an impenetrable thick hedge interspersed with reeds and green sedge trou wel grete .. . Mid iui grene al bigrowe (ON615ff.) tree very big ... with ivy green all overgrown =a very big tree all overgrown with ivy
the description of the 'prepositional syntagm' cf. Chapt.IV. 1
67
Considering the structure of the PDE substantival syntagm, we expect expansions with finite and others with non-finite verb forms. Among the expansions with finite verb forms, relative clauses are the most frequent in English of all periods.^ In EME relative clause expansions do not exclude expansions by prepositional syntagms. Examples; se king of gyus pet was i-bore ( K S 2 6 , 1 4 f . ) the king of Jews that was born =the king of the Jews who was born si glorius seywinge of ure lordes beringe bet us telp pet holi godespel of te day ( K S 2 7 , 1 6 f . ) 1 6 the glorious showing of our Lord's birth that us tells the holy gospel of the day =the glorious heralding of our Lord's birth which the holy gospel of today tells us The co-occurrence of these two kinds of expansions makes us enlarge the model of the substantival syntagm by two more positions. They will be labelled 'prepositional supplement' and 'relative supplement'; the generalized form of the model is now quantifier deicticum ordinal adjectival syntagm prepositional supplement relative supplement
-> substantive,
It allows the analysis of some of the examples quoted above. Analysis: quantifier: prie deicticum: po ordinal: 0 adj.synt.: 0 prep.suppl.: of hebenesse rel.suppl.: 0
15 16
-> subst.: kinges
For the description of this syntactic type cf. Chapt.VII.2.
The syntagm is ambiguous; the relative clause can be interpreted as an expansion of the nucleus seywinge or of the nucleus beringe.
66
quantifier: 0 deicticum: Ο ordinal: Ο subst. adj.synt.: Ο prep.suppl.: in eche songe rel.suppl.: 0 quantifier: 0 deicticum: one ordinal: Ο -> subst.: pipe adj.synt.: Ο prep.suppl.: of one smale wode unripe rel.suppl.: 0 quantifier: 0 deicticum: 0 ordinal: 0 adj.synt.: supe stronge prep.suppl.: of operefs] songe rel.suppl.: 0
-> subst.: plaiding
quantifier: 0 deicticum: pe ordinal: 0 adj.synt.: 0 prep.suppl.: 0 rel.suppl.: pat ous hauep wrout
-> subst.: louerd
quantifier: 0 deicticum: 0 ordinal: 0 -> subst. : dai adj.synt.: 0 prep.suppl.: 0 rel.suppl.: pat he ferde to wude quantifier: 0 deicticum: se ordinal: 0 adj.synt.: 0 prep.suppl.: of gyus rel.suppl.: pet was i-bore
-> subst.: king
69
quantifier: 0 deieticum: si ordinal: Ο adj.synt.: glorius
-> subst.: seywinge
prep.suppl.: of ure lordes beringe rel.suppl.: pet us telp pet hol! godespel The PDE equivalent of the other expansion with a finite verb is called 'appositive clause' in GCE. 1 ^ Whereas in PDE introductory that is obligatory in appositive clauses, EME allows other conjunctions as well, such as hu 'how 1 . Examples; a strong raed pat Riwald kinge iwerd dead (LB3910f.) =a hard fate that king Riwald died tidende ... hu pe king Rodric his raeflac makede (LB9936ff,) =tidings that King Rodric was ravaging Appositive clauses commute with appositive infinitive constructions and do not exclude expansions by prepositional syntagms or relative clauses. Examples; tidende pat him waes saer pat icumen wes Julius Cezar (LB7418f.) tidings that him were unpleasant that come was Julius Caesar =news which was unpleasant to him, that Julius Caesar had arrived Jpst word ... p«t was widene cud paet pe king Latin •jef Lauine his douter Eneam to are brude ( L B l 6 0 f f . ) that word ... that was widely known that the king Latin gave Lavine his daughter to-Eneas as a bride =the news which was widely known that King Latin betrothed his daughter Lavine to Eneas The consequence of this situation is that a new peripheral position, which will be labelled 'appositive supplement 1 , must be integrated into the model of the substantival syntagm. Its generalized form is then
17
cf. S13.16f.
