The Construction ἀπὸ κοινοῦ in the Germanic Languages 0404518117, 9780404518110

Table of contents :
Chapter I. Introduction 7
Chapter II. ἀπὸ κοινοῦ in Old Germanic Poetry 20
Chapter III. ἀπὸ κοινοῦ in Middle English and Middle Dutch 55
Chapter IV. ἀπὸ κοινοῦ in Old Germanic Prose 81
Chapter V. Constructions Similar to ἀπὸ κοινοῦ but with Distinctly Subordinate Second Part 90
Chapter VI. The Origin of the Construction ἀπὸ κοινοῦ in Germanic 110
Bibliography 113

Citation preview

STANFORD

UNIVERSITY UNIVERSITY

LANGUAGE AND

PUBLICATIONS

SERIES

LITERATURE

V olume VI

N umber 2

T he Construction από κοινού in the Germanic Languages

By

HERBERT D E A N MERITT, Ph.D. (Princeton)

,

Assistant Professor o f English Stanford University

STA N FO R D U N IV E R SIT Y PRESS ST A N FO R D U N IV E R S IT Y , C A L IFO R N IA L O N D O N : H U M P H R E Y M ILFORD OXFORD U N IV E R S IT Y P R E SS 1938

S T A N F O R D U N IV E R S IT Y PR ESS STANFORD UNIVERSITY, CALIFORNIA OXFORD UNIVER SITY PRESS LONDON: HUMPHREY MILFORD TH E BAKER AND TAYLOR COMPANY 55

FIFTH AVENUE, NEW YORK TH E MARUZEN COMPANY

TOKYO, OSAKA, KYOTO, SENDAI

COPYRIGHT I 9 3 8

BY THE BOARD OF TRUSTEES

OF TH E LELAND STANFORD JUNIOR UNIVERSITY

Published December 1938 PRINTED IN TH E UNITED STATES OF AMERICA BY STANFORD UNIVERSITY PRESS

PREFACE This study was presented as a dissertation at Princeton University in 1931 and was accepted by the Department of Oriental Languages and Literatures. Since then I have worked through it at various available times and have added some parts and rewritten others. To Professor Harold H. Bender I am especially grateful for his advice and interest. To Professor Robert J. Menner and to Professor Arthur G. Kennedy I am indebted for several helpful suggestions. A discussion of από κοινού is in the main a quite technical subject, but not necessarily one which is difficult to follow. If at first it should seem to be such, the reason lies perhaps more in the name than in fact. The term is probably not so commonly used as other syntactical terms such as ellipsis, asyndeton, and fusion; but as a feature of language it is deserving of equal recognition. And in some respects it is not technical at all; for probably every educated person manifests at one time or another some curiosity about a particular turn of word or phrase which he uses or hears, and if he sought to describe it he might on occasion find that it could be called από κοινού. In general the term applies to that feature of language wherein it seems necessary to understand a word twice although it is expressed but once. The second edition of Webster's N ew International Dictionary gives "apo koinou” and defines it as “a syntactical construction in which a subject or object in one clause stands also, without repetition, in a case relationship in a following clause, as in 'then sprang from his seat Hagen thus spoke’ ( Gudrun, 538).” With this definition mine (p. 16) has no conflict, although mine is in some respects broader and in others imposes filrther conditions. The use of the form "apo koinou” with Latin spelling, rather than with Greek spelling, seems increasingly desirable and it may well eventually fall in with the English pattern as a single word with the accent on the first syllable as it is now sometimes pronounced, although the dictionary gives "apo' koinou'.” Since, however, statements containing the Greek form of the term are quoted here and there throughout this book, it has seemed best in this instance, by keeping the same form throughout, to guard against a certain amount of confusion to which the appearance of the two forms might give rise. This present treatment of the construction consists of a collection and analysis of examples from Old and Middle Germanic languages. Some have been pointed out by others; many have not. It is likely that everyone who has studied an early Germanic poem, such as the Old English Beowulf [157]

[3

4

THE CONSTRUCTION obtÓ XOIVOIJ IN THE GERMANIC LANGUAGES

or the Old High German poetry of Otfrid or the Old Saxon Heliand, has come upon some reference to από κοινού concerning a particular passage and has there thought about the likelihood of the construction. But in consid­ ering this likelihood one naturally wishes to study other examples. Isolated references cannot offer the helpful criteria which a collection and a sifting of examples from numerous poems provide. The abhorrence of anything “ungrammatical” in most modern writing tends to exclude the construction there. It would strike the eye as some­ how an error of omission, since it makes twofold use of a single locution. Such a syntactical short cut we are usually careful to avoid in writing, although it may not infrequently be heard in daily speech and is found in writing of earlier periods. The recognition of από κοινού in the older languages is of value in a number of ways. It enables one to understand text as it stands without feel­ ing that something has been omitted which must be editorially supplied. And applying to a somewhat similar tendency it eliminates some unneces­ sary punctuation which occasionally gives to an early Germanic poem a halting quality with which it was not actually afflicted. It makes unneces­ sary certain grammatical exceptions, such as holding a usually transitive verb to be intransitive because its apparent object is also object of another verb. Quite the opposite from being a burdensome bit of linguistic parapher­ nalia for the student, it helps to bring home to him the realization that the dull printed page before him was once the spoken word as ready on the tongue of man as is his own. The occurrence of the construction από κοινού, though limited, is suf­ ficiently frequent to merit a general understanding of the term. As one of the many vagaries which language exhibits, it is of interest to all. And to those readers of older texts who feel that in certain passages something is syntactically amiss, it is deserving of careful consideration as a fitting in­ terpretation. H. D. M. S tanford U niversity, California August 1, 1938

[158]

CONTENTS PAGE

Chapter

Chapter

I. I n t r o d u c t i o n .................................................... από κοινού a Widespread Syntactical Construction . A Résumé of Discussions of από κοινού in Germanic . An Interpretation of the Term από κοινού . The Plan of Discussion. . . .