70
quantifier
deicticum ordinal adjectival syntagm prepositional supplement relative supplement appositive supplement
-> substantive.
This now allows us to analyse yet more examples used for the lustration of syntagms in 'supplement' positions."* ^
il-
Analysis: The 'appositive supplement1 position is realized by infinitive constructions: app.suppl.: for to gonge deicticum: pa app.suppl.: his dohter pam Casere to gifene
> subst.: time
-> subst.: aoas
The 'appositive supplement 1 is realized by an appositive clause:
deicticum: pi app.suppl.: pat pu me knijti woldest
-> subst.: wile
deicticum: a adj.synt.: strong -> subst. : raed app.suppl.: bat Riwald kinge iwerd dead app.suppl.: hu pe king Rodric his raeflac makede rel. suppl.: pat him waes saer app.suppl.: pat icumen wes Julius Cezar
18
-> subst.: tidende
-> subst.: tidende
From now on empty positions are omitted in the analyses.
71
deicticuin: paet rel.suppl.: pat was widene cud app.suppl.: paet pe king Latin Lauine his douter Eneam to are brude
-> subst.iword
The 'appositive supplement' is realized by two coordinated appositive clauses: app.suppl.: pet he was diadlich -> subst.:bi-liaue pet diath solde suffri for man-ken In PDE relative clause expansions commute with two kinds of participial constructions; the characteristic element of one is a present participle, that of the other, the past participle of a lexical verb, e.g. a famous poet living in Paris, a famous poem written after the Second World War. Substantival syntagms with past participle constructions as postmodifiers are recorded from the OE period onwards. 1 ^ It is therefore not surprising that my EME data provide a substantial number of syntagms which illustrate this postmodifier. Like relative expansions they combine freely with expansions in 'prepositional supplement' position.^0 Example; an god flocc of prestess, Sprungenn off himm, strenedd burrh himm ( O R 5 1 0 f . ) =a good flock of priests, descended from him, begotten by him Visser (1963-73:§1010) also quotes examples with present participle constructions in peripheral positions of substantival syntagms. However, most of his examples illustrate present participles rather than present participle constructions. Both types of expansions are recorded for EME, too - neither very frequently. Present participles realize the same position as adjectives, namely that of 'adjectival syntagm'. Like adjectives they precede the substantival nucleus. Present participle constructions, on the 19
cf. Visser (1963-73:§1141) the description of participial constructions cf. Chapt.VII.3.
72
other hand, follow the substantival nucleus; they commute with relative expansions.21 Examples; mani wepinde eie (FB742) =many a weeping eye Lazarus ... fule pa stincende ( B H 2 6 , 4 f . ) =Lazarus stinking then foully pe children ... Wringinde here hende ( K H 1 1 1 f . ) =the children wringing their hands Apart from ordinary relative clauses the 'relative supplement1 position can thus also be realized by participial constructions. A fourth syntactic type which realizes the same peripheral position is the infinitive construction without subject.22 Example; faire gomes Wip him for to pleie ( K H 2 2 f . ) =noble youths to keep him company This construction occurs also in a reduced form, with one of nominal constituents deleted.
its
Example; scheid To fi^te wib vpon pe feld (KH513f.) =a shield to fight with on the battle-field Consequently, for the analysis of the substantival syntagms not yet presented the model need not be modified. It has simply to be born in mind that the label 'relative supplement1 must not be identified with 'relative clause in peripheral position', but that this position can be realized by a variety of constructions. The application of the model to the remaining examples yields the following analyses: deicticum: ore adj.synt.: uaste jpicke rel.suppl.: Imeind mid spire & grene segge adj.synt.: wel
-> grete
-> subst.:hegge
-> subst.: trou
rel.suppl.: Mid iui grene al bigrowe
21
cf.fn.20 syntactic type is described in Chapt.VII.1.