7 7 10 16 18

II. από κοινού in O ld Germanic P oetry . . . . 20 A. The Common Element Constitutes a Half-Line . 20 B. A Possible Koinon Standing in a Half-Line with the First Part of the Construction and Relation of Such to A s y n d e t o n .........................................................36 C. The Common Element Standing in the Half-Line with the Second Part of the Construction . . . 38 D. . A Dependent Clause between Two Independent Clauses and an Independent Clause between Two Dependent Clauses . . . . . .4 9 Table of Citations . . . . 52

Chapter III. από κοινού in M iddle E nglish and M iddle D utch 55 I. Middle E n g lis h ............................................................. 55 A. The Position of Koinon in Middle English V e r s e ................................ 58 B. Grammatical Nature of the Koinon . 65 C. Context of the Construction . 67 II. από κοινοί in Middle D u t c h ................................... 69 III. Comparison with Middle High German and Old Germanic . . . . . . . · . . 75 Table of Citations . . . 79 Chapter IV. από κοινοί in O ld Germanic P rose . . 81 Infrequency of the Construction in Prose . .8 1 A. Gothic . . . . . . . .8 1 B. Old English . . . 84 C. Old N o r s e ..................................................................... 85 Comparison with Asyndetic Parataxis in Old High German ( T a t i a n ) ...............................................................86 Table of Citations . . . . . . 88 [159]

[5

Chapter

V. Constructions S imilar to από κοινού but with D istinctly S ubordinate S econd P art . . .

90

Relation to Other C o n s t r u c t io n s ..................................... 91 The Grammatical Nature of the Attributive-Clause Construction and Its Position in Verse . . . 95 A. Old High German . . 96 B. Old Norse . . . 96 C. Old English . .97 D. Middle Germanic 97 Review of Examples . .101 Table of Citations . . . . . . 108 Chapter VI. T he O rigin of the Construction από κοινού in Ge r m a n i c ..........................................................................110 B ibliography .

.113

[ 160]

CHAPTER I INTRO DUC TIO N από κοινού

a widespread syntactical construction

The terpi από κοινού has been applied broadly to constructions which are not entirely of a uniform type. The interpretation has varied, depend­ ing partly upon the particular language which happened to be under dis­ cussion.. Thus in regard to Latin the construction is defined in S tolzSchmalz Lateinische Grammatik: “Die Figura από κοινού besteht in der Setzung eines zu beiden Gliedern einer Verbindung gehörigen Wortes erst beim zweiten Gliede.”1 Tobler in his Vermischte Beiträge thus refers to the construction as it appears in Old French: “Die Redeweise, deren Eigentümlichkeit besteht in der gleichzeitigen Zugehörigkeit eines Rede­ stückes zu einem Satze, dessen Schluss und zu einem zweiten Satze, dessen Anfang es bildet............... ”2 Behaghel in his Deutsche Syntax defines the construction with reference to Middle High German: “Unter constructio από κοινού ist die Erscheinung zu verstehen, dass formal und sachlich gleichgeordnete Sätze oder Satzglieder, die nicht durch Konjunktion gebunden sind, ein Glied gemeinsam haben, das in der Mitte zwischen beiden Sätzen oder Satzgliedern steht und grammatisch sich sowohl mit dem vorausgehenden wie dem nachfolgenden Redeglied zur Einheit zusammenschliesst.”3 Although the interpretation of από κοινού varies somewhat in regard to any one branch of the Indo-European languages, in the main the construction to which the term is applied has one distinctive feature. It is a kind of verbal economy: a word or closely related group of words, expressed but a single time, serves at once a twofold grammatical function. And this form of expression is widely spread through Indo-European. The Sanskrit term for this construction is käkäksivat, ‘in the manner of a crow’s eye’—a rather pat term derived from the belief that a crow, while having but one eye, could nevertheless use it on either side of its head.4 A passage from the Mänavadharmagästra illustrates this construc­ tion in Sanskrit. 4, 83:

qirahsnätasca tailena nängam kimcidapi sprget ‘Having bathed his head with oil let him not touch his body'

1 Fifth edition, p. 848.

2 Adolph Tobler, Vermischte Beiträge, I, 115.

3 Otto Behaghel, Deutsche Syntax, III, 534. 4 In connection with από κοινού this term has been pointed out by Η. H. Bender, Journal of English and Germanic Philology, XI, 565.

[161]

[7

8

THE CONSTRUCTION dbtO κ ο ιν ο ί IN THE GERMANIC LANGUAGES

Kulluka’s commentary upon this passage, as cited in the Boehtlingk-Roth Sanskrit Dictionary under käkäksi, is as follow s: atha vä taileneti käkäksivadubhayatra sambhadyate. ‘Thus the word tailena is connected käkäksi vat with both sides’56

The term από κοινού as applied to syntax was first used by Apollonius Dyscolus in his De C onstruc tione.e He there gives as an example the fol­ lowing passage: Βοιωτών μέν Πηνέλεως καί Λήιτος ήρχον || Άρκεσίλαός τε Προϋοήνωρ τε Κλονίος τε,7

where ήρχον, he states, is to be taken as koinon. A construction in which the koinon is probably more keenly felt is Thucydides i. 2 6 .3 : οίκήτορά τε τον βουλόμενον ΐεναι κελέυοντες.8

Here Ϊεναι may stand από κοινού with the two participles between which it occurs. In the field of Latin syntax, the term από κοινού has been used broadly. The following citations have been pointed out as examples of the con­ struction : Liv. XXI . 3 0 .7 : pervias paucis esse exercitibus

Esse is taken από κοινοί. Liv.

x l v .5 .1 0 :

sed cum veneno se malle quam mori ferro dixisset9

M ori is taken από κοινοί. Horace C. 1 .13.4: fervens difficili bile turnet iecur

Difficili bile is taken as ablative of cause with both fervens and tumet. Horace C. 1 .3 .2 8 : ignem fraude mala gentibus intulit 5 Other references to this term in Sanskrit are given by Boehtlingk-Roth under

käkäksi. Also Lanman, in the notes to his translation of Räja-Gekhara’s KarpüraManjan, Harvard Oriental Series, IV, 264. 6 Grammatici Graeci, II, 2, pp. 170-71. 7 Iliad 11.494-95. 8 Cited by A. Biegler, De diversis quibus Graeci et Romani in dicendo usi sunt brevitatis generibus (Lissa, 1851), p. 18. 9 These last two citations are from W . Baehrens, “Beiträge zur Lateinischen Syntax,” Philologus, Supplementband, X II, 281 and 268.