73
deicticuin: an adj.synt.: god prep.suppl.: of prestess rel.suppl.: Sprüngen of him strenedd purrh hinun quantifier: mani
-> subst.: flocc
-> subst.: eie
adj.synt.: wepinde rel.suppl.: fule pa stincende deicticuin: pe rel.suppl.: Wringinde here hende
> subst.: Lazarus
-> subst.: children
rel.suppl.: To fißte wib vpon pe feld adj. synt.: faire rel.suppl.: Wip him for to pleie
> subst.: scheid
-> subst.: gomes
2. The Adjectival Syntagm The label 'adjectival syntagm1 is motivated by the nucleus of the construction, which is realized by an adjective. Adjectives may be expanded in various ways. The most frequent expansion is by adverbs. Examples; to long (ON343) =too long swide fagen (B415) =very glad ful brihte (PM75) =very bright Most of these adverbs have an intensifying meaning, e.g. to 'too', swide, wel 'very', f u l , al 'completely 1 , swa, pus 'so', iliche 'equally', and for this reason the peripheral position they realize in the model of the adjectival syntagm will be called 'intensifier'. The adverbs in this position can themselves be expanded.
74
Example; al to wlonc (ON489) =all too wanton On principle, adverbs in any position are expandable by other adverbs. This recursiveness would theoretically allow an unlimited sequence of adverbs in the 'intensifier' position. In the most complex syntagm of this type in my data, the adjective is expanded by an 1C which consists of three adverbs. Example; swa swide ungemetlice mycel wind (PC1118,16) =a so very terribly strong wind The generalized form of the model of the adjectival syntagm which accounts for the syntactic structure of the above examples is intensifier
> adjective.
Its application to the adjectival syntagms quoted so far yields the following analyses: intensifier: to > adjective: long intensifier: swide > adjective: fagen intensifier: ful > adjective: brihte intensifier: (al > to) > adjective: wlonc intens.: (swa > (swide > ungemetlice))
> adj.: mycel
Very often adjectives are nuclei of comparative constructions, in which case the adjective occurs in its comparative form, i.e. its stem is followed by the comparative affix -er. There is unilateral occurrence dependency between this affix as nucleus and several other ICs, e.g. the particle £e 'the' (as in PDE the more, the better) and constructions which are introduced by panne 'than' (as in PDE purer than gold). 2 3 The affix -er itself is unilaterally occurrence-dependent on an adjective, and the comparative affix is therefore to be analysed as an expansion of the adjectival nucleus as well as the nucleus of expansions like b_e and banne-constructions. 24 ^constructions of this type are described in Chapt.IV.3. ('The Comparative Syntagm'). 24
My treatment of comparative-constructions differs from that Mulder proposed for PDE. He considers the comparative affix (or its equivalent more) and the particle than together as one, discontinuous, 1C. This analysis is based on the fallacious assump-
75
Examples; wisure pane he (ON1250) =wiser than he betere ... panne atter imaingd mid wine (PM144) =better than poison mixed with wine be gladur (ON19) =the gladder As the same adjective can be preceded by an adverb and followed by the comparative a f f i x , it might seem as if the two constituents realized different peripheral positions in the adjectival syntagm. This hypothesis has to be refuted, however, because in constructions of this type the adverb is not an expansion of the adjective, but of the comparative a f f i x . Examples; hundredfealde hotter panne is vre (PM249) hundredfold hotter than is ours =a hundred times hotter than ours muchele more and betere panne alle obre binges · (PM388) =much bigger and better than all other things Thus the model of the adjectival syntagm is also able to account for the analysis of the constructions with adjectival nuclei which are expanded by comparative constructions. Analysis:^ intens.: (-er intens.: (-er
-> adj . : wis pane he] panne atter imaingd mid wine) > a d j . : god be) > a d j . : glad intens.: (-er
-er
-> adj .: hot
I panne is vre
tion of a bilateral occurrence dependency between the two elements: "if we take the item more as basic, the item than is compulsory, and, therefore, not functional, and if we take than as basic, more is not functional" (Mulder 1980:169). In PDE, just as in EME, tHe occurrence of the comparative a f f i x (or more) does not depend on the occurrence of a than/panne-construction , cf. m the gayer ever more ' . ich am be blibur euer more ( ON1 1 08 ) = ' I am 25
The comparative a f f i x is represented in a 'normalized 1 form. The arrangement of the ICs in the analyses is irrelevant; as far as possible, it mirrors the arrangement in the data.