[ 162]

INTRODUCTION

9

Gentibus is taken as dative of indirect object with intulit, dative of refer­ ence with mala. Horace C. 1 .30.2: sperno dilectam Cypran et vocantis | ture te multo Glycerae decoram | transfer in aedem10

Te is taken as direct object of vocantis and transfer, Cic. L eg. agr. 2.95 : ex hac copia atque omnium rerum affluentia11

Omnium rerum is taken in common with copia and affluentia. Tobler12 has cited a considerable number of examples of the construc­ tion in Old French. From his collection the following citations are taken: Marie de France G 313: Mes si vus piest que jeo vus die M’aventure vus cunterai13

M ’aventure may be taken in common with die and cunterai. Rich. 2900: Et Richars aquieut ces paiiens Ochist a milliers et a cens.14

Ces paiiens may be taken in common with aquieut and ochist. Bast. 1840: Chius est tant vertüeus c’on doit recommander Le poissance de lui a tous bons recorder15

Le poissance may be taken in common with recommander and recorder. Tobler also mentions an από κοινοί of sound and of syllable but he states that this may be merely synaloepha. For instance: nuts ne nule ne tent amender ( = a amender) son afaire.16 10 These citations from Horace are from J. C. M. Grimm, The Construction από κοινού in the Works of Horace (Philadelphia, 1928). 11 Cited from Stolz-Schmalz, Lateinische Grammatik (5th ed., München, 1928), p. 848. Here is also given a bibliography of treatments of the construction in Latin, 12 Adolph Tobler, Vermischte Beiträge, I, 115 ff. 13 Die Lais der Marie de France, ed. Warnke. 14 Richars li Biaus, ed. Foerster. 15 Li Bastars de Bullion, ed. Scheler. 16 Vermischte Beiträge, I, 187 note. Similar occurrences in Sophocles have been discussed by J. H. Wright, “Studies in Sophocles,” Harvard Studies in Classical Philology, X II, 137 ff. He finds that often, where a word seems to be omitted, a simi­ lar sound is found in a syllable of a word which is there expressed. For an analogous occurrence in Old English, cf. Anglia, X LIII, 311, sec. 154.

[ 163]

10

THE CONSTRUCTION

άπό κοίνοΰ IN

THE GERMANIC LANGUAGES

There seems to have been no mention of the construction από κοινού concerning the Balto-Slavonic languages, but as it is noted so widely one would expect to find it there. The following passage from a Lithuanian daina contains two constructions in which there is possibly a feeling of a common element. Nenulýdjkit margú skrýniu rasztélj nenudrukúsit bè Tilzés drukorélio

‘Let not the rain fade The fair chests' inscription You will not imprint Without the engraver of T f

Nenuláuzjkit margú skrýniu kojelés nepadarýsit bè |srúts diszérélio1718

‘Do you not break The feet of the fair chests You will not replace Without the Isrut artificer'

In the first verse, the accusative rasztéli stands between two verbs, is nec­ essarily the object of the first, and in thought is the object of the second verb. Similarly with kojelès in the second verse. Constructions referred to as από κοινού have long been noted in some of the Germanic languages.. It is with this branch of the Indo-European family that we are here specifically concerned. A RÉSUM É OF D ISC U SSIO N S OF ά π ό κ ο ι ν ο ί IN GERM ANIC

The most recent work concerning από κοινού in Germanic is a thorough treatment by Fritz Karg of the construction as it appears in Middle High German.1* The central theme of Karg’s work is a comparison of the άπό κοινού construction, such as do spranc von dem gesidele her Hagene also sprach, with the construction which he fitly calls the /wV^-Konstruktion; for example, die worhte ein smit hiez Volkan. In this comparison he investigates the rhythmical and syntactical nature of both constructions. He finds that, while in άπό κοινού in poetry the koinon often begins a line or fills a line and so is separated by a pause from the preceding part, in the /^’^-Konstruktion the assumed koinon often ends a line and hence seems to belong more closely with what precedes than with what fol­ lows. In the latter construction, too, he points out, there is often a diver­ gence of case in the assumed koinon; that is, it stands in an oblique case and as subject of the following clause would have to serve also as nomina­ tive. He further finds that in the speaking of a sentence containing άπό κοινού there is a sudden change in the tone of the voice when the 17 Cited from Wiedemann, Handbuch der Litauischen Sprache, p. 254. 18 Fritz Karg, "Die Konstruktion από κοινοί) im Mittelhochdeutschen," in his Syntaktische Studien (Halle, 1929). First published in Beiträge zur Geschichte der deutschen Sprache und Literatur, X LIX , 1-63.

[ 164]

INTRODUCTION

11

koinon is reached, but that this change does not occur in the enunciation of the /^^-Konstruktion. He distinguishes between complete από κοινού, in which each of the two parts forms a complete sentence, and incomplete από κοινού, where such is not the case; he holds, however, that these two types have essen­ tially the same structure. In regard to the context of the construction it is his observation that both parts usually express an action; these, for example, may stand in a temporal relationship to each other or from a situation conceived of as consisting of a group of actions any two may be chosen. In contrast he points out that the /^’^-Konstruktion prevail­ ingly denotes a quality or state of a person or thing and does not express an action. Where a sentence stands as a possible common element, he decides that this is not a true από κοινού form. In discussing the historical background of the construction he con­ siders a number of passages from the Heliand but concludes that there is no real από κοινού in that poem. The first real από κοινού he finds in the Exodus, about 1120; the first /zi^-Konstruktion in the Physiologus, about 1050. The construction από κοινού, he finds, is most noticeable in the thirteenth and fourteenth centuries, and thereafter is increasingly rare; while the /^^-Konstruktion, on the other hand, begins to flourish as από κοινού tends to disappear. The former, he notes, is rare in religious poetry and prose, but more common in folk poetry and poetry of the court, while the /wVs-Konstruktion is frequent in religious poetry and prose. Asyndeton and από κοινού are, he believes, distinct from each other.. In an asyndetic construction the word which might serve as koinon is in his estimation more closely attached to the first part of the construction and is followed by a pause. The reverse he finds to be characteristic of από κοινοί). Finally, Karg believes that the από κοινού construction in Middle High German literature was taken over from the habits of daily speech. He states: “Man denke etwa an einen Redner, der, wie man herkömmlich sagt, ‘aus der Konstruktion fällt.' Diese lapsus linguae werden in vielen Fällen Konstruktionen από κοινού sein . . . ."

R. Hildebrand notes that the construction occurs in Middle High German, Latin, and Greek.19 He cites a few examples from each of these languages. In the construction he sees a shorthand method of thinking— “eine art Stenographie des gedankens.” The cultural development of a people, he feels, makes necessary a more abbreviated method of expression 19

“Zur Geschichte des Sprachgefühls bei den Deutschen und Römern," Zeitschrift

für deutsche Philologie, II (1869), 260-65. [ 165]

12

THE CONSTRUCTION από XOIVOO IN THE GERMANIC LANGUAGES

which presses to the heart of the matter without employing words which may be dispensed with. In his edition of Erec, M. Haupt assembles a large number of examples of από κοινού from Middle High German.20 Upon the nature of the con­ struction he remarks only that it is found in early and popular Middle High German poetry rather than in the cultivated style of the thirteenth century. The Heliand, he says, is full of it. His examples are grouped according to the grammatical form of the koinon: nominative; accusative; two cases in one form ; verb; phrase; and adverb. Η. H. Bender in a discussion of sixteen examples of από κοινού from Gudrun considers that the construction may sometimes result from the omission of a relative pronoun, that in long and involved sentences it may be occasioned by anacoluthon, and that originally perhaps in most cases it arose from subordination without relative pronouns, a period imme­ diately following parataxis. The possible course of development as he outlines it would be as follow s: 1. Parataxis: do spranc von dem gesidele her Hagene. her Hagene also sprach. 2. Subordination, originally not syntactical but only subordination in thought and expression: do spranc von dem gesidele her Hagene; (her Hagene also sprach) or (der cdso sprach). 3. Coalescence of two words which stood side by side with the same mean­ ing, resulting in από κοινού: do spranc von dem gesidele her Hagene also sprach. He notes the lack of unanimity of opinion concerning the Gudrun examples and the tendency on the part of editors to avoid the construction through varied punctuation and textual readings.21 Hermann Paul states that in the από κοινού construction one sentence may be logically subordinate to the other, so that one might represent it by a relative clause.22 Such a sentence as Gegen Frankfurt liegt ein Ding über heisst Sachsenhausen he considers από κοινού. The koinon too, he thinks, may represent divergent cases as in von einem slangen was gebunden. At a meeting of German philologists in Jena in 1921, B. Delbrück spoke on “Die Konstruktion από κοινού im Hochdeutschen.”23 He dis20 Erec, ed. M. Haupt (2d ed., Leipzig, 1871), pp. 391 ff. 21 “από κοινού in Gudrun ” Journal of English and Germanic Philology, X I (1912), . 565 ff. 22 Deutsche Grammatik (Halle, 1920), IV, 189 f. 23 Verhandlungen der 53. Versammlung Deutscher Philologen und Schulmänner in Jena (Leipzig, 1922), pp. 63-64.

[ 166]

INTRODUCTION

13

agreed with Paul concerning such constructions as von einem slangen was gebunden and sentences containing the verb heissen. Such constructions he considered really two sentences, the second of which could be joined with an anaphorical pronoun, από κοινού originates, in his estimation, from habits of daily speech; as, for instance, the expression, es geht mir etwas besser geht es mir. So arose the construction of a common element with two predicates, and on this model other similar sentences were formed. In an instructive section on από κοινού, O. Behaghel doubts that it originates in anacoluthon.2* Rather it arises, he states, from asyndeton. From a Middle High German point of view such an Old High German construction as the following from Matthew, argengun duo uz pharisara worahtun garati would, he believes, be considered από κοινού; and from such instances with the koinon as subject the construction spread. In Volume IV of his Syntax25 he cites examples wherein a main clause stands από κοινοί) between two dependent clauses. Such an occurrence, he feels, is to be connected with a widespread tendency to express a thought more than once, and may also be influenced at times by som'e uncertainty on the part of the speaker in regard to what he had already said. These constructions he would distinguish from those in which only a word serves as common element. Numerous similar examples from the Heliand had been previously cited by him ;242526 for example, Heliand 1065: ef thu sis godes sunu, be hwi ni hetis thu, ef thu giwald hades. L. Kellner considers από κοινού to be of the same nature with what he calls “the oldest stage of the adjective clause (omission of the relative pronoun)” ; as, for example: H er on þis geare gefor Æ lfred wees œt Baðum gerefa, Chronicle 906 (Parker M S.).27 Such locutions as wees eet Baðum gerefa were in his opinion felt from the beginning as subordinate to the principal sentence, and he would look upon the whole construction as a sentence with one subject and two predicates. The starting point for the construction από κοινού he believes lies in sentences of the type, “There be some sports are painful,” where the common subject occurs in the middle. E. Einenkel considers the simplest form of από κοινού to be the þœt for peet peet construction; for example: he wile nu geleestan peet he lange gehet, where the relative and demonstrative fall together and only one is 24 Deutsche Syntax (Heidelberg, 1928), III, 535 f. On p. 534 he gives the defini­ tion of από κοινού already quoted, which he illustrates with numerous examples from Middle High German, arranged according to the grammatical form of the common element. 25 Deutsche Syntax (Heidelberg, 1932), IV, 289 f. 26 Die Modi im Heliand (Paderborn, 1876), p. 15. 27 Historical Outlines of English Syntax (London, 1892), pp. 61-64.

[167]

14

THE CONSTRUCTION από XOIVOU IN THE GERMANIC LANGUAGES

expressed.28 As the pronoun could assume two functions, one for the main clause and one for the subordinate clause, so could a noun. Such a sentence as se feeder hire sealde ane peowene Bala hatte contains in his in­ terpretation a common element functioning as object of a main clause and as subject of a subordinate clause. In discussing what he terms contact clauses, such as “This is the boy we spoke of,” Otto Jespersen notes that they are frequently explained on the από κοινού principle.2930 He remarks: This may account for some cases, and at any rate may describe the psychological feeling even in Present English sentences like those given above. But the explanation is not sufficient, and if it is admitted, we must at any rate say that the construction was soon extended to cases in which there is no element that can strictly be said to belong to both parts of the sentence.