76
/ muchele intens.:I I panne alle opre pinges
-> -er
> a d j . : (muchel
a d j . : unorne > adj.: long
When an intensifying adverb and a negation particle co-occur in front of the same adjective, the negation particle either modifies only the adverb or a complex nucleus consisting of the adverb and its adjectival nucleus, cf. PDE not very long. Depending on its meaning, this syntagm has to be analysed as ((not > very) > long) or (not > (very > long)). Similarly, a negation particle which precedes the comparative form of an adjective either expands the comparative affix alone or it expands the complex nucleus which consists of the adjectival nucleus and the comparative affix. The first alternative seems to be appropriate in non-contrastive (=neutral) environments. Example; nowiht lesse (PM214) =not less Analysis: intens.: (nowiht
> -er)
> adj.: lytel
Adjectives can also be expanded by prepositional syntagms. 2 ? Such expansions do not exclude an element in 'intensifier 1 position and the model of the adjectival syntagm must be enlarged by comparatives more and betere belong to suppletive paradigms; the corresponding positive forms are muchel and god. 2
^This syntactic type is described in Chapt.IV.1.
77
one peripheral position, called here 'supplement'. "Prepositional supplement' would be too specific a label for this, suggesting that only prepositional syntagms could occupy this position, which is not the case. Pronominal syntagms, substantival syntagms, genitival syntagms, infinitive constructions and pat-clauses all commute with prepositional syntagms in this position. Examples; icundur to one frogge (ON85) =more natural to a frog to gredi of selvere and of golde (PM266) =too greedy for silver and gold oper vniliche (PM358) =unlike the other a meiden ilike ( B 4 4 2 ) =resembling a girl eches godes ful (PM369) =full of every good thing wurdi to hauen same (B346) worthy to have shame =deserving shame iwar bat ho song hire a bisemar ( O N 1 4 7 f . ) aware that she sang her a mockery =aware of her singing to mock her The elements in 'intensifier' position and in 'supplement 1 position do not determine the adjectival nucleus in noticeably different ways. Consequently, the complex tactic relation in the adjectival syntagm is parallel determination and the model of the construction can be represented as intensifier supplement
-> adjective.
This model accounts for the syntactic structure of all presented so far. Analysis: intens.: -er suppl.: to one frogge
examples
-> adj . : icunde
intens.: to suppl.: of selvere and of golde
-> ad j . : gredi
78
intens.: 0 -> adj . : vniliche suppl.: oper intens.: 0
-> adj. : ilike
suppl.: a meiden intens.: 0
-> a d j . : ful suppl.: eches godes intens.: 0 suppl.: to hauen same
-> adj . : wurdi
intens.: 0 -> adj . : iwar suppl.: pat ho song hire a bisemar
3. The Pronominal Syntagm In the description of the substantival syntagm it became evident that some pronouns (e.g. possessive, demonstrative, indefinite pronouns) can function as expansions of substantival nuclei. In this function the pronouns can themselves be expanded. Al- though these expansions have an intensifying meaning comparable to that of the elements in the peripheral position ' intensif ier 1 of the adjectival syntagm, the peripheral position in the pronominal syntagm requires a different label, because there is only partial overlap between the elements which realize the two functions. Moreover, in the pronominal syntagm the choice of possible expansions is governed by the morphological class of the nucleus. Possessive pronouns can be expanded by e Own1 , demonstrative pronouns by ilke 'same 1 , and indefinite pronouns by elements like to " t o o 1 , so ' s o ' . Examples ; pinoße dep (FB200) =your own death heora agene men ( P C 1 0 7 0 , 9 f . ) =their own men
79
on pam ilcan geare (PC1070,3) =in that same year pat like mon (FB251) =that same man so manie flures (FB442) =so many flowers to monie tale (ON257) =too many tales The label 'emphasizer' has been chosen for the peripheral position in the model of the pronominal syntagm on the basis of its mnemonic value. Its nuclear position is called 'pronoun', because it can be realized by a number of different pronouns. The general form of the model is then emphasizer
> pronoun.