He feels it safer to say that in early speech the pronoun was felt to be in­ herent in the verb, that two sentences, originally independent units, if pro­ nounced rapidly after each other, came to be felt as a grammatical unit, and that this usage continued after it had become customary to have a pro­ noun as subject. In his so-called contact clauses he cites examples where the από κοινού feeling is excluded but gives others, for example, “My father had a daughter loved a man,” where he feels that the από κοινού analysis may be applied. Concerning Old English, Klaeber in Textual Interpretation of “Beo­ w u lf JS0 expresses the belief that some passages suffer from overpunctua­ tion. He states : “Certainly a simple ( ‘unvaried’) word or phrase occupying a medial position between two terms of variation should be assigned the από κοινού function, whenever it occurs at the beginning of the line.” Ex­ emplifying this he cites Beowulf 753: he on mode wearð | forht on ferhðe. With a proleptic use of nouns or pronouns the verb may, he states, be said to be employed από κοινού. For example, Beowulf 1180: ic minne can I glœdne H roþulf þœt he þa geogoðe wile \ arum healdan. Fur­ thermore he mentions an instance of the από κοινού construction with an object at the beginning of the second half-line, Beowulf 131: polode ðryðswyð pegnsorge dreah. In discussing variation he mentions the form “noun: infinitive phrase, the former being general in its meaning, the latter specific; they are ‘governed’ by the same verb, which may be said to be used από κοινού.” Exemplifying this, he cites Beowulf 1431: bear him ongeaton \ guðhorn galan. 28 Geschichte der Englischen Sprache, II, Historische Syntax (Strassburg, 1916), pp. 121 ff. 29 A Modern English Grammar (Heidelberg, 1927), III, 132 ff. 30 Modern Philology, III (1905), 237 ff.

[ 168]

INTRODUCTION

15

In the chapter headed “‘Ellips,” F. A. Stoett refers, with many ex­ amples, to από κοινού in Middle Dutch.31 He states: Evenals in het Mhd. is ook in het Mnl. eene constructie από κοινοί! bekend, waar­ bij één element als bestanddeel van twee zinnen dienst doet. Haplologie en slordigheid zullen hiervan wel de voornaamste oorzaken zijn.

On the από κοινού principle he explains such sentences as Here laet ons ter vlucht geven (=: laet ons ons) and Ferguut heft sinen wech genomen in een foreest (in een foreest) es hi comen. A. C. Bouman has a broader interpretation of από κοινού in Middle Dutch.32 He remarks: “Te eerder is een constructie από κοινού mogelijk omdat zeer dikwijls ook heterogene dingen met elkaar worden vergeleken.” As an example of this he cites: Flet was een scip mids in de zee, daer in pelegrimme mee dan ander have in was geladen, “ (Dus pelegrimme en ander have).” It is apparent that the subject of από κοινοί has been touched upon in re­ gard to several of the Germanic languages and treated thoroughly only in regard to Middle High German. Attempts at historical tracing of the con­ struction have been limited almost entirely to the range of Middle and Old High German, or else to that of Middle and Old English. There is also great diversity of opinion, both as to the nature of the construction and as to its origin. Regarding the nature of the construction there has been no attempt at discrimination so far as Old and Middle English is con­ cerned. Karg has done an excellent piece of analysis regarding the con­ struction in Middle High German, but his conclusions about the construc­ tion outside of that field are intentionally limited. The occurrence of από κοινού in Old High German and Old Saxon has been referred to only in a restricted manner; to the construction in Old Norse and Gothic there are a few references concerning individual passages. Concerning the origin of από κοινού, all conclusions have been based upon a relatively limited field. And the theories proposed have been many—everyday carelessness in speech, anacoluthon, asyndeton, subordination without expressed pro­ noun, and omission of the relative pronoun. One would like to know if από κοινού is confined largely to certain mem­ bers of the Germanic branch, appearing differently in each of these, or if it is a common Germanic syntactical structure, possessing throughout some common characteristics. It is this question which has led to the present discussion of the subject. Such conclusions as are drawn are based upon a reading of the following 31 Middelnederlandsche Spraakkunst, Syntaxis (VGravenhage, 1923), pp. 152-53. 32 Bijdrage tat de Syntaxis der “Dat”-Zinnen in het Germaansch (Utrecht, 1918), pp. 92-94.

[169]

16

THE CONSTRUCTION obtO κ ο ινο ύ IN THE GERMANIC LANGUAGES

literature: (1) Gothic: The Gothic Bible and Skeireins; (2) Old High German: Tatian, O tfrid; (3) Old Saxon: the Heliand and the Old Saxon Genesis; (4) Old N orse: the Poetic Edda, Gylfaginning of the Prose Edda, the first six sagas of the Sturlunga group; (5) Old English: the poetry collected in the three volumes of Grein-Wiilcker, Bibliothek der angel­ sächsischen Poesie, Alfred’s translation of Orosius, Alfred’s translation of Bede, the Anglo-Saxon Chronicle; (6) Middle English: some fifty thou­ sand lines of Middle English poetry; (7) Middle Dutch: some sixty thou­ sand lines of Middle Dutch poetry; (8) Middle High German: some Middle High German poetry. For such reference as is made to this last field I have used examples mainly from the collection made by Karg. AN INTERPRETATION OF THE TERM cbtO XOIVOU

Since the term από κοινού will be used henceforth in this discussion only as the writer interprets it, a definition is necessary as a working basis. The definition is not made arbitrarily but is deduced from the observation of numerous similar syntactical constructions. Such constructions do not, of course, always fall into stereotyped forms; but there are certain main characteristics which can be held to. From the observation of these the following definition is formulated. από κοινού is a syntactical construction in which a word or closely re­ lated group of words, occurring between two portions of discourse, con­ tains an idea which completes the thought of the first part, to which it is grammatically related, at once supplies the thought essential to the follow­ ing part, to which it may also be grammatically related, and is not felt to belong more closely with the first part than with the second. It is not always possible to draw hard and fast rules about a phase of syntax, and in some cases the including of a construction in the category of από κοινού or its exclusion therefrom is a matter of degree and not of clear-cut distinction. In the main, however, the definition here given is tenable. For the sake of specific example, the following citations are given: Old English, Beowulf 740: ac he gefeng hraðe forman siðe slat unwearnum

slœpendne rinc

Here slcependne rinc stands between gefeng and slat, completes the thought expressed in gefeng and is object thereof, supplies the thought essential to slat and is object thereof, and stands in a half-line, hence is not felt to be more closely connected with what precedes than with what follows. Simi­ larly in the following examples : Beowulf 131: þolode ðryðswyð

þegnsorge dreah

[ 170]