This model accounts for the syntactic structure of the pronominal syntagms of the above examples. Analysis: emphasizer: emphasizer: emphasizer: emphasizer: emphasizer: emphasizer:
036 agene ilcan ilke so to
> pronoun: pin > pronoun: heora > pronoun: panr^ > pronoun: pat > pronoun: manie > pronoun: monie
If pronominal syntagms are not expansions of substantival syntagms, their potential expansions are not restricted to 'emphasizer 1 elements; hence the model requires an additional peripheral position. The following examples illustrate some of the realization possibilities. Examples; nucleus = personal pronoun we ... sylfe (PC1106,18) =we ourselves vvijp horn pat ne mu e from pe Schilde (ON62) against them that not may from you protect =against those who cannot protect themselves from you heo sat on pe sunne, WijD tieres al birunne (KH653f.) =completely covered with tears she sat in the sun form may either be analysed as an inflected form of the demonstrative pronoun pat or of the definite article. The nuclear position of the model of the pronominal syntagm can also be real-ized by the definite article, e.g. bilke day (FB371) = ' o n the same day'.
80
nucleus = demonstrative pronoun se of France (PC1120,2) that of France =the king of France se ... pet sent hi to heueriche (PM42) that ... that sends it to heaven =the person who sends it to heaven nucleus = indefinite pronoun eall pet he mihte (PC1070,15) all that he might =all he could to feawe of para folce ( B H 2 8 , 2 1 ) =too few of that people feole to fordeme in schynynde wede (SC18) many to condemn in shining weeds =many who are going to be condemned in shining garments In addition to expansions by means of 'emphasizer' elements, pronouns can be expanded by relative clauses and by participial constructions which commute with relative clauses. The syntactic structure of these expansions is identical with that of the corresponding expansions in the 'relative supplement' position of the substantival syntagm and thus it is appropriate to use the same label for the peripheral position in the pronominal syntagm, too.29 AS 'emphasizer' expansions and 'relative supplement' expansions do not exclude each other, the model of the pronominal syntagm with personal pronouns in nuclear position can be represented as emphasizer -> pronoun. relative supplement It allows the following analyses:
^Strictly speaking, the same label should only be used if exactly the same realizations were possible in the two syntagms. There is indeed reason to believe that the investigation of further texts will provide the missing examples, i.e. participial constructions and infinitive constructions without subject in the 'relative supplement' position of pronominal syntagms.
81
emphas.: sylfe -> pron.: we rel.suppl.: 0 emphas.: 0 rel.suppl.: pat ne
-> pron.: horn from pe schilde
emphas.: 0
-> pron.: heo
rel.suppl.: Wip tieres al birunne emphas.: 0
-> pron. : se
rel.suppl.: pet sent hi to heueriche emphas.: 0
-> pron. : eall
rel.suppl.: pet he mihte Unlike in PDE, relative expansions are not restricted to pronominal nuclei which are not expansions. A case in point are possessive pronouns which realize the 'deicticum 1 position in substantival syntagms. Examples; pat spek in chirche pat nes no god To his mupe takep be flod (VP109f.) that spoke in church that not-was good to his mouth goes the flood =the flodd mounts to the mouths of those who spoke ill in church his lif and his soule worpe I-schend pat pe to me
pis herende hauep send (DS213f.)