INTRODUCTION

17

Here þegnsorge is object both of þolode and dreah. Andreas 537 : þa hleoðrade

halgan stefne cyning wyrSude

cempa collenferhð

Old Saxon, Heliand 2238: Segel up dadun letun wind aftar manon ofrar thana meristrom

wedarwisa man

Old Norse, Sigorþarkviþa 4: ne hunskr konungr hefjask at arme fal mege giuka

mey frumunga

Brot af Sigorþarkviþo 19: þa reynde þat moþogr a vit

es riþet hafþe

min at biþja

Old High German, O ff rid I, 25, 9: zi imo sprah tho lindo kundta imo . . . .

ther gotes sun selbo

O tfrid II, 1, 21: Tho er deta thaz sih zarpta

ther himil sus io warpta

Middle English, P H .A . 3, 151 Heo buggeth with heore Iuweles

vr Iustises heo schendeth

Middle Dutch, Walewein 5510: Ende soe scoot in haer ansichte

Met haren naglen trac soe die huut

Obviously the criteria at hand in poetry for determining whether a word is more closely related to what precedes or what follows cannot be applied in prose. If on any ground it can be shown that there is a feeling of a common element, such a construction as the following is considered από κοινού: Tatian 269, 5: Tho antlingita ther ander increbota inan.

Constructions are not considered από κοινού if the word which would have to serve in common must serve two cases, one of which is different in form from the word as it appears; nor if it is distinctly more closely re­ lated to the preceding part of the sentence; nor if the second part is so [ 171]

18

THE CONSTRUCTION α π ό κ ο ίνο ΰ IN THE GERMANIC LANGUAGES

subordinate in idea that a carrying over of thought is not essential.. These three points are illustrated, respectively, by the three following examples: Roman van Lancelot 13872: Dat mi mijn vader te wive gaf Enen kinde was des conincs sone.

Heliand 5100: Tho balg ina the biskop

habda bitran hugi.

Alexander 7, 851: Daer es een boom heet lignum vite.

Since the discussion of από κοινού in the field of Germanic is a broad subject, these examples have been taken for the sake of orientation before specific details are taken up. THE PLAN OF DISCUSSION

Regarding the από κοινού construction, there are four main factors to be taken into consideration: (1) the grammatical nature of the koinon, (2) the context of the sentence in which the construction occurs, (3) the position of the koinon as it occurs in verse, and (4) the possible connection of από κοινού with constructions of a somewhat similar type.33 The ques­ tion of the position in verse divides the subject into two parts—poetry and prose; and these will be considered separately, with a discussion of the grammatical nature of the koinon and the general context of the sentence in each. The question of the relation to other constructions involves a considerations of two constructions: (a) asyndetic parataxis; ( b) attributive constructions in which the second part expresses an attribute or state of the word which would be taken in common. This construction has been referred to variously as the adjective clause, omission of the relative, con­ tact clause, and /^^-Konstruktion. In the matter of asyndeton, the ques­ tion is whether a word is to be taken in common or whether it is more closely related to a preceding part. As the means for determining this are not the same in poetry and prose, the question will be taken up first along with the από κοινού construction in poetry and secondly with this construc­ tion in prose. The attributive construction, however, is not confined to a question of sentence-pause, and its appearance in prose and poetry may be discussed together. To start with the earliest appearance of από κοινοί, the next 33 This method of analysis is similar to that of Karg in his “Die Konstruktion από κοινού im Mittelhochdeutschen/’ in Syntaktische Studien; other writers men­ tioned in the résumé of discussions provided above have individually considered one or several of these points.

[172]

INTRODUCTION

19

chapter considers the construction in Old Germanic poetry. As it continues into the poetry of the Middle Germanic period, the following chapter takes up the subject in Middle English and Middle Dutch, with comparisons also concerning Old Germanic and Middle High German. The fourth chapter is given to a discussion of από κοινοί in Old Germanic prose. The nature of the construction having been examined in these three fields, in the next chapter a comparison is made between από κοινού and the attributive con­ struction. At the end of each of these chapters just mentioned, a table of citations is given. In conclusion it will be possible to add some new evidence regarding the origin of the από κοινοί construction in Germanic.

[ 173 ]

CHAPTER II από κοινού IN OLD GERMANIC PO ETRY In Old English poetry, the Poetic Edda, the work of Otfrid, and the Heliand and Old Saxon Genesis, all included in this discussion, a line con­ sists of two half-lines between which there is a caesura. In an από κοινού construction the common element may take any of the following positions: (1) It may fill a half-line, as, for instance: Heliand 2469: an is briost hledit linod endi lestid

that gibod godes

(2) It may rarely occupy a half-line along with the first part of the con­ struction. Here it is separated by the caesura from the second part and this intervening pause vitiates somewhat the feeling that there is a common element, as it is metrically more closely attached to the first part of the construction; for instance: Beowulf 513: þaer git eagorstream earmum þehton mæton merestrœta mundum brugdon

Here one must decide between από κοινού and asyndeton. (3) It may oc­ cupy a half-line along with the second part of the construction; for instance: Otfrid II, 9, 29: nu will ih hiar gizellen

ein bilidi ginennen

Rhym>e Poem 9: þa wæs wæstmum aweaht

world onspreht

In no case in the material covered does the whole construction fall within a single half-line. A. THE COMMON ELEMENT CONSTITUTES A H ALF-LINE (RARELY TWO h a l f-l in e s )

In this construction the koinon may be either the first or the second half-line. If it falls in the first half-line, it is set off somewhat from the portion in the preceding line by the natural feeling of a slight pause follow­ ing the completion of a full line. It is set off from the succeeding portion by the caesura. For example: Beowulf 56: heah Healfdene

20]

oþ þæt him eft onwoc heold þenden lifde

[174]

IN OLD GERMANIC POETRY

21

If the koinon falls in the second half-line, it is set off from the preceding portion by the caesura, and from the succeeding portion by the feeling of pause at the end of the line. For example: Maldon 29: me sendon to þe heton 6e secgan

sœmen snelle

A reading of the major portion of Old Germanic verse has yielded 128 examples of this type, wherein a half-line (and rarely a full line) is filled by a sentence element that is grammatically and contextually impor­ tant both to what precedes and to what follows. The following examples illustrate this type of construction in the Poetic Edda, the Heliand, Otfrid, and Old English poetry: 1. Nominative case.—Koinon is nominative case (87 in all), for ex­ ample : Beowulf 1013: Bugon þa to bence fylle gefægon1

blœdagande

Beowulf 1119: wand to wolcnum hlynode for hlawe

wœlfyra meest

Phoenix 186: we der condel we arm

suþan bliceS weorodum lyhteð

O tfrid I, 25, 9: zi imo sprah tho lindo kundta im er iz wolta

ther gotes sun selbo iz ouh so wesan scolta

O tfrid IV, 36, 1: ni mohtun noh bilinnen thes armilichen willen thie selbun ewarton thaz ougtun tho mit worton2

Heliand 2238: segel up dadun letun wind aftar manon obar thana meristrom

wedarwisa man

1 The subject of the following half-line, fœgere geþœgon, is not blœdagande, but magas þara. Cf. Hoops, Kommentar zum Beowulf (Heidelberg, 1932), p. 127. 2 In two examples from Otfrid the first verb is singular, the second plural: Otfrid IV, 24, 13: Ingegin riaf tho luto heriscaf thero liuto irscrirun filu gahun Here the singular heriscaf governs the singular riaf, but the plural idea of the subject as a whole induces a plural form for the following verb. Similarly V, 20, 19.

[ 175]

22

the construction

από κοινού

in the

Germanic

languages

Heliand 4050: than werdat fan dode quike thurh maht godes arisat fan restu

mankunneas gihwilik

V filundarkviþa 12: sat a berf jalle bauga talþe eins saknaþe

alfa Ijoþe

Beowulf 1814: eode weorð Denum æþeling to yppan þær se oþer wæs hœle hildedeor Hroðgar grette

In this example it is possible that the group hcele hildedeor is subject of grette and is felt to be closely attached appositionally to œþeling, subject of eode? In the following examples the appositional phrases hang closely together and form a compact koinon: Beowulf 2204: ða hyne gesohtan on sigeþeode hearde hildefrecan Heaðo-Scilfingas niða genægdan nefan Hererices

Finn. 16: ða to dura eodon

Sigeferð and Eaha

drihtlice cempan hyra sword getugon

In the preceding examples the koinon stands as subject. There is here a primary distinction to be made. When the koinon is subject, if it is fol­ lowed by one verb after which a new subject is expressed or a distinctly new idea is taken up, the construction is felt to be more completely a unit than in instances where the second element is followed by another verb with unexpressed subject, as in a group. For example: Heliand 1075: tho bigan eft niuson endi nahor geng odru sidu fandoda is frahon that fridubarn tholoda3

unhiuri fiund

3 Hoops, Kommentar, p. 197: “hcele hildedeor ist από κοινού gebraucht: als Varia­ tion zu œþeling, aber zugleich als Subj. zu g r e t t e Klaeber, Anglia L, 209 : “Whether the phrase is meant as variation of œþeling, pure and simple, looking backward (Kock, Anglia X LV I, 90 f.) or occupies an από κοινοί) position, looking forward also (to Hroðgar grette ), I am not prepared to decide.” Kock, Anglia X LV I, 91, takes weorð Denum, œþeling, and hœle hildedeor as three parallel members subject of eode and in his rendering of the passage joins the last verb by a conjunction. H e feels that hœle hildedeor might go with grette, but does not state that it might then also go with a preceding part.

[ 176]

IN OLD GERMANIC POETRY

23

Here, after the second part of the construction, fandoda is frahon, there is a definite change in idea and a new subject is expressed. In such construc­ tions, the first two verbs with intervening subject constitute a distinct unit of thought. This one may call the closed construction. In such examples as the following, however, the two sentence portions surrounding the com­ mon element are felt to form less of a unit. One may call this the open construction. Andreas 962: þæt me bysmredon

we ras wansœlige

bendum fæstne wordum tyrgdon

slogon 7 swungon

Heliand 832: giwitun im tho eft thanan fan Jerusalem Joseph endi Maria habdun im ti gisidea suno drohtines allaro barna betsta thero the gio giboran wurdi magu fan moder habdun im thar minnea to

Rigsþula 43: upp oxu þar hesta tgmþo

Jarle bomer hlifar bendo

Beowulf 1980: meoduscencum hwearf geond þæt healreced Hœreðes doktor lufode ða leode liðwæge bær

In these examples there is an element which is to be construed with what precedes and with what follows, and it is no moré closely related to one than it is to the other. But the asyndetic addition of one or more other verbs gives a certain looseness to the construction. This asyndetic joining of several verbs is common in old Germanic poetry, and it would not be surprising if it occurred after an από κοινού construction as well as else­ where. It is, however, relatively rare. From 87 examples in which the koinon is the subject, there are only 24 examples where another verb fol­ lows with unexpressed subject. But of these 24 examples, 12 are of a type wherein the first two verbs, with their common subject, form a unit in thought to which the third verb is alien: a) The third verb is joined by a conjunction. In such cases the feeling of unity in the first two verbs is preserved, and the third verb often has the aspect of an addendum. For example: Andreas 537: þa hleoðrade

halgan stefne cyning wyrðude wuldres waldend 7 þus wordum cwæð

cempa collenferhð

[ 177]

24

t h e c o n s t r u c t io n

από κοινού i n

the

Ge r m a n ic l a n g u a g e s

Here the two verbs, hleoðrade and wyrðude, with their common subject, cempa, form a unit of thought from which the third verb, cwœð, is ex­ cluded.4 b) The first verbs stand closely together, but a following verb is sep­ arated from them by numerous intervening words. For example: Genesis 1376: sträng wæs 7 reðe wreah 7 þeahte manfæhðu beam middangeardes wonnan wæge wera eðelland hof hergode5

se ðe wœtrum weold

c) The parts joined by the common element express what can be con­ sidered two parts of a general idea, while the third verb brings in a thought which, while not distinctly a new idea, is nevertheless felt to be detached from the thought of the first two verbs. For example: Heliand 3053 : tho te lat ni warth sprak san angegin eno for im allon habda im ellian god

Simon Petrus

Here the two parts, joined by the common element, Simon Petrus, may be looked upon as a method of expressing the general idea, 'he wasn’t slow in speaking.’ There is a distinct break in thought between this idea and the following expression, 'had good courage.’ Beowulf 234: gewat him þa to waroðe wiege ridan þrymmum cwehte mægenwudu mundum meþelwordum frægn

þegn H roðgares

Here the portions joined by the common element þegn Hroðgares express the general idea of the thane’s approaching the shore in a rather hostile fashion. The third verb introduces the speech which follows. Beowulf 2538: aras ða bi ronde rof oretta heard under helme hiorosercean bær under stancleofu strengo getruwode anes mannes

The first two sentence elements joined by the common subject, rof oretta, 4 Similarly, Chronicle Poem SB1, Whale 27. 5 Similarly, Juliana 614, Heliand 832. In the example above the synonyms wreah and þeahte go together, forming the second verbal expression in the group and fol­ lowed at some length by the third, hergode.