his life and his soul be put-to-shame who to me this errand has sent =that person's life and soul ought to be put to shame, who sent me this errand The pronominal nuclei are his in both examples; they are expansions in the substantival syntagms his mupe ( V P l 0 9 f , ) and in D S ( 2 1 3 f . ) his lif and his soule, respectively. Analysis: emphas.: 0 rel.suppl.: pat spek in chirche pat nes god
-> pron. : his
82
emphas.: Ο -> pron.: his
rel.suppl.: pat be to me pis herende hauep send
his
A second, more elaborate model has to be developed to account for the syntactic structure of pronominal syntagms with a demonstrative or an indefinite pronoun in nuclear position and which are not expansions in substantival syntagms. These nuclei can be expanded not only by syntagms which realize the 'relative supplement1 position (e.g. relative clauses, infinitive constructions without subject), but also by prepositional syntagms. Thus the model must also contain a 'prepositional supplement 1 position, and its generalized form is emphasizer relative supplement prepositional supplement
-> pronoun.
Its application to the relevant examples yields the following analyses: emphas.: 0 rel.suppl.: 0 prep.suppl.: of France emphas.: to rel.suppl.: 0 prep.suppl.: of pam folce
-> pron.: se
-> pron.: feawe
emphas.: 0 rel.suppl.: to fordeme prep.suppl.: in schynynde wede
-> pron.: feole
The early prose texts Old Kentish Sermons and The Peterborough Chronicle show traces of the OE construction with an indefinite pronoun in nuclear position which is expanded by the genitive of a personal or a deictic/demonstrative pronoun. Examples; here eurich ( K S 3 4 , 1 ) =each of them heora aelc (PC1087,15) =each of them
83
sum p*ra ( P C 1 0 9 7 , 1 2 f . ) =one of them These syntagms cannot be analysed with any of the models presented above. Their expansions realize a separate peripheral position which might be called 'partitive'. The construction was already disappearing in EME; it was replaced either by the coordination of two pronouns ( e . g . hi sume (PC1101,8) ='some of them 1 ; pet ... eall (PC1070,19f.) = ' a l l t h i s ' ) or by pronominal syntagms with an expansion in 'prepositional supplement 1 position. It is, therefore, simpler and more appropriate to describe this archaic construction separately rather than to burden the model of the pronominal syntagm with an additional peripheral position. A further simplification of the description of the pronominal syntagm is to be expected from a closer investigation of the syntax of the personal and possessive pronouns. As soon as examples appear in which these pronouns are expanded by prepositional syntagms, the model with the three peripheral positions 'emphasizer 1 , 'relative supplement', and 'prepositional supplement1 will account for all occurrences of the construction, irrespective of the realization of its nuclear position.
IV.
FUNCTIONAL SYNTAGMS
The syntactic function of a linguistic entity can be indicated by several means, such as wordorder or functional markers. For markers which indicate the syntactic function of constituents of an utterance, Martinet coined the term "monemes fonctionnels'. Mulder uses the more general label 'functional' for such syntactic entities. Functionals are the nuclei of "functional syntagms'. The label 'functional syntagm' is thus a cover-term for a whole class of syntagms with the common property that their syntactic function is indicated by a special plereme. The classification of the functional syntagms of EME presented here is based on the nature of this plereme. It is also this plereme which determines the potential realizations of the 1C in the peripheral position.
1. The Prepositional Syntagm The prepositional syntagm consists of two ICs. One of them is a preposition, while the other may be realized by different constructions, the most frequent being that of a substantival syntagm. The two ICs are mutually occurrence-dependent and hence the syntactic relation between them is either that of subordination or of interordination. The following syntagm is used here to demonstrate that it is the preposition which indicates the syntactic function of the prepositional syntagm. Example; pet holi godspel of to day us telp pet a bredale was i-maked ine bo londe of ierusalem ( K S 2 9 , 8 f . ) =the Holy Gospel of today tells us that a wedding was celebrated in the land of Jerusalem The first prepositional syntagm of the example is of to day. The adverb to day 'today' can have different syntactic functions; very frequently it is used as an adverbial complement (of a clause).
85
Here the preposition of indicates that the prepositional syntagm of to day is an expansion of the substantival nucleus godspel. Similarly, the preposition ine indicates that the prepositional syntagm ine fro londe of Jerusalem is an adverbial complement of the clause a bredale was i-maked. Thus the preposition is the syntactically dominant part of the construction, and the relation between its ICs is subordination, with the preposition as nucleus and the other 1C as actualiser. The general form of the model of the prepositional syntagm may be symbolized as preposition
pred.nucl.: habb-
The second example is syntactically ambiguous. The relative construction occupies the relative supplement position in the substantival syntagm with the nucleus men. The peripheral position of the relative construction is realized by a clause with the nucleus wenden pat ani god hefden. The conjunctional syntagm which expands the predicative nucleus wenden as object consists of two ICs which are either the conjunction pat and the predicative syntagm ani god hefden or the conjunction pat and the clause pe ... ani god hefden. The first analysis implies that the example illustrates a relative construction without functional amalgamation and without a relative push-down element, while the second analysis implies that both phenomena are present. The second syntactic structure is mapped in the sequence of analytical steps below.
168
Analysis: deicticum: pa rel.suppl.: pe hi wenden pat ani god hefden relat.: pe subst.: men
rel.compl.(clause): hi wenden bat ani god hefden
subj.: hi adv.compl.: 0
> p r e d . s y n t . ( t y p e ( 9 ) ) : wenden pat ani god hefden
obj.: pat ani god hefden conj.: pat
pred.nucl.: wenden
conj.compl.(clause): ani god hefden
subj . : & ?//
adv.compl. 0
> p r e d . s y n t . ( t y p e ( 9 ) ) : ani god hefden
3. Participial Constructions Unlike most of the inflected forms of the English verb, the present and past participles have survived until the present day. In EME both can realize the nuclear position in the active verbal syntagm as well as a peripheral position in the substantival syntagm. In the latter function they commute with adjectival syntagms and, like adjectival nuclei, participles can be expanded by adverbs. Participial constructions, however, can be described as special forms of predicative syntagms or clauses, as can infinitive and relative constructions. Like these they do not have the normal function of predicative syntagms or clauses, which is to form independent syntagms, but instead usuallly realize the peripheral position 'relative supplement' in substantival and pronominal syntagms, or the function 'direct object' in predicative syntagms. This functional shift is brought about by the participle forms of the verbs.^ The participle endings must therefore be ^Theoretically, the participles of lexical and copulative verbs are possible in this function. The data examined so far, however, contain only participles of lexical verbs.
169
analysed as the elements which indicate the function of participial constructions. The nuclear position which they realize will be called 'participle' in the descriptive model, the peripheral position, 'participial complement'. Generally speaking, the peripheral position can be realized by a special form of a predicative syntagm or a clause. The nucleus of the predicative syntagm, which may be the nucleus of a clause, is not realized by an active verbal or by a passive syntagm, but only by a verbal stem. If the predicative syntagm is expanded by a subject, the latter is not in the subject case, but in the object case. This applies, of course, only to subjects which are realized by a personal pronoun. This special case form of the subject is a feature which participial constructions share with infinitive constructions with subject, and the it is accounted for by the functional in both constructions. 7 The type of predicative syntagm or clause involved depends on the form of the participle (present or past) and on the syntactic properties of the lexical verb the participle belongs to. If it is a present participle, its stem can be the nucleus of predicative syntagms of types ( 2 ) , ( 8 ) , ( 9 ) , ( 1 0 ) or ( 1 1 ) ; if it is a past participle, its stem can be the nucleus of predicative syntagms of types ( 2 ) , (3), (6) or ( 7 ) . The generalized form of the model of the construction can be represented as participle