[ 178]

IN OLD GERMANIC POETRY

25

express the general idea of the warrior’s getting ready to fight. The third verb brings in a new idea, namely, that he trusted in his own strength. Daniel 256: bliðe wæron ofestum heredon drihten on dreame dydon swa hie cuðon6

eorlas Ebrea

The expressions bliðe wæron and heredon, with their common subject, eorlas Ebrea, contain the idea that joyfully the leaders of the Hebrews worship the Lord. The clause, dydon swa hie cuðon, is a distinctly supple­ mentary expression.. Out of 87 cases where the common element forms the subject, there are 75 examples in which the two verbs with their one subject form a syntactical construction which is not only neatly connected grammatically but which also, within the continuous expression of ideas in the poem, forms a dis­ tinct unit of thought. There are 12 examples in which the two verbs with their common subject do not form such a unit of thought. For example: Beowulf 1980: meoduscencum hwearf geond þæt healreced Hœreðes dohtor lufode þa leode liðwæge bær

Here the general idea is that of ministering to the men in the hall, and the third verb is as essential to this thought as are the other two. M etra 28, 15: on t5ære ilcan

eall ruma rodor

eaxe hwerfetS recene scriðeð

suðheald swifeð

The movement of the firmament is the idea contained in this whole passage. The first two verbs with their common subject do not express an idea essen­ tially different from that expressed by the third verb. The 12 examples of this type, however, are distinctly in the minority compared with the 75 cases in which the first two verbs with their common subject constitute a unit of thought. The only essential difference between the open construction and the closed construction is that the former, in some instances, does not possess the compactness of expression which is charac­ teristic of the latter. 2. Accusative case.— Koinon is accusative (15 in all), for example: Christ 484: 7 fulwiað folc under roderum hweorfað to heofonum 6 Similarly, Elene 61, and Beowulf 2430.

[ 179]

26

THE CONSTRUCTION

άπό κοινού IN

THE GERMANIC LANGUAGES

Here Christ is giving commands to his disciples as to what they are to perform. That folc under roderum must be the object of hweorfað is clear from the context. The passage runs on, 'break up idols, abolish hostility, sow friendship/ The idea is plainly to turn mankind toward heaven. Under hweorfan Bosworth-Toller Dictionary cites this passage a s : fulwiaþ folc hweorfaþ to heofonum: 'baptize people and turn them to heaven/ M etra 27, 2: swa swa mereflodes yða hrerað iscalde see wecgað for winde

The waves stir and move the ice-cold sea because of the wind. Both hrerað and wecgað need a direct object.7 Salomon and Saturn 395: cristnað 7 clænsað wuldre gewlitigað

cwicra manigo

Christ and Satan 154: þær we ymb hine utan ealle hofan leomu ymb leofne lofsonga word drihtne saedon

Lofsonga word is object both of hofan and sœdon. Leomu is used figura­ tively here, 'limbs of the lord/ parallel with we in the preceding line.8 Heliand 2630: ubile endi gode

endi fahid bediu tiuhid up te stade

lidod sie te lande

It is significant that the pronoun is expressed after lidod but not after tiuhid. The expression ubile endi gode is object of fahid and stands in a position to serve as object of tiuhid, so that no further expression is necessary. Sigorþarkviþa 4: ne hann kono kyssa gjzfrþe ne hunskr konungr hefjask at arme mey frumunga fal mege Gjuka 7 Cf. Daniel 388: þa ðe lagostreamas | wœterscipe wecgað, and the prose Boethius 39, 1: swa swa yða for winde þa sœ hrerað. 8 Cf. Bosworth-Toller Supplement under lim. E. Kock, “Jubilee Jaunts and Jot­ tings/’ Lunds Universitets Arsskrift, N.F. avd. 1, Bd. 14, nr. 26, p. 70, takes leomu as object of hofan: “lifted up our wings.” In the context the raising of wings seems less likely than the raising of songs of praise, and leomu, ‘w ings/ is less likely than leomu, ‘limbs (of the lord )/ Lofsonga word is taken as object of both verbs by G. Krapp, The Junius Manuscript (N ew York, 1931), p. 235.

[ 180]

IN OLD GERMANIC POETRY

27

The Hun king is Sigurth, and the passage refers to his lying with Brunhild chastely, since he was winning her for Gunnar, the son of Gjuki. In English it might be possible to translate this construction, ‘he never had to his arms the maiden he kept for Gunnar/ and explain it as a relative ellipsis, the relative ‘whom' having merely been left out. But it is doubt­ ful if this applies here. There is really no relative subordination in the second part of the construction. There are rather here two independent ideas, ‘he never had the maiden in his arms' and 'the maiden he kept for Gunnar/ M ey frumunga is definitely the object of hefja and was probably felt to be directly the object of fal.9 Guþrunarkviþa II, 8 : þa heyrer þu hrafna gjalla prno gjalla cede fegna varga þjota

Here the adjective phrase, œzle fegna, modifies both hrafna, grno in the preceding section and varga in the following section.10 Beowulf 1807 : Heht þa se hearda Hrunting beran heht his sweord niman

sunu Ecglafes

Interpreting this ‘then the brave one bade the son of Ecglaf bear Hrunt­ ing, bade him take his sword/ sunu Ecglafes stands από κοινού subject accusative of beran and niman.11 Beowulf 1020 : Forgeaf þa Beowulfe beam Healfdenes segen gyldenne sigores to leane hroden hiltecumbor helm ond byrnan mœre maþþumsweord manige gesawon beforan beorn beran.

It is usual to punctuate here with a semicolon after byrnan or a colon after -sweord.12 But -sweord surely belongs in the group segen, helm, byrnan, all objects of forgeaf. The process of enumeration is rather lengthy, and 9 Cf. B. Sijmons and H. Gering, Die Lieder der Edda, Kommentar (Halle, 1931), II, 246: