Root-Determinatives in Semitic Speech: A Contribution to Semitic Philology 9781463212933

Citation preview

Root-Determinatives in Semitic Speech

Analecta Gorgiana

33 Series Editor George Kiraz

Analecta Gorgiana is a collection of long essays and short monographs which are consistently cited by modern scholars but previously difficult to find because of their original appearance in obscure publications. Carefully selected by a team of scholars based on their relevance to modern scholarship, these essays can now be fully utilized by scholars and proudly owned by libraries.

Root-Determinatives in Semitic Speech

A Contribution to Semitic Philology

Solomon Theodore Halevy Hurwitz

-äk

1

gorgias press 2009

Gorgias Press LLC, 180 Centennial Ave., Piscataway, NJ, 08854, USA www.gorgiaspress.com Copyright © 2009 by Gorgias Press LLC

All rights reserved under International and Pan-American Copyright Conventions. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning or otherwise without the prior written permission of Gorgias Press LLC. 2009

1

ISBN 978-1-59333-627-1

Printed in the United States of America

ISSN 1935-6854

NOTE During the last fifty years the horizon of our view upon the extent and development of the Semitic languages has been singularly enlarged; on the one hand by archaeological investigations that have brought to light unknown ¡Semitic civilizations, and on the other by a study of the most modem developments of these languages. The origin and early history of these languages offer a problem that is at once interesting and perplexing. I t has been in the mind of scholars since the first attempts were made in Europe to grasp the extent and the importance of ancient and mediaeval Semitic culture. Even in the Early Middle Age», essays were put forward to solve the question, at least in regard to one of these tongues, the Hebrew. I n modern times such undertakings have been quite numerous. The difficulties in the way of proving any theory that will account for this development are, of course, great: the pitfalls are numerous. Jl»r. Eurwitz has tried to overcome the one and to avoid the other by a strict adherence to a scientific method and by choosing his examples with great care. I t is idle to hope that he will convince all students of so vexed and contentious a question; but he has offered a solution that deserves the careful attention and scrutiny of his fellow-workers in the same field. May, 1913.

.RICHARD GOTTHEIL

TO MY MOTHER

HANNAH

LANGLEBEN-HIÏRWTTZ

PREFACE The "biliteral theory", sometimes attacked, at other times defended, and almost entirely rejected in recent years, forms the subject of these studies. They were begun in 1907 in a somewhat empirical fashion, but the author became deeply interested in the mass of literature which had accumulated on the subject, although these researches had apparently resulted in hopeless confusion rather than in even an approximate solution of their problem. As a result of his investigations, the writer was very glad to find that able grammarians had not labored for centuries in vain, contending blindly over something that had no existence; but that, on the contrary, the biliteral root in Semitic was an actual entity. In 1910, encouraged by the friendly interest shown to him at Columbia University, the author presented a thesis in which he briefly outlined the possibility of the assumption of an original biliteral character of the Semitic root as a key to the solution of many vexing problems in Semitic etymology, although he has been obliged, from the very first, to modify his views, and to adopt an entirely conservative position. His results, which have been attained independently, have shown that he has been anticipated in almost every point by such scholars as Pappenheim and Lagarde; but this very fact only renders him the more sure that his position is justified. In fact, he does not claim originality in any one of his conclusions, for they have, each and all, been the historical and logical result of the labors of the ix

X

PREFACE

scholarly world before him. -He has merely combined, sifted, and profited by the researches—and even by the errors—of former generations, and he believes that he has brought the biliteral theory within a restricted field, and that he has placed it upon a scientific foundation. His method has necessarily been conservative; he has been obliged to consider each of the Semitic languages as an individual unit, the vocabulary of which can best be explained by itself, i. e., by a comparison of kindred words within the same dialect having identical fundamental elements. Only when such means were unavailable was he compelled to tolerate cross-comparisons of words in different dialects, which may have been separated from each other by many centuries of usage i. "Whatever light semasiology has thrown upon the development of words and their usages he has liberally used. Doubtless a more generous equipment of lexicons (especially for Arabic and Ethiopic) would have aided much, but this defect has long been felt, in the Semitic world of scholarship, and can be remedied only by degrees. On the whole, the fundamental idea conveyed by each root in the various languages is quite evident from a study of all its derived forms. The subject naturally falls into two divisions: the grammatical and the lexicographical. The first is considered in the opening chapter; the last, which is by far the most important, in the following chapters. A survey of the various theories of former Jewish biliteralists would have been interesting, but would have added nothing of perman1 The rule has been generally applied that it is safest to compare one North Semitic form with its phonetic equivalent in another North Semitic dialect, while South Semitic has been generally compared with South Semitic (Of. Noldeke, Die semitisehen Sprachen, p. 14, and

W r i g h t , Comp. Gram.,

p. 10).

PREFACE

XI

ent value to the subject; since all such theories were based on an imperfect and unscientiiic knowledge of the laws which govern Hebrew grammar 2 . It was therefore thought advisable not to enter into the matter at the present moment. The author's thanks are due for the constant aid and advice tendered by his teachers at Columbia University, Professor Richard J . H. G-ottheil, Professor John D. Prince, and Professor A. V. Williams Jackson. To Dr. Louis H. Gray he is grateful for many valuable suggestions. S. T. H. H. 2

For literature on the subject consult the appended bibliography.

CONTENTS CHAPTER

PAGE BIBLIOGRAPHY

I.

THE I

XV

SUBCONSCIOUS B I F I T E R A L R O O T OR T H E W E A K

II. DENOMINATIVE

II.

STEM

7

TRILITERAL STEMS

III. ROOT-DIFFERENTIATION

30 34

A

S T U D Y OF P L U R I L I T E R A L S

AND THE R E S U L T I N G T H E O R Y

.

.

37 40

II. DEVELOPMENT BY PREFORMATIVES

55

I I I . I N F I X A T I O N IN S E M I T I C S P E E C H

60

IV. DEVELOPMENT

64

V. T H E

BY SUFFORMATIVES

TRILITERAL

EXAMPLES I.

A DEVELOPED STEM

OF P R I M I T I V E

STATIVE

SEMITIC STEM-FORMATIONS

STEMS

66 74 75

II. INTENSIVE

STEMS

85

III. PURPOSIVE

STEMS

85

I V . C A U S A T I V E AND K I N D R E D V. R E F L E X I V E

AND K I N D R E D

STEMS

89

STEMS

95

V I . S T E M S WITH P O S T - D E T E R M I N A T I V E S IV.

19

IV. B I B L I C A L FOLK-ETYMOLOGIES I . SEMITIC PLURILITERALS

III.

1

T H E T R I L I T E R A L I T Y OF T H E W E A K V E R B — C A Ü S E AND O R I G I N

105

S U M M A R Y OF R E S U L T S

107

I N D E X OF R O O T S

ILL

xiii

BIBLIOGRAPHY A. PERIODICALS American Journal of Semitic Languages and Literatures (formerly Hebraica). Chicago 1884 ff. Beiträge zur Assyriologie und vergleichenden Sprachwissenschaft. Leipzig 1890 ff. Giornale delia Società Asiatica Itaiiana. Firenze 1887 ft'. Journal of the American Oriental Society. Boston and New Haven 1849 ff. Journal asiatique publié par ¡a société asiatique. Paris 1822ÎT. * Journal of the Royal Asiatic Society of Great Britain and Ireland. London 1834 if. Literaturblatt des Orients. Berichte, Studien und Kritiken für jüdische Geschichte und Literatur. Leipzig 1840—1851. Revue des études juives. Paria 1880 ff. Wiener Zeitschrift für die Kunde des Morgenlandes (Vienna Oriental Journal). Wien 1887 ff. Zeitschrift für die alttestanü-ntiirhe Wissenschaft. Gießen 1881 ff. Zeitschrift für Assyriologie und verwandte Gebiete. Leipzig 1886 tf. Zeitschrift der deutschen morgenländischen Gesellschaft. Leipzig 1847 ft'.

ABBKLiVrATION AJSL. BA.

JA OS. JA. JRAS.

IIEJ. WZKM. ZA TW. ZA. ZDMQ.

B. COLLECTED WORKS Jewish Encyclopedia. lü volumes. New York and London 1901—1906. JE. Kleinere Schx-iften von Dr. H. L. Fleischer. 8 Bde. Leipzig 1885—1888. Morgenländische Forschungen. Festschrift Herrn Professor Dr. H. L. Fleischer. L eipzig 1875. Morgenl. Fovschngn. Orientalische Studien, Theodor Nöldeke . . . . gewidmet. 2 Bde. Gießen 1906.

XVI

B I B L I O G R A P H Y

Semitic Studies in memory of Rev. Dr. Alexander K o h u t edited by G. A . Kohut. Berlin 1897. Symmicta von Paul Anton de Lagarde. 2 Bde. Goettingen 1877—1880. C.

G E N E R A L

SEMITIC

ABBKEYIATIO»

Sem.

Studies

P H I L O L O G Y

BARTH, J., Etymologische Studien zum semitischen insbesondere zum hebräischen Lexicon. Leipzig 1893. — Die Nominalbüdung in den semitischen Sprachen. Leipzig 1S94. — Sprachwissenschaftliche Untersuchungen zum Semitischen. 2 Teile. Leipzig 1907—1911. BROCKELMANN, CARL, Grundriß der vergleichenden Grammatik der semitischen Sprachen. 2 Bde. Berlin

1903-

yß%

Grundriß

Abhandlungen f ü r semitische W o r t forschung. Leipzig 1844. H U I Z I N G A , A. H . , Analogy in the Semitic Languages. Baltimore 1891. K Ö N I G , E D U A R D , Hebräisch und Semitisch. Prolegomena und Grundlinien einer Geschichte der semitischen Sprachen. Berlin 1901. L A G A R D E , P A U L A N T O N [DE], Übersicht über die im Aramäischen, Arabischen und Hebräischen übliche Bildung der Xomina. (Abhandlungen der königlichen Gesellschaft der Wissenschaften zu Güttingen xxxv). DIETRICH, F R A N Z ,

Göttingen 1889.

Bildung

M., Etudes sur la formation des racines sémitiques. Paris 1858. M A R T I N , S I R V . , Inquiries concerning the Structure of the Somitic Languages. 2 parts. London 1876— 1878.

LEGUEST,

E R N S T , Die Bildung und Bedeutung des Plural in den semitischen und indogermanischen Sprachen. Mannheim 1846. N Ö L D E K E , THEODOR, Die semitischen Sprachen, Leipzig 18b7. — Neue Beiträge zur semitischen Sprachwissenschaft. Straßburg 1910. R E N A N , E R N E S T , Histoire générale et système comparé des langues sémitiques. 4« édition. Paris 1863. R U C Z I C K A , R , , Konsonantische Dissimilation in den MEIER.

Abhandlungen

d. Nomina

BIBLIOGRAPHY

xvii ARBKEVIATION

semitischen Sprachen. (Beiträge zur Assyriologie v i ' H e f t 4). Leipzig 1909. WRIGHT, WILLIAM, Lectures on the Comparative Grammar of the Semitic Languages. Cambridge. 1890. Z I M M E R N , H . , Vergleichende. Grammatik der semitischen Sprachen. Berlin 1898. D.

H E B R E W

a) Jewish Bilitei alists and their DONNOLO,

SHABBETHAI

BEN ABRAHAM

Writings

'JLÜSN

1SD

II commento di S. Donnolo sul libro della creazione . . . da David Castelli. Contains a cabbalistic exposition of the value of the letters in the roots and is the principal source for all such later investigations. Firenze 1880. FÜRST, JULIUS BHpn JNS^ "ISIM Librorum sacrorum veteres testamenti concordantiae. Follows biliteral principles. Lipsiae 1840. H E R Z F I E L P , L £ V I , Einblick in das Sprachliche der semitischen Urzeit. Han novel' 1883. J U D A H A K T E H BEN Z E B I H I R S C H OF C A R P E N T R A S

MW*

^NS

An etymological dictionary on biliteral principles. Jeßnitz 1719. LEVINSOHN,

ISAAC

BAER

DK/ ^."IK

Etymological studies on biliteral principles. Warsaw 1893. — «Bn» Etymological studies on biliteral principles. Wilna 1841. nw ritten — Etymological studies on biliteral principles. Warsaw 1877. LEVISOHN, GEOKOE rbio nroin A commentary on Ecclesiastes on biliteral principles. Hamburg 1784. MARCUS,

AHRON

^PO

Sprache als Schritt der Psyche. Berlin 1905. Follows the older cabbalistic exposition of the value of letters in roots. M E N AH EM B E N

SHLOMO

31B

Composed 1130. Pentateuch commentary on biliteral principles (see introduction to Levisohn's iliuß nnsin).

Ccm_p. Gram.

xviii

BIBLIOGRAPHY

rnano » i n nuwm onio A Hebrew lexicon edited by H. Filipowski. London 1854.

MENAHEM

BEN SARUK AND DUNASH BEN

ABBREVIATION

LABEAT

PAPPENHEIM, SOLOMON

NOB®

PM

First part of a biliteral dictionary containing the letters S and S. Breslau 1802. ntab» w r v — A book of Biblical synonyms in three parts on biliteral principles. Dhyrenfurth 1784—Roedelheim 1812. S T E I N B E R G , JOSHUA

12»

A Hebrew grammar with an exposition biliteral theory. Wilna 1891.

—

A dictionary on biliteral principles. 1903. b) Recent Grammars,

'SUMS

of the

amt»N

BBBD

8*!1 ed. Wilna

Lexicons,

etc.

W., Anfänge der hebräischen Grammatik (Zeitschrift der deutschen morgenländischen Gesellschaft xlix. pp. 1—62; 335—392). B A R T H , J . , Wurzeluntersuchungen zum hebräischen und aramäischen Lexicon. Leipzig 1902. BÖTTCHER, J . , Ausführliches Lehrbuch der hebräischen Sprache. 2 Bde. Leipzig 1866. CRAWFORD , F . J . , Horae Hebraicae. An elaboration of the prepositional theory. London 1868. D E L I T Z S C H , FRANZ, Iesurun, sive isagoge in grammaticam et lexicographiam linguae hebraicae. Grimmae 1838. DELITZSCH, F R I E D R I C H , The Hebrew Language viewed in the Light of Assyrian Research. London 1883. — Prolegomena eines neuen hebräisch-aramäischen Wörterbuchs zum alten Testament. Leipzig 1886. D ' O L I V E T , F A B R E , La langue habraïque restituée. Exposition of Hebrew roots according to cabbalistic value of letters. Paris 1815. EWALD, HEINRICH, Ausführliches Lehrbuch der hebräischen Sprache des alten Bundes. 8*« Auflage. Göttingen 1870. F Ü R S T , J U L I U S , Hebräisches und chaldäisches Hand-

BACHER,

Lehrbuch

Prolegomena

Lehrbuch

xix

BIBLIOGRAPHY

ABBREVIATION

Wörterbuch über das alte Testament. 2 t e Auflage. Leipzig 1863. — A Hebrew and Chaldee Lexicon to the Old Testament. Translated by Samuel Davidson D.D. New York 1867. G E I G E R , A., Lehrbuch zur Sprache der Mischnah. Breslau 1845. GELLOT, E Ü O E N E , The Monosyllabism of the Semitic Languages. New York 1912. G E R B E R , "W. J., Die hebräischen Verba Denominativa. Leipzig 1896. GESENIDS, W I L H E L M , Geschichte der hebräischen Sprache und Schrift. Leipzig 1815. — Hebräische Grammatik. 11 to Auflage. Halle 1831. G E S E N I U S - K A Ü T Z S C H , Hebräische Grammatik. 2 7 * 0 Auflage. Leipzig 1902. GESENIUS, W I L H E L M , Thesaurus philologicus criticus linguae hebraeae et chaldaeae veteris testamenti. Bditio altera. Lipsiae 1835. G E S E N I U S - B R O W N - D R I V E R - B R I G S S , A Hebrew and English Lexicon of the Old Testament. Boston 1906. G E S E N I O S - Z I M M E R N - B U H L , Hebräisches und aramäisches Handwörterbuch über das alte Testament. 14 t e Auflage. Leipzig 1905. K O H U T A L E X A N D E R , Aruch Completum. 8 parts. Vienna 1878—New York 1892. K Ö N I G , E D U A R D , Historisch-kritisches Lehrgebäude der hebräischen Sprache. Leipzig Bd. i. 1881. Bd. ii. 1. 1895. Bd. ii. 2. 1897. L E V Y , J A C O B , Neuhebräisches und chaldäisches Wörterbuch. Leipzig 1876—1889. M E T M A N N , L E O , Die hebräische Sprache. Jerusalem 1904. N E U M A N N , C A S P A R , Genesis linguae sanctae veteris testamenti. Eyposition of the cabalistic value of letters in roots (cf. Gesenius, Geschichte d. heb. Sprache, p. 126). Norimburg 1696. OLSHAUSEN, J U S T U S , Lehrbuch der hebräischen Sprache. Braunschweig 1861. S T A D E , B., Lehrbuch der hebräischen Grammatik. Leipzig 1879.

Lehrbuch

Hehr.

Gram.

Lexicon

HWB.

Lehrgeb. NHbCWB.

Lehrbuch Lehrbuch

BIBLIOGRAPHY

XX

ABBESVIATIOK

und S I E G F R I E D , Lehrbuch der neuhebräischen Sprache und Literatur. Karlsruhe und Leipzig 1884.

STRACK

Lehrbuch,

E. A R A M A I C , S Y R I A C , A N D D I A L E C T S BROCKELMANN,

CAKL,

Lexicon Syriacum, Edinburgh et

Berlin 1895. DALMAN,

Grammatik des jüdisch-palästinischen Leipzig 1894. R U B E N S , Traité de grammaire syriaque. Paris GUSTAF,

Aramäisch. DUVAL,

1881. KAUTZSCH,

Gram. EMIL,

Syr.

Grammatik des Biblisch-Aramäischen.

Leipzig 1884. C A S P A R , A Grammar of the Aramaic Idiom contained in the Babylonian Talmud. Cincinnati 1900. Aram. Gram. L E V Y , J A C O B , Chaldäisches Wörterbuch über die Targumim. Leipzig 1881. CWB Targum. M A C L E A N , A. J., A Dictionary of the Dialects of Vernacular Syriac. Oxford 1901. — Grammar of the Dialects of Vernacular Syriac. Cambridge 1895. Vern. Syr. M A R G O L I S , M A X L . , A Manual of the Aramaic Language of the Babylonian Talmud. München 1910. Babyl. Talmud Gram. M E R X , A D A L B E R T S , Chrestomathia Targumica. Berlin 1888. Chrest. Targ. •— Grammatica Syriaca quam post opus Hoffmanni. Ilalle 1867. Gram. Syr. NÖLDEKE, THEODOR, Grammatik der neusyrischen Sprache. Leipzig 1868. Neusyr. Gram. — Kurzgefaßte syrische Grammatik. Leipzig 1898. Kurz. Syr. Gram. •— Mandäische Grammatik. Halle 1875. Mand. Gram. P A Y N E - S M I T H , R., Thesaurus Syriacus. 2 vols. Oxonii 1879. Thesaurus P E T E R M A N N , J . H . , Brevis linguae chaldaicae grammatica, litteratura, chrestomathia cum glossario. Editio secunda. Carolsruhe 1872. Ling. Chald. LEVIAS,

F. A S S Y R I A N DELITZSCH, FRIEDRICH,

Assyrian Grammar.

Translated

from the German by A. R. S. Kennedy. Berlin 1889. — Assyrisches Handwörterbuch. Leipzig 1896,

BIBLIOGRAPHY

xxi ABBBEVIATIOH

MUSS-ARNOLT, W . ,

rian Language.

A Concise Dictionary of the Assy2 parts. Berlin 1905.

Dictionary

G. ARABIC AND D I A L E C T S DOZY, R., Supplément aux dictionnaires arabes. 2 tomes. Paris et Leyde 1881. FRAËNKEL, SIEGMOND, Beitrüge zur Erklärung der mehrlautigen Bildungen im Arabischen. Leiden 1878. Mekrl. Bildungen F R E Y T A G , G. V . , Lexicon arabico-latinum. 4 parts. Halis 1830—1837. LANE, EDWARD W., An Arabic - English Lexicon. Lexicon 8 parts. London 18Ö3—1893. SCHWAUZLOSE, F . G . , De linguae arabicae verborum pluriliteralium derivatioiie. Berlin 18^4. SPITTA - B E Y , WILHELM, Grammatik des arabischen Vulgärdislectes von Aegj^pten. Leipzig 1880. Gram. STUMME, H A N S , Grammatik des tunisischen Arabisch nebst Glossar. Leipzig 1896. Gram. d. Tunis V Ö L L E R S , K . , The Modern Egyptian Dialect of Arabic. Translated by F. C. Burkitt. Cambridge 1895. W R I G H T , W . , A Grammar of the Arabic Language. Translated from the Gorman of Caspari. 3 d edition. 2 vols. Cambridge 1896—1898. Arabic Gram. H. E T H I O P I C Grammatik der äthiopischen Sprache. Zweite Auflage von Dr. Carl Bezold. Leipzig 1890. Äthiop. Gram. — Lexicon linguae aethiopicae. Leipzig 1865. Lexicon K Ö N I G , E., Neue Studien über Schrift, Aussprache und allgemeine Formenlehre des Äthiopischen. Leipzig 1877. STADE, B., Uber den Ursprung der mehrläufigen ThatWörter der Ge'ezsprache. Leipzig 1871. Ursprung d. Ge'ez DILLMANN,

AUGUST,

I. IXDO-GERMANIC PHILOLOGY, ETC. Essai significations). Paris BRUGMANN, C., Grundriß der indogermanischen burg 1897 ff. BRÉAL,

MICHEL,

de sémantique (science des 1897. der vergleichenden Grammatik Sprachen. 2. Auflage. StraßGrundriß

xxii

B I B L I O G R A P H Y ABBREVIATION

C., Kurze vergleichende Grammatik der indogermanischen Sprachen. Straßburg 1902 ff. C U R T I U S , G., Grundzüge der griechischen Etymologie. 5. Aufl. Leipzig 1879. H I R T , H E R M A N N , Der indogermanische Ablaut. Straßb u r g 1900. Indogermanische Forschungen, herausgegeben von C. Brugmann, etc. Straßburg 1892 ff. JOHANSSON, K. F., Beiträge zur griechischen Sprachkunde. Upsala 1891. L A G A R D E , P A U L A N T O N DE, Reliquiae juris ecclesiastici antiquissimae, graece edidit. Leipzig 1856. M Ö L L E R , H E R M A N N , Semitisch und Indogermanisch. Kopenhagen 1906. — Vergleichendes indogermanisch-semitisches "Wörterbuch. Göttingen 1911. P E R S S O N , P E R , Studien zur Lehre von der Wurzelerweiterung und Wurzelvariation. Upsala 1891. S K E A T , W A L T E R W . , An Etymological Dictionary of the English Language. Oxford 1882. W A L D E , A L O I S , Lateinisches etymologisches Wörterbuch. Heidelberg 1910. BRUGMANN,

IF.

Reliquiae

CHAPTER I THE SUBCONSCIOUS BILITEIIAL ROOT I t was J e M d ä ben David Hayyug who, following the moro scientific methods of A r a b grammarians, first laid down the principle that Hebrew roots are invariably triliteral, and thereby developed Hebrew grammar beyond its previous empirical stage. While the predecessors 3 of Hayyug were still in utter uncertainty as to the true nature of the Hebrew root, and held that it might be uniliteral, biliteral, triliteral, or even quinqueliteral, as occasion might require, it was reserved for him and for succeeding generations to establish law and order where confusion and doubt had reigned, and to bring all apparent irregularities in the Hebrew verb within the scope of the triliteral theory. The work of the succeeding schools, both in the Jewish and in the Christian world of Hebrew scholarship, was thus, for many centuries, purely analytic, I t was in the labors of Wilhelm Gesenius that theoretical Hebrew grammar first found a sound exposition, Among the varied results of his studies in Semitic philology the '•root-theory," formulated by him, but more fully developed by his successors, holds a conspicuous place. This hypothesis found its widest acceptance and interpretation in the early seventies of the nineteenth century; but its firmest 3 DnJO rnani? p. 39; see also Bacher,

Anfänge

der

Grammatik,"p. 80 ( Z D MG. xlix, 352). 1

hebräischen

2

ROOT-DETERMINATIVES I N SEMITIC SPEECH

adherents have seen its foundations shattered, and many have surrendered the ground that had been gained through decades of toil. I t is true that the "root-theory" has been weighed in the balances, and found wanting, but this failure is not due so much to the unsoundness of the theory as to the faulty methods employed by its defenders. Efforts are constantly being made to throw more light on the difficult problem of the original character of the Semitic root, and the biliteral theory still claims many staunch advocates. Nöldeke has only recently taken up the question of the long-contested biliteral nouns and weak verb-stems (Neue Beiträge zur semitischen Sprachwissenschaft, Straßburg 1910, pp. 109-207); but while it is true that such scholars as he and Eduard K ö n i g 4 have not despaired because of former failures and still seek to solve the problem, other investigators such as Brockelmann and the younger Delitzsch 5 hold firmly to the triliteral theory and oppose as futile all attempts to support or to recast the biliteral hypothesis. Thus Brockelmann, in his recent Grundriß der vergleichenden Grammatik der semitischen Sprachen, Berlin 1910, p. 285, summarily dismisses the biliteral theory in the following words: "Schon im Ursemitischen wiesen die meisten Wortformen einen festen Grundstock von durchweg drei Konsonanten auf. Vielfach schlössen sich diese "Wortformen wieder zu Gruppen zusammen, die bei verwandter Bedeutung zwei, meist die beiden ersten Konsonanten gemeinsam hatten. Man hat daraus schließen wollen, daß diese zwei Konsonanten einmal eine selbständige Existenz geführt, und daß aus ihnen erst die dreikonsonantigen Stämme sich entwickelt hätten. Dieser Schluß ist aber keineswegs zwingend. Es ist ebenso4 JE. x. 468, and Lehrgebäude ii. (1). 370—374. Delitzsch, Uliwerth und Gefährlichkeit der Wurzeltheorie, in his Prolegomena, 188—198. 6

T H E SUBCONSCIOUS B I F I T E R A L ROOT

3

gut denkbar, daß an dreikonsonaritiae Wurzeln durch analogische Neuschöpfung unter dem .Einfluß von bereits bestehenden Wörtern verwandter Bedeutung neue Gruppen sieh angeschlossen haben. Dieser Vorgang, der sich in einzelnen Füllen sehr wahrscheinlich machen läßt, kann in älteren Phasen schon in viel weiterem Umfang aufgetreten sein; auch auf indogermanischem Gebiet hat ja Bloomfield (IF. iv. 66) ähnliche Erscheinungen nachgewiesen."® There is no doubt that the tendency to form triliteral roots was already strongly implanted in proto-Semitic, but Bloom field's theory, which ascribes the origin of the so-called root-determinatives in Indo-Gennanic to the process of congeneric assimilation, fails to satisfy the student of Semitics; nor has it, even within the sphere of IndoGermanic, succeeded in entirely replacing the part played by the root-determinative. The principle of analogy, upon which the operations of this theory depend, has long been recognized in the science of linguistics, but to suppose that it has been the sole cause of those external and internal root-mo'lificati.ius by affixed elements which are so characteristic of the Semitic languages, is to lay upon this hypothesis a burden more heavy than it can bear. This theory may, indeed, be invoked iri explaining the trend of specialization taken by certain determinatives in the course 6

According to Bloom field'j theory "ers and vers, 'to flow,' may have clashed in some such manner as this: there is a root *er, 'to go,' and this was specialized to connote 'to flow,' having come under ¡he influence of iters to such an extent that it adopted its congeneric character and became *er.s, aided by partial formal similarity; or conversely, a root *ers may have aUVcted a root uer, especially as words denoting 'water' begin with n (IF. iv. 68). Fraenkel has shown how congeneric assimilation lias its counterpart in Hebrew (cf. BA. ill. Hi— P>2), and attributes to such adaptation all changes which others may seek to explain by the theory of root-determinatives.

4

ROOT-DETERMINATIVES IN SEMITIC SPEECH

of root-development, but the origin of the root-determinative itself demands a more tangible theory. Sufficient evidence has been found to justify a belief in the independent existence of a biliteral stem in the hypothetical proto-Semitic at a period shortly before the prehistoric separation of the various Semitic stocks. Before this, however, a few words are necessary concerning the older "root-theory," its errors and shortcomings, and its relation to the present investigation. Hottcher, 7 August Miiller, 8 and Stade, 9 in their endeavors to determine the true character of the primitive Semitic root, began by studying the existing form of the weak-verb and by deducing certain generalizations from it. This method is followed by many recent investigators, 10 aud forms worn down in course of time, and in reality of a regressive character are still held to be the progenitors of all other existing forms. The present writer, on the contrary, has found himself obliged to surrender this theory, and he is convinced that by the time when the proto-Seinites were divided into the stocks known in history, the verb-roots had already become wholly or partly triliteralized, while at least some roots had even then been made pluriliteral; and, at the same time, he is compelled to view with scant favor the theory of the survival of a biliteral, based upon the simplified weak stem. Philippi 1 1 was here perhaps the most accurate of former investigators. He applied the biliteral principle to the strong verb and weak verb indiscriminately, 7

Lehrbuch ii, 476 ff. 8 ZDMG. xxxiii. 698—700. 9 Lehrbuch i. 109 ff. i'1 So Wolfenson, J AOS. xxvii. 003—316, and Alircw ZDMG. lxiv. 161-194. i ' Der Grundstamm des starken Verbums im Semitischen, in Morgenl. Forschng., pp. 71—106.

THE SUBCONSCIOUS BILITERAL ROOT

5

and regarded the latter as the outcome of a further r a t h e r than of a retarded growth. Gesenius, 1 2 Dietrich, 1 3 F ü r s t , 1 4 the eider Delitzsch, 1 5 and, in fact, all the older theorists, confused two distinct phenomena in the early history of Semitic speech, so t h a t they usually regarded as an identical development both root-differentiation and the modification of roots by determinatives. The grammar of Gesenius-Kautzsch has led many others into this error with its conclusion (ij 30) that, "Weitere ^Modifikationen derselben Wurzel erfolgen endlich dadurch, daii sowohl die Wurzelkonsonanten, als der Zusatzbuchstabe nach den Gesetzen des Lautwechsels in einen verwandten L a u t übergehen." I t must be emphasized, from the very first, that the rootdeterminative (with which alone this investigation is concerned) is a constant; and that any apparent variation from the principle is due to the affixing of other distinct determinative elements, which may or may not be related to the first determinative. This confusion of regarding two distinct and separate processes as one and the same has caused L a m b e r t 1 6 to postulate a vocalic origin for all determinatives; and, reasoning in like manner, Duval, 1 7 with strict adherence to the triliteral principle, has made all Semitic roots pass through a graded series of changes, both in form and in meaning, before they assumed their present character. 15

Ilebr.

Gram.

(11th edluonj. p. 65, and the discussion «..¡' tiw roots

VP, 02, t:. r,n, «.-t«. Ki

Abhandlungen, p. 97 und the v n , b ~ ¡ , bt, bs, bo, t, Ki. Kl, m, eic. 15 Jemrun, p. 142, and the comparison bcrwt 'u SB, no, and SB. is De la formation des racines triliieres ioru» in Sem. Studien, p. 358. 17 Gram. Syr-, pu. 32—36. 14

6

BOOT-DETERMINATIVES IN SEMITIC SPEECH

The older theorists had, moreover, no idea of the science of semasiology and would very ingeniously connect, as Friedrich Delitzsch justly shows,18 any one root with any other at pleasure. As a matter of fact, however, the investigator in this field must be rigidly governed by the meanings and uses of words as recorded by the latest and best lexicons. Finally, the older school always sought for analogies in Indo-Germanic, and if they could correlate their results in Semitic with similar phenomena in the other field, they felt doubly confident. The consequences of such analogies were that Julius Fürst and Franz Delitzsch 19 built up a SemitoIndo-C ermanic prepositional theory; while Philippi20 formulated a theory, based on Indo-Germanic, that the base of the strong verb *Jcatab was almost invariably developed from a primitive *kat plus a post-determinative; and Ernst Meier 20» propounded a reduplication theory based on IndoGermanic. Unlike Indo-Germanic, however, where practically only post-determinatives are employed, the number of pre-determinatives, in-determinatives, and post-determinatives is almost equal in Semitic; and these three classes of Semitic determinatives consist, it should be noted, of almost the same group of consonants. To avoid all risk of superficial analogy and the consequent danger of a 18

Prolegomena, p. 189: "Man braucht ja nur die Grundbedeutung einer "Wurzel recht zu verallgemeinern und zu verflößen und jede Wortbedeutung wird aus ihr hergeleitet werden können." i» They identify pre-determinative ti with Sanskrit sam, 'together with,' 3 with Sanskrit abhi, 'to,' etc. {Jesurun p. 143). See the criticism by Renan, Hist. d. lang. Sem., p. 448 ff. Cf. also Fürst, Concordantia, preface p. 8. 20 Morgenl. Forsch'ng., p. 84, It is interesting to note that early Arab grammarians had a similar notion about the Semitic root; cf. Baidhawi in his comment to Sura ii. 4, pp. 18—19. 2,,a Cf. Die Bildung und Bedeutung des Plural, p. ivff.

THE SUBCONSCIOUS BILITERAL ROOT

7

false impression of actual, kinship between Semitic and Indo-Germanic—a kinship which, though often postulated and theoretically possible, is yet entirely unproven 21 —it seems best to disregard Indo-Germanic phenomena in the main body of the discussion of the Semitic root-determinative; although analogical Indo-Germanic developments may be cited by way of illustration after the theses of the present study shall have been established from the point of view of Semitic philology.

i The Triliterality of the Weak Verb—Cause and Origin of the Weak Stem. Early theorists, as has already been noted, for the most part began their investigation of the character of the biliteral root from a consideration of the weak verb and, as a result, they invariably reached the misleading conclusion that the weak verbs in question were surviving specimens of older biliterals, Ewald 2 2 was of the opinion that the weak consonant in verbs T'J? was vocalic in origin; and that the form *qum was probably the type from which these verbs were derived. Böttcher, 23 following a similar line of thought, formulated a series of able arguments by which he sought to prove that forms like 3D and riD antedate forms like and nin, and that they are extant remains of biliteral stems. The arguments of Böttcher have, however, of late been refuted by Lambert, 2 4 who has clearly shown that Böttcher confuses the concept of a primitive biliteral with an historic biliteral, which was the 2 ' Cf. Hirt, Indogermanen, pp. 83—84; and see Möller, Semitisch und Indogermanisch, Kopenhagen, 1906: Vergleichendes indogermanischsemitisches Wörterbuch, Güttingen 1911. 11 Lehrbuch, pp. 293—295. » Lehrbuch, ii. 476—478, 492 ff. 24 La triliteralite des racines, BEJ. xxxv. 203—212.

8

ROOT-DETERMINATIVES IN SEMITIC SPEECH

product of linguistic decay: sab has arisen from *sababa, as qam from *qaivama. Stade 2 5 and almost simultaneously August Müller contend that the strong analogy exerted by the pure triliteral has forced the weak verb into its present irregularities. The latter scholar says in p a r t : 2 6 "Ursprüngliche zweiradikalige Wurzeln sind im Semitischen bei dem Durchdringen der Analogie der dreiradikaligen, den letzteren dadurch gleichwertig geworden, daß entweder der "Vokal oder das zweite (gelegentlich auch das erste) konsonantische Element in der Aussprache verstärkt wurde: qäm—qam. oder sab—sabb; yäqüm—yaqüm oder yasub—yasubb, yissub." Both Nöldeke 2 7 and Wellhausen 2 8 later adopted this theory, and Gesenius-Kautzsch 2 0 have accorded it a certain sanction. I t must, however, be remembered that Wright, 3 0 Philippi, 3 1 König, 3 2 and Brockelmann 3 3 refuse to accept this view, and insist, in conformity with the labors of A r a b and later Hebrew grammarians, that any deviation of the weak verb from a triliteral character is due to the attrition of speechforms which constitutes a well-known linguistic phenomenon. Brockelmann 3 4 has summarized the principal arguments against this most recent school of biliteralists, and clearly shows that the media y class in Hebrew was independent of the media w, and was not, as Ewald thought, an abbreviation from a h i f i l form of media w; that forms like *mawt, "qaivm, and *bayn are older than the forms *qum

« Lehrbuch, pp. 109-110, 230-291. m ZDMG. xxxii;. 700. " ZDMG., xivi. 77t>, and Kurz. Syr. Grata., p. X. Skizzen und Vorarbeiten vi. 250. 23 Rehr. Gram., p. 175. 30 Comp. Gram., p. 243. 31 Morgenl. Forschung., pp. 71—106. s2 Lehrgeb., ii. (1) ¿73. 31 " Grundriß, i. 605 ff. op. cit., pp. 606—607.

THE SUBCONSCIOUS BILITERAL ROOT

9

and * bin, which are derived through ablaut and secondary contraction; that the hypothesis of Stade and Miiller misunderstood the character of the w and y sounds, which are not, as Stade maintained, 35 labial and palatal spirants, but are consonantal vowels; and, finally, that the arguments, adduced by Stade, that true X'V and V'V verbs exist only in such forms as rawah and 'aya'Ai, were practically annihilated by Noldeke, who has shown that all these verbs are secondary denominatives. The very latest attempt to regain the vantage point of the last biliteral school and, at the same time, to extend its conclusions to the whole category of irregular verbs, is represented by Ahrens, in his study "Der Stainm der schwachen Verba in den semitischen Sprachen' (ZJDMG. Ixiv. 161—194). So novel are some of his arguments and so misleading are some of his deductions, that a detailed consideration of his theory becomes advisable, as well as an attempt to discover the main source of its error. He proposes, at the outset, to find in the weak verb the solution of the problem of the biliteral root in Semitic speech, and he proceeds to sum up clearly the well-known facts of the close affinity of all the existing classes of weak verbs, N"B, n'S, 2"S, ">rj?, yy, or He shows that the Hebrew of the Old Testament often has a double conjugation for the same verb, with the same meaning, and in the very same passage, among such cases being, iP'lD and Bteto, 'to feel', in Gen. xxvii. 21—22; T.3 and 112 in in^STl« vnsi, 'I will be an adversary unto thine adversaries' (Exod. xxiii. 22); Jip"1 and 'to be removed,' in Ezek. xxiii. 18; a metaplastic nittf, 'you abide,' in Jer. xlii. 10; and S2>n and BHK in UtfiT 'he will be threshing it' (Isa. xxviii. 28). 35

op. cit., p. 65.

10

BOOT-DETERMINATIVES IN SEMITIC SPEECH

One class of verbs is actually confused with another in many of its forms. Thus, several hof'als of 1"B and V'B are identical in Hebrew; the Syriac 'af'el of N"B is analogous to V'B,7 so that r** 'to be lost' becomes A7 and the Biblical A r a m a i c has r n a i n (Dan. vii. 26); HiJ 'to rest' forms two hifils, n ^ n and n\Sn, as if from two conjugations, etc. A h r e n s very correctly draws the conclusion that, underlying all these interrelated forms, there must be a common biliteral. Besides this permissible inference, however, he seeks to find traces of an original biliteral stem which he holds to be still extant, and this biliteral, he believes, constitutes the base of all the reduplicated (pilpel) forms. Y e t the fact t h a t all these forms are related to one or another class of weak-verbs, should cause us some hesitation in accepting this explanation as final, especially since the loss of the weak letter in these reduplicated forms is of the same nature as its loss in certain forms of the weak verb. I t is true that many of these pilpel forms, as A r a b i c hudhud, 'hoopoe,' Assyrian laqlaqu, 'stork,' sarsaru, 'grass-hopper,' and Hebrew baqbuq, 'flask', are purely onomatopoeic in origin, yet this is no certain criterion of their antiquity. Such formations, based on older analogies, may have been made at all periods. Of all Arabic dialects, the Egyptian is especially rich in such forms and many of its reduplications are certainly of relatively modern d a t e 3 6 . Popular language is fond of such formations as is shown by the dialect of the Babylonian T a l m u d 3 7 and by Modern Syriac; 38 Spitta-Bey, Gram., p. 190. 31 New pilpel formations in the Babylonian Talmud include 1313, 'to expend,' 'to confuse,' pSJi?, 'to sprout forth,' Dlfil, 'to glow (of the twillight), be semi-conscious' (cf. DIN, 'to be red,' Dan, 'to be silent'), plIJI, 'to divide into small parts, go into details,' "irnn, 'to reflect upon,' "irnn, 'to kindle, burn,' Btatfo, 'to bespatter,' 'to mock,' etc.

THE SUBCONSCIOUS BILITERAL EOOT

11

and it must also be borne in mind that the vocabularies of the various dialects always include weak verbs which keep alive in the Semitic mind the specific meaning that the reduplicated forms are intended to convey. All that can be conceded safely in this argument is, that the Semite invariably rejects, whether consciously or subconsciously, the weak element as subordinate, when such reduplications are formed. Ahrens then applies his results to an analysis of the weak verb, and finds that the monosyllabic forms are primitive, while the dissyllabic forms are of later formation. He feels himself obliged to draw this conclusion because of the inability of historic grammar to account for the total disappearance of waw or yodJi in the perfect of the hollow verb; because of its failure to explain the aphaeresis of ivatv, yodh, nun, and 'ate/ in the various imperatives of the

defective verb; and especially because of the meaningless confusion of stems that is so characteristic of the weak verb in general, and that has been hitherto entirely ascribed to analogy. This, perhaps, is tlie crucial point of the whole problem, and if sufficient reason can be found for the changes in question, the long-mooted problem will be finally solved. Ahrens raises the further question of the primitive noun, and rather daringly premises that n"i? verbs, such as TO, 'to command' and fllj?, 'to wait for,' are denominatives from biliteral nouns, such as IS and Ip. I t is true that ¡"IT, 'to throw,' looks very much like a denominative from T^s, 'hand'; yet it might be alleged with equal plausibility that ¡VII, 'to live,' is a denominative from T!> 'living,' HIJJ, 'to bend,' from y , 'heap,' or JT3, 'to be parched,' from

38 Another example often quoted is rnn, 'to be pregnant,' associated by some with in, 'hill' (cf. Levy, NHbCWB).

12

ROOT-DETERMINATIVES IN SEMITIC S P E E C H

'beast of the desert.' Either the whole class of H"1? verbs is denominative in character, or none at all. Finally, Ahrens reaches the conclusion that the narrative tense (the consecutive imperfect in Hebrew) and the jussive are the oldest existing forms of the verb, to which may be added the imperative; whereas the perfect stem is a later formation. These three are monosyllabic in character in the Hebrew weak verb. This conclusion is, however, open to question. In the regular verb, as well as in the weak verb, the jussive displays at least vocalic shortening 39 ; and the imperative and the jussive, as being forms most frequently used in the living language, might be expected to suffer aphaeresis sooner than forms of less frequeut occurrence. Moreover, the shortening of the consecutive imperfect in Hebrew seems to receive an adequate explanation in the theory of Brockelmann, 40 that the consecutive imperfect has been developed from the pure imperfect through a shift of accent, which still rests on the prefix in an open syllable, as in "insl, ]B;i, etc. On the whole, Ahrens has admirably summarized all the arguments in favor of the hypothesis of Stade and Müller, and has given these arguments their widest interpretation; but while it must be confessed that the explanation of the irregularities of the weak verb is still beset with certain difficulties, it is equally clear that the theory of a surviving biliteral is inadequate to interpret the phenomena presented by the weak verbs in question. To understand the development of the weak verb in its true light, and to grasp the real nature of all the changes M Ci'. Wright, Comp. Gram., p. 191. Tiie jussive lias ordinarily 110 characteristic vowel, and uses the shorter terminations. Ungnad (BA. YI. iii. 57—58) believes that the imperative and jussive originate from the second or intransitive form of the verb (qatil). Grundriß, i. 557.

THE SUBCONSCIOUS BILITERAL ROOT

13

and deviations which it has undergone during its existence, a more consistent solution is necessary, and one which will not only be linguistically sound, hut which will also explain each successive step in the evolution of the weak verb. Briefly to summarize the situation, it can scarcely be denied that a comparison of such kindred forms as ^¡VT, 1J5"!, tO"l and ¡"DT, 'to beat down, trample,' din, Dfiil, nan, and on J, 'to be turbulent, roar,' Dil, DI2T, and HQ"!, 'to be silent,' and QOn, nan, and DIT, 'to be hot,' legitimately leads to the postulation of a common biliteral base for these various interrelated forms. We may then hold, with König 4 ', that this biliteral base, as the fundamental root, is a theoretical abstraction which never actually existed in the living language; or we may prefer the other alternative, maintaining that the biliteral root once had an independent existence, and that it developed into its present state by the affixation of formative increments or determinatives. In either case, the weak element in the forms under discussion is certainly the determinative, as is obvious from the fact that certain verbs in one dialect may belong to one class of weak stems, while closely kindred weak verbs in another dialect may belong to another class having only a slight difference of meaning. Thus aaT, 'to be many,' in Hebrew is o^.; in Syriac; |ril, 'to give,' in Hebrew is in"1 in Phoenician; 21Ö and 313}, 'to be good,' in Hebrew coincide as 2KB in Aramaic and Syriac. Nöldeke has given a very full list of such related forms from all Semitic idioms 42 . Some of these are post-Biblical "1DJ, 'to saw,' Syriac jjju, but Ethiopic and Arabic idem; Hebrew "IjJJ 'to hew out, pierce,' Syriac 'to dig,' but Arabic 'rent,'

« Lehrgeb. ii. 370, and JE. x. 469. « Neue Beiträge, pp. 179—206.

14

ROOT-DETERMINATIVES IN SEMITIC SPEECH

and Ethiopic Old: 'barefoot,' Syriac

or a)i'to

pierce, dig'; Hebrew i]IT,

'to go barefoot,' but Arabic

idem; Hebrew 2(53, 'to bore a hole,' also

'cellar,' but

Arabic iJ^Sj, 'to intrude'; Syriac a o j , 'to adhere to,' and Arabic uas^, 'to stand firm,' etc. Even in the same dialect one stem of a verb sometimes follows the analogy of one conjugation, and another stem of the same verb that of another conjugation. Thus, the Hebrew ^¡1, 'to go,' conjugates the imperfect as ^ and, in like manner, 233, 'to stand,' is found in the hitpa'el as 'to station oneself,' (Aramaic 'firm, right'); Syriac primae I verbs pass into priinae » in the 'af el and ' e t t a f ' a l (and in the Saf el of 'to delay,' which becomes *£&*); in Syriac the verbs «-f., 'to know,' and 'to sit,' are identical with 3"S in the imperfect, as is JJT, 'to know,' in Aramaic; the Syriac vei-b -¿u, 'to learn,' becomes .a\» in jthe pa'el participle; the Hebrew rn3, 'to rest,' forms two hifils on the analogy of V'J> and 3"S. I n the dialect of Tunis 4 3 , Stumme has shown that ljda, 'to take,' has four root variants, while kla, 'to eat,' has five: J ^ j , JX*, j y , and ^ K All these changes have thus far been attributed solely to the action of analogy upon the various weak forms, yet it is doubtful whether analogy alone has caused so hopeless a confusion of forms. The different classes of weak verbs are not merely related formally, through the possession of a common weak element, but are a h o related semasiologically, since this common weak element gives an identical connotation to their biliteral base. I t may, indeed, be " WZKM. viii. 262, and Oram. d. Tunis., p. 21. « For a complete list of such cases of root-confusion, see Huizinga, Analogy in the Semitic Languages, pp. 16—20.

THE SUBCONSCIOUS BILITEIU.L ROOT

]5

argued that this weak element (or determinative) gives no connotation whatsoever to the biliteral base bat that it is a mere meaningless accretion, yet the first position is far more logical and is confirmed by a close study of the semasiological value of each determinative. I t would accordingly appear 'that the various weak elements at first had somewhat different functions as determinatives of the biliteral base, and that they gradually acquired, in course of time, and through the constant interaction of Semitic words upon each other, the closest identity of meaning, while, on the other hand, their individual primary connotations vanished at an early period from the consciousness of the Semites, If this theory is correct, all weak forms which eventually coincide after having started from different points can be explained only by assuming an identity of meaning in their determinative elements, at least in later usage. Thus the pol'él forms in Hebrew in the two conjugations V'JJ and JTJ>, and the identical hof'als in V'S, VJ?, and JJ"}?, cannot be merely the result of their analogous possession of weak elements '14a, but must also be due to the subconsciousness of the identity of meaning conveyed by the prefixed y, the infixed w, or the gemination of the final letter. I n other words, we must at once assume that the biliteral root still made itself felt in the consciousness of the people, at least in the weak verb. K ö n i g 4 5 has suggested such a possible solution of the question: „Hängt mit dem secundaren Ursprung des einen von den drei Stammconsonanten nicht auch dies zusammen, daß die schwachen Yerba in ihrer Flexion im Verlauf der Sprachentwickelung vielfach in einander übergingen?" 44a Cf. Stade, Lehrbuch, p. 122. Lehrgeb. ii. 370.

46

16

ROOT-DETERMINATIVES IN SEMITIC SfEECH

A t first this seems a very radical conclusion but, when the peculiar phenomena displayed by the weak verb are borne in mind, such a theory is alone seen to be able adequately to account for every irregularity. This hypothesis is, moreover, apparently confirmed by evidence from many sides. The disappearance of a nun in the imperative of 18*53, 'to approach,' of a lamed in npS, 'to take,' or p^>D, 'to go up,' of an 'elif in the Arabic 'to take,' 'to eat,' and 'to command,' is not due merely to the strong analogy exerted by the imperfect of the verb, as Earth and Philippi maintain 4 6 , but is still more indebted to the subconscious recognition of the secondary character of the weak element, or determinative. The tendency to form triliteral bases, strong though it was, could not yet completely dominate the consciousness of the people who spoke the living language, but the weak elements were mercilessly dropped, not because they were weak, but because they were subordinate. I t is true that the stronger consonants were better fitted to survive, yet some even of their number were still subconsciously felt to be subordinate in character; and it is equally true that each Semitic dialect asserted its individuality in syncopating or in retaining these elements. The strong tendency to assimilate and to syncopate the various consonants in the verbs N D, V'1S, 3"B, V'1J>, JTJJ, and 'T'b has its origin mainly in such subconscious identification of the underlying biliteral element; and phonetic law alone will find it difficult to account for all these involved changes. The consonants n, I, and h are not necessarily weak; but when a subconscious force, recognizing that the determinative stands in a secondary relation to the biliteral root, is once posited, the tendency in the weak verb will be " BA. ii. 365, and ZDMG. xliv. 692.

THE SUBCONSCIOUS B I L I T E R A L ROOT

17

exerted in the direction of reversion to the original biliteral element. The conclusion which we thus reach is, therefore» the reverse of the view expressed by the biliteral school: SD and no are later than *sababa and *ma%ut,i'1 and their biliteral character is, biologically speaking, a reversion to type. A number of other considerations lead to the same conclusion. Ethiopic has most faithfully preserved the original triliteral type of the weak verb 48 - When it is remembered that the Ethiopic was separated from the Arabic branch of the Semitic family at a very early period, it is not strange that it displays the older forms Hlfl : and CR:, whereas the Arabic has already suffered contraction (JUXo and ^Ú). An isolated group, surrounded on all sides by languages of an entirely different structure, is more liable to retain an older family characteristic than is a language which has kindred dialects on every hand, and which is constantly exposed to the encroachments of cognate idioms. Friedrich Delitzsch* 9 has shown similarly that certain peculiarities common to Assyrian and Ethiopic can be explained only as common survivals from proto-Semitic, and Haupt expresses a similar view with regard to the priority of the Assyrian present and Ethiopic imperfect. 50 There is no doubt that the marked tendency of Ge'ez for new rootformations, as Stade 5 1 has shown, is likewise a survival « Similarly Scerbo in the Giornale della Società Asiatica Italiana, xxii. 299. Dillmann, being a firm adherent of Ewald's views, and not finding in Ethiopic a biliteral form the rule but rather the exception, assumes that this dialect has gone further in its development, and that it has given the original vocalic element a consonantal character (Athiop. Gram., pp. 113—115). i» Hebrew Language, p. 18. so JBAS. (New Series) x. 246. 51 JJrsprung d. Geez, p. 3. 2

18

ROOT-DETERMINATIVES IN SEMITIC SPEECH

from the older period.

Again, since the longer forms of

the weak verb occur most frequently in poetry (cf. VDH, V^J), they are clearly to be regarded as archaisms52.

The

consonantal character of the w and y is, moreover, shown by the fact that all the other root-determinatives are conFinally, the method of the later

sonantal in character.

formations of reduplicated verbs from earlier triliteral stems conclusively proves that a subconsciousness of the original biliteral root still survived within historic times.

This is

especially the case in the modern Arabic and Syriac. Thus, in the dialect of Egypt 5 3 , xjtj, 'to be bright, sparkle,' is connected with

and y b ) , 'to shine'; i_ÄJüLS, 'to shudder

with cold,' is formed from l J j , mourn, groan,' from Syriac 65

'shiver';

'woe!

woe!,' 54

and J p ^ ,

etc. In modern

'to be feverish,' is formed from

hot'; v^il, 'to smoke,' from JL'iJL, 'smoke'; bloody,' from 14 j, 'blood';

'to be

y*»;,

'to

be

'to tread down', (cf. tJ>n,

'to tread'); and mi, 'to push' (cf. T { , 'to squeeze'). verbs may be cited

'to

Prom

'to be gratified,' from 1j¿¡, 'to

be pleasant'; and -¿A¿.V, 'to lick up,' from JL¿¿£, 'to lick'. Sometimes even W

verbs lose their weak letter

as

'to cut up,' from "»-A-®, 'to cut'; and xKxl, 'to push, beat' (cf. J?¡?n, 'to thrust').

I n all cases, the biliteral has been

faithfully preserved, while the determinative has suffered attrition, the result being what we have ventured to call reversion to type. The problem of the biliteral noun is in place here, and the solution seems to be similar to that of the biliteral 52 König, Lehr geb. i. 524—525. 53 Spitta-Bey, op. cit., p. 190. The etymology of the particle being forgotten, the y was treated as if it were the weak element of a hollow verb. 55 Nöldeke, Neusyr. Oram,., pp. 189—190.

THE SUBCONSCIOUS BILITERAL ROOT

19

verb. Philippi 56 had already attempted to prove that such apparent biliterals were in reality triliterals and Barth 5 7 later confirmed his results in greater detail. König s s expresses similar opinions, although Nöldeke 59 has collected a large number of forms, drawn from all the dialects, to support the possibility of the survival of the biliteral noun. The problem need not, however, be discussed here, since its solution is largely dependent upon the explanation of the biliteral character of the weak verb. ii Denominative Triliteral Stems. Not only does a study of the life and development of the weak verb give evidence of the subconscious existence of the biliteral element, but a study of the life and development of the secondary denominative still further strengthens the theory here maintained. I t accordingly becomes necessary to make a study of the secondary triliteral stems in Semitic, as well as an attempt to elucidate the way by which these secondary triliteral denominatives have been derived from older triliterals, and the means whereby these secondary formations still seek to retain a triliteral character. It will thus become possible to ascertain both readily and accurately what elements in each root were still felt, whether consciously or subconsciously, to be primary, and what were regarded as secondary; what formative elements became radical in the process of development, and what others still retained their formative character. In this manner, we may gain a clear and well-defined view of the formation of new triliterals 5« ZDMO. xxxii. 73 ff. w lb. xli. 603 ff. 58 Lehrgeb. ii. 372. Neue Beiträge, pp.

59

109—178. 9*

20

ROOT-DETERMINATIVES IN SEMITIC SPEECH

•within the historic period, and one which will aid materially in casting light upon prehistoric methods of development. In Biblical Hebrew is found the very curious form ^ruTl 'mocked,' (I Kings xviii. 27). Although its secondary character may still be called into doubt by many, who attribute the peculiar vocalization to Aramaic influence 60 and view the däges as complementary, 61 yet it is more correct to assume, as does Ewald, 62 that ^riH was no longer, in the consciousness of later generations, a hif il from —a stem dead, to all appearance, in the other conjugations—but rather that it had gradually assumed an independent character, as the Masoretic pointing, in its various stages, indicates. Thus iVnijn occurs with the dage§ retained in Job xiii. 9; but Jer. ix. 4 reads ^ruj, (pause); in fibrin, in Judg. xvi. 10, there is no trace of the dage§; and the derivative substantive D^nrj, in Job xvii. 2, completes the chain of development. 63 Similar in origin is the aira£ Aeyofitvov "inn in the expression in Isaiah xlvii. 13 where the Ketib reads r o n . This "an is generally connected with the rare Arabic root "to cut in large pieces," and hence the passage is rendered 'the dividers of the heavens.'64 The root is best explained, however, as a denominative from a hif il of 113 'to separate, sift, purify,' which is found in the infinitive form "OH1?! sift' (a synonym of ni"l6, 'to winnow') in Jer. iv. 11. The expression

so si 62 63

Gesenius-Kautzseh, Hebr. Gram., p. 181. König, Lehrgeb. i. 352. Lehrbuch, p. 337. Of. also Stade, Lehrbuch, p. 112.

6* So Gesenius-Zimmern-Buhl HWB., p. 155, and Franz Delitzsch, Commentary on the Prophecies of Isaiah ii. 225.

THE SUBCONSCIOUS BILITERAL ROOT

21

taken from such a figurative use of the word "15065 would thus mean 'those who sift the h e a v e n s a n appropriate metaphor for an astrologer; and the form would thus be connected with a recognized Hebrew root, instead of being an independent offshoot from a dubious source. Hitzig 66 is right in the main when he attempts to read VlSH, 'to sift,' although the change in the traditional text 6 7 is unnecessary. A similar o.ira$ Xeyo^evov is ^IH, 'to tread down,' Job xl. 12, which is generally compared with a vulgar Arabic ¿¡SJi, 'to tear down,' while Budde reads '¡Jin, from *[3"[,68 although this again is unnecessary. Here also a secondary denominative hif'il stem may be postulated, either as actually alive in the earlier period, or as alive to the consciousness of later tradition. Contrary to the opinion of Barth, 69 , pJJ, 'to fix, establish,' is a pure type of denominative from HiOft, 'equipment, establishment,' and its formation, as will be shown below (p. 22), is very frequent in the later language. I t is very probable that 'to be alienated,' is a nif'al denominative from (Ezekiel xxiii. 22, 28), which occurs in the form in Ezekiel xxiii. 18. Nöldeke comments on this verb as follows70."Es ist unwahrscheinlich, daß die Sprache selbst das perf. 65 The Talmudic rnari, "syllable" is a contracted form of ¡TUM, 'enunciation,' as Levy (NE'bC'WB i. 449) correctly maintains, and despite Franz Delitzsch (op. cit. ad loc.) is not from this root. 66 Das Buch Daniel, p. 29. 67 Among some of the very curious emendations of '"jail, that of Diilmann, Der Prophet Jesaia, p. 422, is interesting. He reads which he connects with the Arabic 'to know.' «8 Gesenius-Zimmern-Buhl HWB. ad loc. 69 Wurzeluntersuchungen, p. 52, but Stade, Lehrbuch, p. 113 and Lambert, Sem. Studies, p. 358, hold the view which is favored by the present writer. 70 Neue Beiträge, p. 198.

22

ROOT-DETERMINATIVES IN SEMITIC SPEECH

Ton J"S, das imperf. von Y'fi gebildet hätte," and he thinks that both verbs are independent formations. Similar nif'al denominatives in the making occur in forms such as 'to melt,' Judg. v. 5, which is best derived from Vbt with a loss of the compensative dages 7 1 through the influence exerted by 1"B verbs, and which would thus stand for the usual iVt}. In like manner, n ^ J l (Isaiah xix. 3) is formed from ¡"¡¡?55, and means 'to become empty,' and HSS3, 'to be overspread,' Gren. ix. 19, is from ^SB, 'to break up.' 1 2 The loss of the däges in these and many similar forms can be explained only by assuming both a strong inherent tendency to simplify weak stems, and the counter-influence of the i"fi verbs. In Neo-Hebrew, the language of the Mishna and Midrashic literature, similar phenomena occur with great frequency. 73 An interesting development is presented by the new denominatives with preformative fl, these including 'to ' i Cf. König, Lehrgeb., i. 343; Ewald, op. cit., p. 511. " Cf. König, loc. cit.; Gesenius-Kautzsch, Hebr. Gram., p. 182; Ewald, loc. cit.; Olshausen, Lehrbuch, p. 592; Böttcher, Lehrbuch, ii. 488; and Stade, Lehrbuch, p. 241. Further examples are raw (Ezek. xli, 7) and Dr6ö2? (Gen. xvii. 11). T3 This phenomenon is of considerable importance in determining the much disputed connection between Neo-Hebrew and the Biblical dialect. When it is borne in mind that new denominatives of the type under consideration are found everywhere in the later popular Semitic dialects, that the controversy in question depends partly upon this very denominative type, and that the formations are in all cases analogous, the only possible inference appears to be that Neo-Hebrew was the result of a long process of development from the Biblical language (cf. Segal, JQR. xx. 647 ff.). I t would, therefore, seem that the dialect of the Mishna constituted a vernacular, as was argued by Graetz (Liter aturblatt des Orients, 1844, p. 824; cf. Luzzatto, ibid., 1846, p. 830 ff.), and that it was not merely a scholastic and artificial language, as has been maintained by Geiger (Lehrbuch zur Sprache der Mischnah, p. 2) and Strack and Siegfried (Lehrbuch der neuhebräischen Sprache, p. 4 ff.).

THE SUBCONSCIOUS BILITERAL ROOT

23

offer up a heave-offering' (riaiin), JTinn, 'to blow a broken blast' (njttin), ^nnn, 'to make a beginning' (¡"6Tin), and g^fin, 'to become wormy' (y'j'iPi).74 In like manner, with preformative a, "l?a, 'to be spoiled, blemished,' may be a denominative from Old Hebrew Tita 'wound,' 75 while DITO, 'to be turbulent,' is formed from Haina, 'turmoil.' Similar quadriliterals are DJ^fl, 'to interpret, translate,' probably from *D11 (cf. Arabic 'to conjecture'), and ]3Dßnn, 'to become poor' from )3Dö, 'pauper,' but these will more appropriately be considered in the discussion of quadriliterals. Denominatives from feminine nouns are filSK, 'to verify,' from nax, 'truth' (*]ÖK), and f\Vp, 'to sift flour,' from 'fine flour' 7 e (*^D). 77 In Aramaic are found "ISPI, 'to carry on trade,' a socalled taf'el from 13«, 'wages, pay'; 7 8 "IHJJ, 'continuous,' from W , 'to turn'; 7 9 nin, 'to sigh,' a secondary denominative from niN (Hullin 57, a); and "IM*, 'to leave over,' has suffered attrition from the safel of *ini, 'to Root (Levy. NSbCWB, iv. 632); J^ln has already developed a denominative verb in the Biblical D^nD (Nahum ii. 4), but Friedrich Tx Delitzsch does not connect it with the Arabic ¿J 5, 'to lick,' but rather with an Assyrian form signifying originally 'to gnaw' (Hebrew Language, p. 66). " So Metmann, Die Hebräische Sprache, p. 22, but the word may be most likely related to the Biblical *nn, 'to be blemished,' from which 11ÖO, 'bastard,' is derived (cf. Arabic 'to be foul'). Similarly the Biblical in Psalms cxix. 103 is not, as is generally believed, from Arabic 'to slip', but is a pure denominative from ns^n. * ™ Levy, NHbOWB., iii. 538. " For lists of such denominatives see Metmann, Die Hebräische Sprache, pp. 22—23, and Strack and 8iegfried, Lehrbuch, pp. 72—73. '8 Levy, NHbCWB., iv. 627. '» So Merx, Chrest. Targ., p. 295, but Kautzsch, Aram. Gram., p. 112, holds that the root here is Tltf, 'to send.'

24

ROOT-DETERMINATIVES IN SEMITIC SPEECH

overflow'. Finally, from JH)2, 'knowledge,' are derived the denominative jnti, 'to have knowledge,' and the form jnififltfNI, 'he recognized' (Onkelos and Pseudo-Jonathan on in Gen. xlii. 8).«o 'to trade,' is found as Similarly in classical Syriac the taf'el from ri^l; while o-j!, 'to sigh,' becomes u^U, from which is formed the derivative noun JLSJqI, 'sigh'. 81 So, too, the assimilations 'to he left over,' for filK*^ and jJil.ll, 'to be taken,' for ¡¿IUI both show the process of denominative formation in its incomplete stages. 82 Noldeke supposes 83 that which is found only in the pa el in the meaning 'to defile,' may be a Saf'el denominative from 133. The feminine noun 'wrath,' forms the denominative 'to fill with wrath';®4 and in like manner, kx», 'to sweat,' is formed from the feminine noun t^-oj, 'sweat.'85 Other denominatives are j, 'to be made head,' from JjJuj, •first' (cf. JU-i, 'head'), and v ' t o be made lord,' from lyi, 'lord.' The modern Syriac dialects furnish abundant material. Denominatives from substantive forms are v 'to be foolish,' from Jij'i-*, 'devilish' (cf. Ijl*, 'devil'); 'to be peaceful,' from Old Syriac jiXwt, 'peaceful' (cf. 'peace'). 86 New taf'el formations are ?!/, 'to come to oneself,' from Old Syriac h^UI; 'to become like a stone, numb,' from ^oUI. 8 ' The new causative formations from the old af el participle in » are also of interest. Here there is so Dalman, Gram. d. Jiid.-Aram., p. 201, but Merx, Gram. Syr. p. 227, however, thinks that the safel yiDE> is a dissimilation for in Syriac, from which jniDntfKl is derived. si Noldeke, Kurz. Syr. Gram., p. 113 note. 82 Cf. Wright, Arabic Gram., i. 77. 83 0p_ c j( t) p, 127 note. si ibid., p. 42; cf. also Payne-Smith, Thesaurus, i. 1299. 85 Noldeke, Neusyr. Gram., p. 188, note. 86 Noldeke, loe. cit. Noldeke, op. cit., p. 195.

T H E SUBCONSCIOUS B I L I T E E A X ROOT

25

generally an assimilation which, according to Noldeke,8S is to be attributed in the main to phonetic changes, although these changes are insufficient to account satisfactorily for all the phenomena in question. The causatives under consideration retain their » throughout, so that it becomes almost an integral part of the root. The Jl"a causatives generally lose the I, as in 'to make captive,' from 'to bind'; j« which is derived from ¿1>, 'smoke.' Finally, in Ethiopic,

Dillmann 9 9

derives the

following

denominatives from a causative stem in h , which may be hardened to 0 through the influence of the following radicals: Afn4»:, 'to encircle firmly,' from ¿ L i , 'to become narrow'; O i d : , 'to wrap up, bind,' from JlS, (V, V I I ) ; ÄHil: from 'to

flow';

Odi.: from ' i j , HBin,

from 'to weaken,

Dip, 'to arise'; relax.'

Some

and forms

retain the old causative prefix fl and are thus made

de-

nominatives, such as tlUd:, 'to measure with a span,' from r n j , 'span'; 100 A i m ; , 'to seduce,' from nnB;

and

([¿A:,

'to be superior,' as compared with DUS, 'face, rank.'

Many

are formed through the addition of a final 1", these being substantive denominatives.

Such are Rill": and R f l t : ,

«6 Quoted by Nöldeke, Neue Beiträge, p. 180. 96 ZD MG., xxv. 685. 97 ibid., 1. 622. 98 ibid.. xlix. 507. 99 Äthiop. Gram., p. 118.

100 c f . Praetorium, BA., i. 44.

'to

28

ROOT-DETERMINATIVES IN SEMITIC SPEE

CB

swim'; ftihl" :, 'to have full power over,' from fl®vh 'power'; and hV"!";, 'to uncover,' from HD3, 'to cover.' Praetorius 1 0 1 shows that 'to take pity,' is a denominative formation from 'mercy,' a word of frequent occurrence in the Sabsean, from the root j>r*] 1 0 2 and similarly w ^ g 'to simulate,' is a denominative formation from and is derived from the root ííi, 'to stray.' This brief exposition of the triliteral denominative verb has revealed an interesting process at work in all the important divisions of the Semitic family. Two distinct and diametrically opposite forces have cooperated in the formation of the new denominatives under consideration: decomposition and growth. On the one hanl, the weak triliteral verb, under the external pressure of a slowly encroaching formative increment, was gradually forced to contract, and to surrender its weak element, thus becoming biliteral; while, on the other hand, this secondary formative element was amalgamated with the quondam triliteral root, and became an integral part of it, at times even obscuring the original root. The two forces which have brought about this phenomenon appear, at first sight, to be the inherent trend towards triliteralism, and the impulse towards simplification or phonetic decay. Yet this is only partly true, for though a small proportion of the quadriliteral denominatives have thus been shortened, by far the greater number have withstood the tendency towards triliterality. I t may practically be said that such contraction has been restricted to the weak verb. Here the weak elements were ruthlessly dropped, scant distinction being observed in such elision as to the initial, medial, or ultimate position of the weak «i BA., i. 21—22. «o» So also Hommel, ZDMQ., xlvi. 566.

THE SUBCONSCIOUS BILITEBAL ROOT

29

element in the root. Thus ¿"B, K"E, yy, TJ>, and il"1? verbs have all lost their weak consonant at one time or another.103 The trend towards triliteralism was plainly of itself not sufficiently strong to displace pure radical elements, and to incorporate servile ones in their stead. There is no inherent reason why the n of the J"B verbs should yield before the encroachments of m or i; nor does the principle of phonetic decay explain why a waw does not change into its equivalent vowel instead of vanishing utterly. These changes can be satisfactorily explained only on the hypothesis that the weak consonants, unlike the others, had no tangible existence in Semitic linguistic consciousness. In other words, we must return to our old hypothesis, and must posit a subconscious identification of the biliteral root, which has always been preserved. As in the case of the weak verb, the original portion of the new denominative stem reverts to a simplified biliteral form through the subconscious dropping of the older servile elements, or determinatives—again a case of reversion to type. The most instructive feature here is the new formative increment of the denominative stem. Within the historic period itself, subordinate elements with definite values have become primary and radical, and have lost their original connotations. This fact might seem to contradict the results which we have already attained, since the weak elements of the weak verb have survived as subordinate, while the later subordinate elements have become primary. Yet there is really no contradiction, for the weak elements were more pliant than these later accretions, and yielded more quickly to phonetic decay, falling at once into the subordinate positions which they have continued to hold. There is every 103 This is especially the case in the maf'el formations in NeoSyriac.

30

ROOT-DETERMINATIVES IN SEMITIC SPEECH

reason to believe that this adoption of formative elements into the root has been going on unnoticed for ages, and that it was an extensive factor in the original triliteral formations. ill Root-Differentiation. In the process of investigation of Semitic roots, Semitic scholars have from time to time noted 104 a phenomenon which can best be described under the term root-differentiation. Root-differentiation involves gradations in the intensity of the consonantal sounds of a root for the purpose of modifying its meaning. As has already been noted (p. 5), Gesenius was one of the most important of early theorists to give wide attention to this phenomenon. Besides pointing out the affinity between groups of roots which differ from each other in the intensity of their root-consonants, he also noted a corresponding change in the meaning of such roots: 105 "In der Regel ist mit einem solchen Lautwechsel zugleich eine Nuancierung der Bedeutung verbunden." He gives as example the differentiation of the well-known 'to hew down,' DD¡?, 'to cut down,' 'to peel off.' The same root appears with a dental in 3BJ?, 'to cut away, destroy,' and Y]¡?, 'to split,' and softens into *D3 in HD3, 'to cut off.' "With further softening the root becomes NJ, 'to shear,' then flattens to TU, 'to make an incision,' and finally is gutturalized into BOH, 'to pierce,' Tin, 'to split,' nnn, 'to break up,' etc. The early biliteralists extensively employed this theory in the solution of many etymological problems. Thus the lexicow* Of. Geaenius-Kautzsoh, Sehr. Gram., p. 99; Renan, Bist. d. lang. Sem., p. 97; and Duval, Gram. Syr., p. 32 ff. 105 Cf. Geseniua-Kautzsch, ibid., Hebr. Gram. (11th edition) p. 66.

THE SUBCONSCIOUS BILITERAL ROOT

31

graphical labors of Fürst 1 0 6 and the elder Delitzsch 107 abound, for the most part, in this somewhat dangerous association of roots presumed to be of near kinship. While there are many difficulties in the way of accepting a theory based upon such exceptional linguistic phenomena, traces of whose existence have been but casually noted in another linguistic field, the Indo-Germanic, 108 still many contemporaneous scholars subscribe to such a view. König calls attention to its importance in the following words: 109 "Ein hinreichend sicherer dritter Pfeiler jener Brücke (des Ursprungs der Sprache) ist der Parallelismus, welcher zwischen der wahrgenommenen Stärke einer Tätigkeit oder Empfindung und dem Grade der Anspannung der Sprechmuskeln beobachtet wird." I t will not be difficult to find many instances in Semitic to prove that the primitive Semite characteristically expressed the intensity of natural acts and sensations by means of modifications of sound, although this is quite outside of the scope of the present work. A few well-known examples from Hebrew will make the point in question clearer. Thus, the verb ^33, iio 'to break apart,' dentalizes 106 In his Hebräisches und chaldäisches Handwörterbuch über das alte Testament. 107 In his Jesurun. 108 Cf. Indo-Germanic *ltier-, 'to twist, turn,' and 'to bend, turn;' *gel-, 'to swallow up,' and *ger-, 'to eat up' (Latin vorare for *guorare). See Persson, Wurzelerweiterung und Wurzelvariation, p. 30. Brugmaun ascribes these parallel forms to the phonetic phenomenon of dissimilation in reduplicated forms (e. g. * gergel, for *gerger, gave rise to both *ger- and *gel-) (Cf. Grundriß, i. § 282). Some other examples may be *plu-, 'to flow,' and *bhlu-, 'to flow' (Persson, op. eit, p. 131), *Jcer-, 'to howl,' and *hel-, 'to howl;' *mu-, 'to groan,' and *bu-, 'to groan' (op. ext., pp. 12—14). •09 Hebräisch und Semitisch, p. 7. no Renan, op. cit., pp. 96—97.

32

ROOT-DETERMINATIVES IN SEMITIC SPEECH

to nriB, 'to crumble'; softens to

in "1X3, 'to crop,' and

then to N3, 'to tear away'; and dentalizes to nna, 'to cleave,' or "H3, 'to separate.' DDI,

and

T o a like category belong

'to split, crumble, crash'; as well as

'to lap,' and ijn1?, 'to lick'; while another instance is furnished

by pjn, 'to cry,' pgj, 'to complain,' Mtí, 'to roar,'

and niS, 'to howl.'

A still better example is f)BÇ, 'to beat

or drum,' and IBB, 'to drip,' which softens to *3t5 in 'to beat in, impress'; appears with a weaker dental in a? 1 !, 'to drip, prattle, mutter,' and is found, with change to a sibilant, in 21T, 'to ooze, in the other l a n g u a g e s , 1 1 2

flow.'111

Parallels may b e found

for the phenomenon,

doubtless,

' n Geseiiius is of opinion that the roots with stronger consonants are the older, and that those with weaker consonants are of later development (Hebr. Gram., p. 66, 11*1» edition). Yet, granting that such an impulse for the formation of roots existed on a large scale, it is difficult to determine which type is older and which is younger. It would perhaps be more correct to assume that many of these forms existed sido by side from the very start. i n Taking these phenomena as a basis, Duval has elaborated a theory of consonant gradation intended to solve the problem of the biliteral root. He holds that one group of consonants has been changed into another to give an additional shading to the original meaning of the root, and he is accordingly enabled to connect (Gram. Syr., pp. 34—35) l»p, i ^ , o , ^ , -ibj>, Ami-, n»j>; ij*», i s » , ^qju, ••. ^ ; ¿ce!, IB»«, "isp; and »íj, i^» as direct descendants from a single triliteral root. His explanation of the consonantal changes involved runs as follows : "La palatale qof (de H?p) passe en la gutturale », celle-ci s'affaiblit en aleph, aleph lui-même se consolide en youdh ou tombe et disparaît de sorte que la racine "TOp peut en dernier lieu devenir "iB> ou avec renforcement de la sifflante "fit. La deuxième radicale sin, peut permuter avec une autre sifflante plus faible ou plus forte, se fortifier en la muette dentale .¿, après une emphatique. La troisième radicale, la liquide r, permute avec la deuxième qui est une consonne faible ou s'aspire, et en s'aspirant s'affaiblit, ou se consolide, suivant les exigences du sens, en une gutturale ou autre consonne." The writer regards this explanation as too confused to deserve any further consideration.

THE SUBCONSCIOUS BILITEKAL EOOT

33

dates back to the prehistoric period, inasmuch as the demonstrative roots arrange themselves in similar groups of graded sounds. We thus find, throughout the Semitic dialects, demonstrative elements in I, m, and ». "We should naturally expect one in r to complete the series, and this is actually to be found, in all probability,113 in the Mishnaic and Talmudic particle nn, 'here, behold,' and in many other dialects, especially in the Syriac 'höre,' and 114 'there.' Dillmann has shown that a similar gradation exists in Ethiopic, which has the demonstratives fa, da, sa, and za in all four forms. Interesting as a study of these phenomena would be, they can be considered in the present study only in so far as they serve to check and correct the results obtained in the identification of the consonants which function as rootdeterminatives. For it must be noted, although Gesenius and his followers often overlooked the fact, that in the process of root-differentiation at most, only two consonants of the first root can shift to form a new root, while the third remains constant; 115 and this principle affords an 113 Barth, Sprachwissenschaftliche Untersuchungen, ii. 27, The present writer is unable to agree with Brockelmann (Grundriß, p. 324), and Nöldeke (WZKM., viii. 265, note) (but cf. Hommel, ZDMG., xxxii. 714, note) who regard this r as remnant of the imperative of ra'd, "to see." 114 Äthiop. Gram., p. 104. u s This oversight has lead investigators into the error of arranging both the determinative and the biliteral into groups of so-called related forms. Thus Lambert, Fürst (Lexicon, pp. 2, 377, 528), and others believe that the vowels a, i, and u gave rise to all determinatives: a to K and the gutturals; i to 1 and the palatals, and u to 1 and the labials; while, in addition, Lambert believes that the 9"9 verbs were differentiated to final r and formed the source of liquids, dentals, and sibilants (Sem. Studies, p. 358). The correspondences between the different roots that have the same determinative must be noted.

3

34

ROOT-DETERMINATIVES IN SEMITIC SPEECH

additional proof of the existence of an original biliteral element. Thus, in the examples already cited, the root 'to slay,' shifts to 'to plunder'; 13(5, 'to cut, shorten,' to -ita, 'to cut'; j>ej>, 'to cut off,' to via, 'to lop off'; rra, 'to cut,' to na, 'to chop'; 'to bruise,' to J>?3, 'to crop'; ^SB, 'to peel,' to ins, 'to divide'; Aramaic t(?J, 'to pierce,' to D?3, 'to slay'; 'to pluck off,' to fpa, 'to wound, revile,' etc. It is quite clear that, while the root contains at the most two variants, there is a fixed constant which may be either prefixed, infixed, or suffixed to the variable root. This constant is the root determinative.

iy Biblical Folk-Etymologies. Kenan was the first to note that in the folk-etymology of names in the Bible, the biliteral element of a root played the important part. He accordingly says: 116 "II est remarquable qu'on y suppose presque toujours la biliteralite primitive des radicaux: ainsi ]'|5 joue avec WIR (Gen. iv. 1), ni avec Dnj (Gen. v. 29), etc." König 117 has returned to the subject and has called attention to its importance. Little faith can of course be placed in folk-etymologies, yet they certainly have one value; that the derivative form is an approximation, at least in sound, to the original with which it is compared. Accordingly, the Biblical etymologist, confining himself to Thus, n^j, 'to lift up,' and r6n, 'to hang,' Aramaic Btpn, 'to pierce,' and Tin, 'to sharpen'; Hebrew ontf, 'to slaughter,' and intf, 'to bribe' (bestechen). »6 Eist. d. lang. Sem, p. 126. i » JE., x. 468. »8 An example of such a popular etymology is the German Sündflut (flood of sin) corrupted from Old High German sinfluot, 'big flood' (cf. Weigand, Deutsches Wörterbuch, Gießen 1910). In English, sparrow-grass for asparagus, and rarebit from rabbit are examples (cf. Century Dictionary and Cyclopaedia, 4974 and 5798).

T H E SUBCONSCIOUS B I L I T E R A L ROOT

35

the evidence of sound, conforms in his derivations to strict phonetic law. Anything that is harsh to the ear must necessarily be avoided. If, therefore, the y in )"!¡? had been a phonetic obstacle to its etymology from í7J¡5 such an etymology would not have been made. It is evident that the author did not consider his etymology to be a phonetic impossibility, especially as he may have had in mind the analogous formation of J5^, which, though etymologically connected with W is, nevertheless, closely related to ilM. If, then, a subconsciousness of the biliteral root once be granted, the derivation of )"!¡5 from rtijj, of ni from Dfti, 119 and the pun on JIB} from ílfiB (Gen. ix. 27) are all phonetically plausible. 120 As a summary of the results thus far attained, it may be said that evidence adduced from various quarters indicates that the Semitic triliteral root had its origin in a former biliteral root. A subconsciousness of the older bins E v e n here t h e Biblical etymologist was in e r r o r ; for, while a root ~\/nh, 'to rest' exists, yet. the root oru, when associated with Dm, Dili», and nan, shows a common biliteral root yhm.

H a u p t thinks

that n m is a privative pi'el f r o m Dm, to sigh (cf. BA., vi. 2, note 22). The Rabbis feel the difficulty in t h e etymology and ask very pertinently: " E i t h e r let the text read 12¡vy, 'he will relieve us,' or let it be innJ, who will comfort us ( u a n r ) " (Genesis Rabba, xxv. 2). 120 I t is remarkable t h a t the Rabbis, as a general rule, in their popular etymological derivations in the A g g a d a always seek out the biliteral characteristic of the word to be explained and generally fail to consider t h e third radical of any importance (an exception is in the passage quoted in the above note). They seem to follow in this the precedent set by the Biblical etymologist. Thus, in Genesis Rabba, xvi. 1, the etymology of the names of the great streams in the garden of E d e n is as follows: "Pishon, because it rears flax (piston); H i d dekel, because it is keen in sound (had qol); and P r a t h , because its waters constantly increase ( p a r a ) . " Instances of such etymologies are so numerous that it will not be necessary to enter into the matter any f u r t h e r . 3*

36

ROOT-DETERMINATIVES IN SEMITIC SPEECH

literal character still survives in the weak verb. Analogy itself, to which alone the tendency towards triliteralism is due, could not obliterate all traces of the former biliteral character of Semitic roots. Even the strong verb, where the determinative has become a fixed and permanent element, still shows evidence of the subordinate nature of some of its consonants, especially in the case of the liquids I and n. The recognition of these facts affords a new point of view in the treatment of the weak verb: that it was the product of a regressive rather than of a retarded development, as the older biliteral school maintained; and that its present biliteral form is simply the skeleton of its quondam biliteral form. The existence of such a prehistoric biliteral can no longer be doubted. The problem now before us is the investigation of the way in which this biliteral developed into its present triliteral form.

CHAPTER II A STUDY OF PLURILITERALS AND THE RESULTING THEORY

Many have maintained that the only method by which the investigator may safely determine the early process of root-formation in Semitic is detailed study of the growth and development of triliterals into a pluriliteral form. As a matter of fact, there can be no doubt that, in the majority of instances, the pluriliteral is a direct development from the triliteral. There is, accordingly, justification for the belief that the triliteral was developed from the biliteral in a manner analogous to that by which the pluriliteral was evolved from the triliteral; for, if the roots were expanded beyond the triliteral limits to which a rigid analogy had confined them, we may reasonably suppose that the growth in question was, from the very first, continuous along the same lines, so that the present triliteral form of the root was a logical sequence of its former biliteral character. I t was Dietrich, 121 one of the older biliteralists, who was among the first to recognize clearly the unmistakable affinity between the root-determinative in the pluriliteral and the root-determinative in the triliteral. Gesenius-Kautzsch divides the quadriliteral stems in Hebrew into the following classes: 122 —a) those which have an added or developed 121 Abhandlungen, p. 326: "Daß mit der bisherigen Beschreibung (der pluriliterae) zugleich die Entstehung der triliterae aus den biliteris vorgezeichnet ist, liegt zutage." 122 Hebr. Gram-, p. 101.

38

ROOT-DETERMINATIVES IN SEMITIC SPEECH

consonant in some part of the root; b) those which result from the fusion of two triliteral stems ; and c) those which are evidently survivals of verb stems which once had an independent existence (e. g. iilVti', 'to blaze,' an old âaf'êl from 'to burn'). As regards class b, the Hebrew examples of which are jn"]BS, and 'Ubbs, Dietrich has shown 123 that such a fusion is contrary to the laws of Semitic philology, and that it occurs only in popular expressions where the two fused elements are in constant use, as in ,ito,7!$ ^ B , 'a certain one,' which are combined into "•ibbs;12* whereas the other two examples are, no doubt, the result of direct development from single triliterals. 1 2 5 This class of pluriliterals may thus be entirely eliminated from the discussion, which will be restricted to the other two classes, those with a developed consonant, and those which show survivals of older stems. 12 « These two classes are, in all probability, identical; for, it can scarcely be supposed that the developed elements of the pluriliteral root are of comparatively late origin, or that they do not date from the early Semitic period. The fact that the same formative elements are to be found in every branch of the Semitic family, instead of being restricted to a particular dialect, strongly militates against such an hypothesis. Moreover, the generic terms for well-known domestic animals 123 op. cit., p. 304. 124 Merx (Gram. Syr., p. 147) is of the opinion that many quadrillerais originated from such a fusion of two triliteral stems, and Leguest (Études sur la formation des racines sémitiques, Paris 1858) thought that he could reduce all triliteral roots to a fusion of two biliteral elements. The Arabic furnishes the greatest number of instances of such fusion (cf. etc.). 125 Dietrich, op. cit., pp. 284, 296. 128 The reduplicated pluriliteral forms ( b o ^ , ctc.) will be omitted from the following discussion (cf. Gesenius-Kautzsch, op. cit., p. 101).

A STUDY OP PLURILITERALS

39

and cultivated plants show these formative elements in the earliest period, so that they may be regarded as survivals of older stems, the special connotation of which had vanished at an early date. The forms that display an infixed nasal or liquid (m, n, I, or r) are everywhere regarded as dissimilated intensive stems, and their nasal infix may be considered the characteristic of the stem. Lagarde has expressed the opinion,127 which is supported by much evidence, that such methods of conjugation were very extensive in proto-Semitic, and that they also included all the various conjugations now found in the separate languages. For example, the VIII form of the Arabic is also found in the Moabite of the Mesa stone in the form Dringt. Lagarde writes: "Einer besonderen Untersuchung muß vorbehalten bleiben festzustellen, ob nicht in hitpa'el und den parallelen und entsprechenden aramäischen Bildungen eine weitgehende Entstellung des alten Besitzes der Sprache vorliegt."12§ In the following discussion, in which, for the sake of exact terminology, the pluriliteral stems will always be referred to a triliteral base pä'al, both nominal and verbal stems will hold equal rank; since, in all likelihood, the formative elements in both are identical in origin, just as they are of identical form. Moreover, primary and secondary pluriliterals are of equal importance, for the latter display, in even clearer fashion, the process of root-formation. 127

Bildung d. Nomina, p. 10 (Abhandlungen K. Gesellschaft der Wissenschaften zu Goettingen, xxxv. 1888). 126 See also Lagarde, Symmicta, i. 125, note 2. A better example is furnished by the name of the divinity Istar, Phoenician mntrp, Palmyrene in», Sabsean "lilTO. This is best derived fro_m a common protoSemitic VIII form of IE»», Aramaic in», Arabic 'to be luxuriant, abundant,' with the usual metathesis of t$ to it (Of. Hoffmann, Abhandlungen K. Gesellschaft d. Wissenschaften zu Güttingen, xxxvi. 22) (Of. s i s t f r n i w ? , Deuter, vii. 13).

40

ROOT-DETERMINATIVES IN SEMITIC SPEECH

i Semitic Pluriliterals. A.

Hebrew.

The

two

puzzling

forms i n H e b r e w u p o n which S t a d e h a s b a s e d a n e w causative stem, the taf'el, i. e. ^ J I R

( H o s . xi. 3), 'I t a u g h t to

walk,' and r n n n n (Jer. xii. 5), 'thou wilt compete,' are b e s t derived from nouns with preformative R , 1 2 9 precisely as the loan-word Dä"]Pl, 'to interpret,' is a denominative formation from t h e A s s y r i a n targumänu,130 from ragamu,

w h i c h is p r o b a b l y derived

'to shout, call'; the t is in all c a s e s a n orig-

inal reflexive element.

Of saf'el forms the H e b r e w p o s s e s s e s

t w o 1 » 1 — r D n b t f , 'flame,'132

aDd

r n m ^ p t f (Lev. xiv. 3 7 ) ,

pressions,' the root of t h e l a t t e r b e i n g seen in y o ,

'de-

'to b e

deep,' and also in r n j $ , 'flat-dish,' while a saf'el form with a

privative sense is doubtless p r e s e n t in "l}5p, 'to deprive

of

light, b l i n d , ' 1 3 3 which is b e s t derived from 113, 'to be

light,'134

Syriac

'crocus,' must still b e con-

ns Such nouns as Neo-Hebrew nnnn, 'competition,' and an hypothetic (cf. Neo-Hebrew ni^J"), 'habit'). See Barth, NB., p. 279, note. So Delitzsch, Hebrew Language, p. 50, note 1, but Muss-Arnolt (.Dictionary, p. 1192) thinks that targumänu is in tarn a loan-word from the Aramaic löfifl; there is no root *Dn, however, in Aramaic with the meaning of 'to speak,' but Arabic has 'to conjecture' (p. 23). 131 Cf. Stade, Lehrbuch, p. 122. Wright, Comp. Gram., p. 204. 132 Barth, Etymologische Studien, p. 49, seeks to connect this with the Arabic cLil-V"', 'flame,' but this is untenable for phonetic reasons (¡J* = Hebrew s 3 ), and his objections as to the non-existence of Hebrew causative 8 disappear when we assume that the form is a survival from a very early period of the language when 'to 'enflame,' existed. Nöldeke, however, thinks it is an Aramaism (Sem. Sprachen p. 13). 133 So Hoffmann, ZATW., 'ii. 68, note 1; but Ewald thinks that it is a causative, from "I1JJ, 'to blind' (Lehrbuch, p. 320), while Halevy (BEJ., xi. 66) considers it to be from a root *"l?p. 134 Another saf'el derivative is cy^Di 'locust,' literally 'one who swallows' (König, Lehrgeb., ii. 404). 130

A STUDY OF PLURILITERALS sidered a haf'el derivative from

41

and the etymology

of Delitzsch, 1 3 6 which connects it with the Assyrian latu,

'reed,' must be abandoned.

habasil-

An 'af'el form is seen

in the common Semitic term for 'mouse,' Hebrew "I3DJ?> Arabic from

Syriac 137

the root being generally derived

Infixed elements are still more

common.

Such are the par'el forms t s ^ l ^ , 'sceptre,' 1 3 8 as compared with t s ^ , 'rod'; DDIS, 'to tear and

off,' from

DD3, 'to clip';

'bemantled' ( I Chr. xv. 27), as compared with

*^>33, 'to bind'; while 'locust' from

, Syriac

Jf-l^fi, Arabic

bop,' as well as ¡"JEJ2"ID, 'bough,'

from HBJ?D, 'bough,' may be similarly explained. Delitzsch 1 : 1 9 likewise believes that

Syriac JAj;I,

Arabic

'hare,' is derived from *3JX, 'to jump, leap,' inferred from the Assyrian word for 'hare,' annabu.

Among pal'el forms

may be mentioned I J i J ^ , 'quiet,' as compared

with

'at ease, quiet';

HBJjbl, 'raging heat,' from f)JJJ, 'to rage'; 0 9 and Utt1?}, 'hard, barren,' Arabic 'rock,' from the

root

'to become hard.' 1 4 0

classed as dissimilated pi'els,

141

These forms are generally although it is difficult to

discover their particular intensive connotation in many of the cases noted.

With sufformative b, we

find

Ara-

maic K^B'lJ?, Syriac I»**-, Assyrian erpu, 'cloud,' from *)"]J7, 'to drip, 1 4 2 rain,' and 'plantation,' from D^S, 'vine135 Of. 'onions'; but see BA., VI. iv. 104 (from *yßn). 136 Prolegomena, p. 82 and Halevy's criticism in BEJ., xiv. 149. »si Levy, NHbCWB, iii. 645, Barth, NB., p. 207. 138 Of. Lagarde, Reliquiae, p. xlvii, Brookelmann, Grundriß, p. 510, and Gesenius-Kautzsch, op. cit., p. 238. is« Prolegomena, p. 114 and Barth, NB., p. 222. Of. König, Lehrgeb., ii. 472—473. 14' So Gesenius-Kautzsch, op. cit., p. 101, 238 and König, loc. cit,, who terms these infixed liquids "Übergangskonsonanten." i " Völlers {ZA., xvii. 311) thinks that the suffix here is but this is untenable, for the I is as much a formative element as m or r.

42

ROOT-DETERMINATIVES IN SEMITIC SPEECH

yard.' Quadriliteral substantives with sufformative D or 3 are fully treated in all standard grammars, 143 these including nVp, 'ladder,' from ^D, 'to lift up'; D^K, 'porch,' from 'to be in front'; and the adjective D1"^ 'naked,' from Ity, 'to be exposed.' Substantives with sufformative 'nail of finger,' 3 are of frequent occurrence, such as ]pj, 'axe,' etc. 143 Quadriliterals with sufformative $ include t^MP, the common term for 'spider,' Syriac Uoiij., 'asp,' Aramaic NJVMS, 'spider or asp,' Arabic (Spider,' the root apparently being 'to spin'; 144 t^lO^n, 'flint,' Assyrian elmesu,145 'diamond,' from D^n, 'to be hard, sound'; ^»"111, 'sickle,' from D"in, 'to cut away'; and CBD1, 'to be young, fresh,' is, according to Konig, 146 akin to 'to be moist.' Among the derivatives with a labial sufformative may be cited 'scorpion,' Syriac Arabic Jlyt*, which is possibly derived from *"lpy, Arabic Jis, 'to tear, pierce'; 147 and ^BJJ, 'bat,' best taken 148 from Arabic J L i , 'to be cloudy, dark.' B. Aramaic. In his study of the Aramaic of the Talmud Babli, Levias 149 has reached conclusions with regard to the quadriliteral verb which are very similar to the theory which is here defended; and he justly regards these quadriliterals as survivals of primitive stems that at one »3 Cf. Gesenius-Kautzsch, op. cit.. pp. 237—238. i " So Levy, NHbCWB, iii. 645. Delitzsch, op. cit., pp. 85—86, note, Noldeke, ZDMG., xl. 728. Ryzicka, on the other hand, thinks that both B^Bjin and tttoin are dissimilated from * »art (cf. BA., VI. iv. 79). 146 Lehrgeb., i- 203. I t is, however, much more probable that BteB") has developed from *B>£t3, 'to be wide, expand,' with preformative r. Gesenius-Zimmern-Buhl (p. 691) and Gesenius-BrownDriver (p. 936) regard the form as a metathesis from Levy, op. cit., iii. 689, Barth. op. cit, p. 207. 148 Levy, op. cit., iii. 634; cf. the Hebrew HQ^S, 'darkness' (Genesis xv. 17). no Aram, Gram., pp. 158—163.

A S T U D Y OF P L U R I L T T E R A L S

43

time had a more extensive usage. Such stems are formed by prefixing n, D, 5, or R, or by infixing 3, 1, B*, or Fi. The stems with n, Vf, or D, are regarded by Levi as 149 as old causativos, and those with 3 or fl, although now active in force, as old reflexives. The saf'el forms include ^nW, 'to change,' DflDB', 'to shut in,' a r f r a * , 1 5 * » 'to flame,' and 0I5SJI2', 'to confuse'; saf'el forms are 3rnp, 'to hasten,' 'to rule,' and *?3(?D,151 'to meet'; the haf'él is seen in pavj, Syriac 'to hold as true'; 1 52 the taf'él in nc'sn, 'to chop off,' and |>3"1PI, 'to sprinkle'; and the naf'él in JñBfa, 'mingere,' Hebrew J ^ H . Par'él forms are 32"]^, 15 3 'to 154 'to sift through.' cause to drag,' pHH, 'to close in,' 'to be naked' (Arabic Jl»£, 'to be destitute'), and p3"lD, 'to be choked'; pan'él, Y ] 3 2 , 'to go beyond bounds'; while the pas'él is seen in "Iptf?, 'to inquire into' from "l¡53, 'to seek.' Forms with sufformatives 1 5 5 include the adjective in¡?»p, Syriac Aj^uaro, 'reddish,' as compared with NjJOp, 'red'; the nouns N'jpS, 'iron,' KPi^D"!«, i»« 'widow,' and Syriac ¡ K * ; ^ , Arabic J X I i , 'candlestick'; while u s Aram.

Gram., p p . 158—163.

•so aniiS, 'to be bright, aflame,' is not, with Petermann, (Ling, chald., p. 12) a cansative form of arA, akin to anbtf, but rather a pal'el form of 'to shine, be like gold' (cf. Mechilta on Exodus xvi. 5), which fell under the influence of inbvf, 'to flame up,' and assimilated itself to it, thus acquiring the meaning of 'to burn.' See also Barth, Etym. Studien, p. 49, note. 151 Of. P e t e r m a n n , Ling. Aram,., p . 201.

Ohald., p . 12, a n d D a l m a n , Gram.

d. jüd.

152 But this may perhaps be a Hebraism (Cf. Yollers, ZA., xvii. 326—327). 153 Levias, loc. cit., and Petermann. loe. cit. 154 On its connection with S5J, see Kohut, Aruch Completum, ii, 349. 155 Cf. Dalman, op. cit., p. 133. is« Cf. Margolis, Bab. Talm. Gram., p. 25, and Ryzicka, BA., VI. iv. 104.

44

ROOT-DETERMINATIVES IN SEMITIC SPEECH

1(21B, 'to lie on the back' 157 (cf. Kijpns, 'vertebral column'), perhaps has a suffixed 0. Syriac. The Syriac shows similar conditions. 168 Among the äaf'el forms may be cited V^t», 'to let loose,' 'to complete,' r-«*, 'to delay,' and 'to inform'; and the saf'el forms include nw, 'to bring before,' and »¿nffl, 'to giye nourishment,' (cf. Aramaic "©g, 'to heal'). A denominative taf'el is found in y^-l, 'to teach,' from 'disciple'; an 'af'el, the causative connotation of which was no longer felt, is 'to find'; and a secondary maf'el is 'to pawn,' from 'pledge.' Among po'el and pay'al forms are ¡-»"^-M, 'to be twisted,' from r*.-v, 'to twist'; f^'oiU, 'to lean on a staff,' a denominative from (^¿«ii, 'staff'; and s^ucp, 'to support,' from 'to think'; 159 and the par'el forms include "Hs,;^, 'to roll,' 'to lop off,' 'to twist' (from Vax), and u ^ j j , 'to rejoice' (cf. 'to be delighted'). A new pan'el formation occurs in ¡ p i ^ , , 'mighty,' from ij^Ss., Hebrew 113J, 'hero.' Most interesting of all are the pa'ali 160 forms, in which the sufformative is Such are -fenj, 'to estrange,' which is evidently a denominative from an adjective J-jp«^ 'of strange descent'; 'to put under,' from the adjective 'lowest'; and - U , 'to take into domestic service,' from 'belonging to the household.' Similar formations occur also from pure roots which have no adjectival form associated with them, such as 'to be astounded, confused' (cf. Hebrew 'to frighten'), - ä ^ , 'to exile,' etc. I t would seem from these examples that 'si A better example with suffixed d is perhaps Hebrew mbBl, 'raft,' from *DB1 (König, Lehrgeb., ii. 404). is» Cf. Nöldeke, Kurz. Syr. Gram., pp. 125—127. 159 Brockelmann, Grundriß, p. 510. wo Cf. Merx, Gram. Syr., p. 233.

A STUDY OF PLURILITERALS

45

the adjectival y was analogous in origin and in signification with the formative y in the quadriliteral, and that both gave the root a denominative force which converted it, as occasion demanded either into a substantive or into a verb. Noldeke shows instances of other denominative verbs with various substantive suffixes, such as ,xm»£, 161 'to sieze firmly' (substantival w), and la»?, 'to shape, form,' a denominative from lUAj, 'form, image.' C. Modern Aramaic and Syriac. These dialects contain a large number of pluriliteral formations, especially denominative forms, though only a few can be noted here. In Mandaean, Noldeke 1 6 2 includes among the saf'el forms tilts', 'to bring into wrath,' amis', 'to broaden,' tmif, 'to disappear,' etc. Saf'el forms are 2!"!"©, 'to hasten,' and bpDD, 'to smooth, stroke'; pan'el and pam'el forms 163 are "UiiO, 'to cut down,' a^O^n, 'to heat oneself' (cf. .-.^S, 'to love'), and 'to destroy' (cf. 'to destroy'); while a par'el form is seen in the adjective ^tDl«, 'naked' (cf. Arabic j £ e , 'to be bare'). In modern Syriac, 1 " we may regard the form .aN»«,, 'to change,' as a Saf'el; t-^i«», "to make to totter,' from ^ i ' , 'to shiver,' as a saf'el; «-.lit, 'to be boiled to pieces' from ^.li", 'to seethe,' as a taf'el; and ««nalt*, 'to be proud,' from 'to shine,' as an 'eStaf'al. Maf'el formations are the regular causative stems, 166 as '¿A**, 'to draw out,' x^ajp, 'to make to sparkle,' etc. Among the pan'el and pam'el forms may be cited «»¿a, 'to So Noldeke, loc. cit.; but Payne-Smith (Thesaurus. i. 1335) thinks that it is a pam el form from ^.m*, 'to prevail.' Mand. Gram., p. 212. 163 Noldeke, op. cit., p. 76, and Brockelmann, loc. cit. 184 Noldeke, op, cit., p. 128. is' Noldeke, Neusyr. Gram., pp. 188—199. 166 Maclean, Vern. Syr., p. 106.

46

ROOT-DETERMINATIVES IN S E M I T I C SPEECH

roll' (cf. Aramaic TU3, 'to roll'), and 'to hew out,' from ouo^, 'to hollow out'; and the par'el forms include •asjrö,167 'to sprout forth,' 'to embrace' (cf. Hebrew pan, 'to embrace'), V W , 'to twist,' 'to pulverize,' from u a i , 'to grind to powder,' etc. A peculiar pa'mel form exists with an infixed » after the second radical. Examples are 'to fold up,' as compared with r °U around,' as compared with 'to wind'; and V*«^» 'to roll.' The denominatives from substantives with sufformatives include 'to give light,' from jjio^, 'giving light, shining'; and 'to be leprous,' from U a ^ , , 'leprous'; and forms with sufformative I are seen in 'to give,' from ocm, and ^ j i - , 'to be afraid,' from 'to fear.' «s D. Assyrian. In Assyrian, Brockelmann 169 thinks that the quadriliterals balkatu, 'to tear down,' harmatu, 'to destroy,' and parSadu, 'to flee,' are instances of par'el and pal'el forms, although parSadu is rather a development from paraSu, 'to fly,' with a suffixed d. An interesting phenomenon is presented by certain pa all formations in which the ending has a denominative force, these including roi?B, 'to be wide, open,' from pallcu, 'wide,' and HD1B, 'to cease,' from paraku, 'to shut off.'i™ Delitzsch 1 '! thinks that the stem KTTS, to be bright,' may be an instance of a new formation with sufformative K. 167 is not, as Payne Smith believes {Thesaurus, 3956) a saf'el from but rather a par'el from *1»D, 'to spread apart' (So Nöldeke, op. cit., p. 191). 168 Only in the phrase JL^'.j' 'do not fear,' in the dialect of Urmiah, but in Ashita and Lower Tiari in all tenses (Maclean, Dictionary of Vern. Syr., p. 84}. Grundriß, p. oil. 170 Cf. Delitzsch, Assyr. Gram., p. 319. 171 hoc. cit.

A STUDY OF PLURILITERALS

47

E. Classical Arabic. The Arabic language has retained the impulse towards the formation of quadriliterals to a much greater degree than any of the languages of the North Semitic group. The subject has been ably dealt with by Dietrich, 172 later by Schwarzlose (De linguae ardbicae verborum plurüiteralium derivatione, Berlin, 1854), and at considerable length by Fraenkel (Beiträge zur Erklärung der mehrläufigen Bildungen im Arabischen). Both Dietrich 1 7 3 and Schwarzlose, 174 who naturally followed the older biliteral school, have seen in the mode of formation of these quadriliterals striking analogies to the way in which the triliteral root has been developed from an older biliteral type. The present investigation corroborates this view. A n a f a l form 1 7 5 is seen in t l ^ s v i , 'to pierce holes in' (of a canker-worm), f r o m i l j ^ - , 'to perforate.' Fraenkel 176 holds that the e of the 'af'al forms has arisen from an original hamza, and it is possible that the development may have taken place at a very early period. Here belong 'yellow,' and ytxa», 'to paint yellow'; 'to leave a camel unguarded,' and 'to permit one to do as he pleases'; and J i o ^ , 'thick', the same as J^O'l, 'thick.' Fraenkel 1 7 7 cites a large number of hafal forms including 'voracious,' from gJo, 'to swallow'; and ^ j i , 'fleet wolf or stork,' from ^Jj", 'to glide.' Although Fraenkel 1 7 7 reaches no definite conclusion as to the origin of the s, "ä Abhandlungen, pp. 305—326. "3 „p. cit., p. 326. »« op. cit., p. 3 ff. Several more examples are given by Levias, Aram. Gram., p. 162, note. 176 op. cit., p. 1 ff. op. cit., pp. 13—16; see Völlers, ZA., xvii. 326—327 concerning the origin of causativc prefixes; also Wright, Arabic Gram., i. 36.

48

ROOT-DETERMINATIVES IN SEMITIC SPEECH

nevertheless, in view of the fact that many of these haf'al forms are merely substantive formations of late date, while a few appear to be old causatives, it seems safe to infer that an early tendency to form causative derivatives in * was here retained from the older language; and although the * died out in Arabic as a causative element, its use as a formative element was retained because of the influence of older analogies. Similarly retained forms in the safal include17» 'to throw down flat on the back,' 'to dash to the ground,' and UtilHi, 'to swallow.' Infixed formations are frequent in Arabic, and it is especially interesting to note that the infixes in question may stand either after the first or after the second radical. Par al forms 179 are 'heel-tendon,' as compared with L ^ J L S , 'heel,' and J¿¿f, 'to contract, curtail,' as compared withk^S, 'to tie up.' Pal'al 179 forms are luls, 'to expand,' from J»», 'to make broad,' and »-¡¡J;, 'to swallow,' from fSj, 'to gobble up.' Pan'al forms 180 -are J y ^ U , 'bunch' (of grapes), from JJis, 'to twist into a knot,' and j j J j L , 'black of the eye,' as compared with ijj^L, 'eye'. Pam al 181 forms are JjuiL-^, 'very sour' (of milk), as compared with 'sour milk,' and 'glistening,' ^ as compared with 'glistening.' A^ pa'lal 182 form is i i J ^ i . , 'black of the eye,' the same as ¿sSJL, 'black of the eye,' and a pa mal 183 is seen in kXl4-i 'to shave the head,' from k i i , 178 Wright, op. cit., i. 46. 179 Of. Dietrich, op. cit., pp. 307—308, and Fraenkel, op. cit., pp. 16—25. no Lane, Lexicon, ad loc. isi Dietrich, op. cit, p. 308 and Lane, op. cit., ad loc. Dietrich, op. cit., p. 311, and Lane, op. cit., ad loc. t83 Wright, op. cit, p. 47. The infixation after the second radical may be due to the influence of p ^ - , 'to Bhear,' upon the root.

A STUDY OP PLUBILITERALS

49

'to scratch.' All these forms are very often paralleled by geminated intensives, and this fact clearly shows that to the Semitic mind these infixed forms have an intensive connotation. The liquid infix is generally considered to be a dissimilation of the gemination of the older intensive stem,i84 although there are many difficulties in the way of such an explanation. A similar intensive type is one with a guttural infix. This differs from the geminated intensive in that, while the gemination is usually believed to have been resolved by dissimilation into a liquid or nasal, the guttural infix apparently originates from the lengthening of the vowel of the geminated stem. 185 The forms actually found are p'a'al and pah'al. P'a'al forms are ^-¿.s, 'thick, strong,' from 'to force, overpower,' ^-oXisl, 'to contract until the ground,' from ^¡S, 'to become short.' Pah'al forms are 'fat,' from J^Jj, 'to shave, smooth,' and 'thick,' from J - ^ , 'onion (having many coats).' 186 Fraenkel also enumerates many cases of a guttural infix after the second radical. A ^ infixed after the second radical 1 8 7 is found in jjpi-J, 'to make retire,' from J ^ - j i 'to withdraw.' In the pluriliterals with sufformatives, we again find liquids, gutturals, sibilants, and labials, and there are also So Fraenkel, op. cit., pp. 20—21. He believes that the dissimilation of the gemination was due mainly to the influence of guttural and emphatic sounds in the root. But how are the roots without any guttural or emphatic sounds to be explained? 1S5 Fraenkel (op. cit., p. 27) believes that the p'a'al forms originate from the presence of an emphatic sound in the root: seven out of the twelve forms quoted display an emphatic sound; while only the pah'al forms may originate from an intensive stem (p. 35). 186 0p. cit., pp. 27—34; 36—38. »7 Wright, loc. cit. 4

50

ROOT-DETERMINATIVES IN SEMITIC SPEECH

occasional palatal sufformatives. The sufformatives are used with especial frequency in the formation of substantives. Thus the following derivatives occur with sufformative 1: 188 j i j i , 'young hyena,' from ¿Ji, 'to sprout forth'; 'small cup,' from 'cup,' while Wright 1 8 9 quotes J ' t o be scattered,' from g-i^, 'to be scattered.' With sufformative m,190 the following may be mentioned: ¿ J U S a , 'young' from ¿S-4-, 'stripling'; and i X ^ i 'meadow, open yard,' as compared with , 'uncovered valley.' Substantives with sufformative n occur with the utmost frequency in all the Semitic dialects, and need no , 191 'to special mention. Sufformative r is found in collect,' as compared with £4-, 'to be abundant,' and gXi, 'to collect'; 'to be proud,' from ^Cii, 'to be high'; 192 sufformative h in 'hard, strong,' from 'to sink on the knee' 'thick, rough'; sufformative 'ain in 192 (camel), from ¿ y , 'to kneel'; sibilant sufformatives in ^IXiji., 'to deceive with soft words,' from •SS^L, 193 'to deceive,' and ^JLi^L, 'to take away everything,' from i j ^ . , 'to remove everything'; labial sufformatives 193 in 'to eat like a glutton,' from 'to eat up everything,' tJj>j), 'to strangle,' from 'to swallow,' and , 'to push back' from JJL), 'to withdraw'; while an occasional 189 Fraenkel (op. c i t p . 49) considers the I of diminutive force and Brockelmann ( G r u n d r i ß , p. 402) thinks that it is borrowed from the Indo-Germanic; but the many examples in all the dialects warrant no such general deductions.

>sa loc. cit. 180 Fraenkel (op. cit., p. 46) notes that practically all derivatives in m are of substantive character. i9i Wright, loc. cit. 132 Dietrich, op. cit., p. 318. 193 Wright and Dietrich, loc. cit.

A STUDY OP PLUBILITEBALS

51

palatal is seen in 'to roll down, slip,' which is also derived from J Á j , 'to move back, withdraw.' F . Modern Arabic. Turning to the modern Arabic dialects, that of Egypt has the very curious form k^ssJ, 'to mix, confuse,' derived from ^

9

'to knock,' with

a

^

preformative I, for, although k^svj is usually considered to be a metathesis for *kIJÁ.,i94 a preformative I is quite as possible as a preformative n or»». 1 9 5 In fact, preformative I occurs frequently in proto-Semitic, as is clear from a study of the root-determinative in the triliteral verb. 1 9 6 áaf al derivatives in Egyptian Arabic are 'to overturn,' from 'to upset'; and ¿J-i-^o, 'to suspend,' from ¿ í » , 'to hang up'; and there are also peculiar maf'als derived from nouns in m, such as oj 'to nail,' from jl^JUt, 'nail'; and jkJLú, 'to rule,' from a^kJL^, 'ruler.' Forms with infixed 'ajn include ¿six*s, 'to make the flesh creep,' from 'skin,' and 'to climb,' from k^io, 'to take hold of.' Par'al, pan'al, pal'al, and pam'al forms 19 abound, these including or ? 'to run, burst in,', from («-^v*, 'to rush at, attack'; 'to push down,' from 'to press'; 'to be wide and flat,' from ^k*», 'to make flat'; and 'to tear to pieces,' from ly8 1 9 9 L j JCherbonneau j , 'to tear' (a pa'mal form). has given a complete list of quadriliterals is* Spitta-Bey, Gram., p. 191. 195 Por an apparent case of preformative r in Hebrew, cf. BteBn above, note 146. 196 See examples p. 97. is' But Brockelmann (op. cit., p. 510) thinks that this is a loanword from the Persian hangama, 'concourse.' 198 For complete lists, see Spitta-Bey; loc. cit, and Toilers, Modern Egyptian Arabic, pp. 33—34. '99 J A., V. xviii. 375—384. 4*

52

ROOT-DETERMINATIVES IN SEMITIC SPEECH

for Algerine Arabic. Besides pan'el par'el, and pal'el forms, of which there are numerous examples, such as i ^ J ^ L , •to sprout forth in foliage' (cf. 'wood'), 'to work with energy,' J->y, 'to be in tatters,' 'to prop up,' etc., there are many others which have an infixed guttural or semi-vowel y or ^ such as J-05», 'to be curious,' 'to send an envoy,' 'to turn things topsyturvy,' etc. In the dialect of Oman, Yollers 200 considers the pluriliterals tdelhem

(pal'el) and G. Ge'ez.

and selheb to be derived from

(saf'el) respectively. The pluriliterals in Ge'ez have been discussed

at length by Stade in his monograph Ueber den der mehrlautigen

Thatworter

der Qe'ezsprache

Ursprung

(Leipzig 1871);

and a few examples will suffice to show that the formations in this dialect were exactly similar to those in the other Semitic languages. Thus, the gaf'al formations include (\73 p. 43 and

y

p. 241.

y l


% 2«o Cf. Lagarde, Abhandlungen K. Gesellschaft der Wissenschaften zu Göttingen, xxxv. 10. 26i The Indo-Germanic root was usually dissyllabic. The determinatives were added to these dissyllabic bases—e. g. *ore, 'to move' (Latin orior), *ore-d, 'to disurb' (Sanskrit ard), *ore-s, 'to flow' (Sanskrit ars). Of course there were also monosyllabic, and plurisyllabic bases (Cf. Hirt, Der indogermanische Ablaut).—Dr. Gray.

CHAPTER III EXAMPLES OF PRIMITIVE SEMITIC STEM-FORMATIONS

In seeking to analyse the Semitic triliterals according to their formatives, the fact must first be recognized that a large number of these roots have probably come into existence at a relatively late period, and that they may have been formed, from their very inception, upon the analogy of triliterals—a fact to which Lagarde has already called attention. 262 Notwithstanding this necessary reservation, the presentation of a comparatively small number of selected examples from the classic languages of the Semitic group, will amply prove the correctness of the hypothesis here maintained—that the Semitic triliteral is developed from a biliteral stem by means of a root-determinative; and that the root-determinatives are identical with the well-known stem-formatives. The chief obstacle which confronts the investigator oí the Semitic root-determinatives is the frequent difficulty (if not, indeed, the impossibility) of distinguishing between pre-, in-, and post-determinatives in certain obscure formations; and here the only hope for a correct solution of the problem lies in keen judgement and in close comparison of kindred 26S Lagarde, Symmicta, p. 122: "There are in Semitism triliteral roots which do not yield to any attempt to reduce them to biliterality . . . We may expect that it also formed new roota which from their very cradle were triliteral."

EXAMPLES OF PEIM. SEMITIC STEM-FORMATIONS

75

forms. Thus, in studying the etymology of the H e b r e w pR2, 'to slash, cut to pieces' (Ezekiel xvi. 40), t h e first impulse would be to connect it with *iiro, 'to cut,' A r a b i c ¿ - o , 'to sever,' and to regard the q as a post-determinative; but since q is a very r a r e determinative, it would b e more plausible to connect it with 'to cleave' (seen in y^a, 'to cleave,' "Ij52, *'to divide,' etc.), and to consider its t as an infixed reflexive element. The reason for such infixation would lie in the fact that the root came under the influence of *nn3, to which it sought to assimilate itself. Every case must, therefore, be judged individually; no general rules can be laid down. I t may, however, be regarded as axiomatic t h a t a primitive biliteral will invariably be found in the intensive )!"]) verb or underlying the hollow verb; and these biliterals can be used as a basis for further investigation.

i Stative Stems. These stems, which fall into the two categories of a) V'B or V'S, and b) V'b or have, for the most part, become identical in meaning with verbs in 1"J?, V 'V, the latter having assumed a stative connotation because of the common determinative V"1 which the two classes possess. They generally denote a) an intransitive or stative force of the root; or /3) an abstract condition. [The attempt is made in the following examples to restore in some of the less evident cases the hypothetical original meaning, these meanings being indicated by a prefixed asterisk.] 2 6 3 263 All verba which are not found in the first (or qal) form are also regarded as hypothetic, and their roots are invariably indicated by a prefixed asterisk.

76

ROOT-DETERMINATIVES IN SEMITIC SPEECH

A. HEBREW a) Y'B or i"B 2 4

l l j , '*to quiver, ' be in a state of fear.' Ygr, '*to harrow, drag'; cf. 11}, 'to drag' (intensive), 264 whence ITIIB, 'saw,' " n u r n , 'making itself drag swiftly along' (Jer. xxx. 23); "113, primarily, '*to drag along, tarry, sojourn temporarily'; 2 6 6 secondarily, 'to harass,' 'to be in a state of fear'; T]J, 'to harrow, scrape.' DIJJ, 'to be hot.' yhm, '*to burn, make hot'; cf. Don, 'to become warm,' n§n, 'heat, sun.' *f)IT, '*to be uncovered' (privative), whence 'barefoot.' y l i f , '*to cover'; ^sn, 'to enclose, cover,' ItBty 'canopy'; 'to cover.' '*to be comprehensive, able.' '*to encompass, surround'; cf. A r a m a i c 'crown,' and 'to surround'; Hebrew ilV?, '*the wreathed one, bride'; '*to surround, shut up, restrain'; '*to com267 prehend, include, complete'; n"73, 'to be complete' (second stative stem), ijp;, '*to sweep in, add.' / a / , '*to sweep'; 2 6 « cf. '*to be swept off, come to an end, cease'; HBiD, 'storm-wind'; More literally 'to make the flesh creep' (cf. Latin horrere, 'to bristle, shudder'). Similar metonomous expressions of fear are 13bJ (Joshua iv. 24), 'they melted from fear,' and DDJ a!> (Nahum ii. 11), 'faint-hearted.' 2«5 For analogical developments to this root, cf. Latin trahere, 'to

draw, drag,' in Walde, Lateinisches etymologisches Worterbuch, p. 788. 269 The Biblical verb "in always means 'to tarry' in contrast to atfj which means 'to dwell.' See the comment on Deuteronomy xxyi. 5 in the Sifri. 267 The meaning 'to include' is usual in Neo-Hebrew. Haupt (AJSL., xxii. 205) thinks 'to hold' is more original. 268

Cf. Martin, Semitic Language», ii. 16.

EXAMPLES OF PRIM. SEMITIC STEM-FORMATIONS

77

nSD, 'to sweep away'; HDD, '*to sweep towards, join, attach to'; '*to sweep together, gather' (causative). 'to designate, appoint.' Y'ad, '*to be ready'; cf. Ity, 'to testify,' denom. from "% 269 '*one who is ready, *on hand, witness'; *WJ|, 'to prepare oneself, be ready' (reflexive).270 '*to be high,' whence hif'il 'to profit.' / ' a l , ""to elevate'; cf. 'to ascend.' 'to lay snares.' Yq§, '*to strike at, *aim at'; cf. B^JJ, 'to lay bait'; 'to strike, aim'; Neo-Hebrew tfptfp, 'to strike, knock.' '*to wave,' whence hif'il B^lil, 'to extend.' '*to wave to and fro'; cf. B1#, 'whip'; Bltf, primarily, '*to be wafted about, rove about'; secondarily, '*to wave aside, despise' (figurative, stative); BKtf, 'contempt.' '*to be attentive,' whence hif'il JTSftn, 'to deliver.' / ^ a ,

flow';

'overflow' (causative), '*to

shine, b e

pure';

cf.

H e b r e w ijst, 'to b e bright,' H3J, 'to b e pure.' ¡Tin,

haru, 'to

'to

dig,

naldru, HD1, rabii,

bore';

yhr, pn,

be

'to

dig';2s4

harasu,1285

cf. T i n ,

'to

great'.

yrb,

'*to

cut

hararu,

into,

dig';

grow, i n c r e a s e ' ;

'to be g r e a t ' ; 3 T , erebu, E.

ARABIC.

a)

V'fi or >"t>.

' t o face, encounter.' ili,

'*to

'nostril'; H e b r e w T i t , 'hole.' 'to

rababu,

kL}

dig.'

Ygh,

cf.

'to increase.'

'*to face';

cf.

'face';

' t o encounter.' mark, brand.'

yhn^i), '*to

mark';

cf.

'to

b e marked, eminent, high, lofty.' 86 2

'to

hasten.'

Yfz,

'*to

hurry, b e s t i r oneself';

' t o b e excited, aroused';

p,

cf.

' l i g h t , active m a n ' ;

ji, ¿ji,

' t o be excited, frightened.' 'to turn out right', jjli,

' t o b e superior';

yfq,

' * t o come

¿¿s,

'above';

forth

high';

cf.

, 3 i j , 'to be made

high, h o n o r e d . ' 2 8 7 283 The roots in and U (cf. Hebrew nbtf, 'to be quiet') are originally differentiated forms. 284 The roots hr, '*to dig,' and hi, '*to dig out,' are, likewise, primitive, differentiated forms. 285 The s was developed through the influence of hs(2), '* to cleave, divide' (cf. hussu, 'destruction, hasabu, 'to cut off'). 286 Both words may be denominatives from Old Semitic 'name.' 287 Cf. Dozy, Supplement,

ii. 704.

EXAMPLES OP PRIM. SEMITIC STEM-FORMATIONS 'to enter into.'

83

'*to keep on going, engage in';

ylg,

cf. g-J, 'to persevere, persist' (intensive); lasJ, 'to have recourse to, to betake oneself to.' 'to become weak, infirm.'

yhn, '*to wane, be defi-

cient; cf. ¿Ia, 'to be low, abject, vile'; '¿¿¿a, 'small, little thing.' •.J^ki, 'to be good.'

ytb,

'*to be good'; cf. CjL,

'to be

sweet.'

b) 1, !? or

bore';

'*to dig'; 2 9 2

tender';

Arabic

cf. A r a b i c

yo,

'to hollow out' (p. 13).

' * t o use words in heat,

be hot';

thin,

'to empty, spit'.

cf. r f i T t : ,

'gall';

curse.' ih '*to sigh, be afflicted'; cf. the preceding discussion. '*to turn in on one, befall (good or bad fortune).' \idr, '*to turn'; cf. fô, 'to turn, revolve'; 'to turn back, repel'; 'to turn on one suddenly.' 'to rub, press.' ydk, '*to beat, crush'; cf. 'to break, crush'; 'to bruise, pound.' iJuLà, 'to smear, anoint, overspread.' ySf, '*to smear'; cf. 'to polish'; 'to be translucent, shining.' 302 The meaning 'to measure' is a development from the idea of extension (So Muss-Arnolt, Dictionary, p. 514). 303 The meaning has passed through a similar line of development and accordingly, = j li".

EXAMPLES OF PRIM. SEMITIC STEM-FORMATIONS

ftKli,

F. E T H I O P I C . 'to be weak, impotent.'

'*to quiet,

89

lull';

cf. Hebrew JJKltf, «at ease'; Jtf;, 'to be asleep.'so4 rtOrt: or rtrhrt'to lick.' Y IS, '*to lick'; cf. A r a b i c J j , 'to lick'; ¿ l ^ J , 'tongue'; Hebrew )1b6, 'tongue'; M t h '*to lick off, become insipid, inefficacious'. MN1:, '*to try to encompass, be powerful, able.' 3 0 5 '*to encompass, surround'; cf. h r t r t ' t o wreathe, surround'; flhrt:, 'to effect, accomplish' (causative). (10.£¡, '*to cut off suddenly, change.' ybd, '*to separate, divide'; cf. fl.C'fl.fi:, '*to be suddenly changed, perish', (intensive); ( U t f : , 'to become like dead' (denominative); (1 ñ a ) ' * t o be divided from, alone.'

iv Causative and Kindred Stems. The pure causatives are formed in Semitic by prefixing ('), or h to the biliteral root, and their original force is still evident. Related to the causatives are the stems with pre-determinative (') and ifi, some of whom have a causative meaning. A. H E B R E W . natf, '*to put a stop to, rest.' ybt, '*to cut off, sever' 3 0 6 ; cf. *nri3, 'to cut off, sever'; "ins, 'to cut in two'; A r a b i c c - j , 'to sever'; C-JLCÓ, 'to interrupt.' 'to twist.' Yzr, '*to t u r n ' ; cf. lit, 'to wring, compress' (Judges vi. 38); *n, 'circlet'; TUS, 'girdle.' íintf, '*to bring to the ground, ruin.' ylit, '*to undermine, '04 So Dillmann, Lexicon, p. 376. 308 So Haupt (AJSL., xxii. 205). 30« The idea of rest originates in Semitic generally from the idea of interruption (cf. Bj?^ and 'jbj below).

90

KOOT-DETEKMINATIVES IN SEMITIC SPEECH level'; cf. flTin, '*to b e s h a t t e r e d , d i s m a y e d ' ;

Aramaic

fini, '*to c o m e d o w n t o t h e g r o u n d , d e s c e n d ' (reflexive). {Dptf, '*to p u t a stop to, cease,' hence, 'to b e a t rest.' '*to cut o f f ' ; cf. *at3j?, 'to c u t

Yqt,

off, d e s t r o y ' ; i?ttjj, 'to

slay'; )bj?, '*to b e c u t off, s h o r t ' ; tttp, '*to b e c u t ' (figuratively). hence, ' t o l o a t h e , feel d i s g u s t f o r . ' 3 0 7 '*to b e c u t off, perish.'

ybd,

'*to s e p a r a t e , d i v i d e ' ; cf.

"¡13, ' t o b e s e p a r a t e ' ; '"Via, ' t o b e divided.' "iB'lrt, '*to go in a s t r a i g h t line, a d v a n c e . '

'*to go

s t r a i g h t ' ; cf. "titf, 'to j o u r n e y , t r a v e l ' ; A r a b i c m a r c h , j o u r n e y ' ; N e o - H e b r e w r n ^ , ' a row,

'to file';

TBfy

'to b e s t r a i g h t , u p r i g h t . ' 'to t h r u s t , push, b e a t off.'

y d f , '*to b e a t , k n o c k ' ; cf.

* n s i , ' * t o b e wounded,' w h e n c e ""B^, 'blemish' ( P s . i. 20); p a n , 'to b e a t , knock';308 ^ IJDJJ, 'to t u r n , o v e r t u r n . '

«to p u s h , drive.'

Ypk, '*to p o u r ' ; cf. 1JB$, 'to spill';

'*to spill,' w h e n c e iJB, 'flask'; N e o - H e b r e w 5JB2B, 'to t r i c k l e . ' B i n , 'to b e a t out.'

ybt,

'*to b e a t ,

strike';309

cf. to?a, 'to

kick'; ^123, '*to b e b e a t e n off,' hence, 'to desist, c e a s e ' ; *tsai, p e r h a p s 'to d a r t (a glance),' hence, 'to look.' ^SJJ. '*to roll a r o u n d . ' Ygl,

'*to r o l l a b o u t ' ; cf.

n^J, 'to roll off, u n c o v e r ' ;

'to roll';

'*to b o u n d with joy, r e -

joice.' 3 io B. A R A M A I C , pats', ' * t o t e a r away, l e t g o free.'

ybq,

' * t o cleave'; cf.

Similarly pip, 'to loathe,' is developed from qs, '*to cut,' the causative form of which, also means 'to abhor.' 308 The q was developed through the influence of dq, '*to pound.' 809 Cf. Lagarde, Reliquiae, p. xlvii, and Martin, op. cit, ii. 12. »1« Cf. note 275.

EXAMPLES OF PRIM. SEMITIC STEM-FORMATIONS

91

'to cleave'; "1(53, '""to cut into, make an investigation'; 311 Hebrew pj?3, p}3, '*to be cleft, empty.' "11$, '*to make to trickle, shed.' Ygr, '*to harrow, drag'; cf. "113, 'to draw, drag'; Arabic 'to run along, flow'; Hebrew 1JK, '*to drag together, gather.' 'to drive, pursue.' y d f , '*to beat, knock'; cf. 'to beat'; *)"], '*something planed, plank'; ^"ll, '*to thrust oneself forth, spread' (reflexive). DOti*, '*to burn to brownness,' hence, 'to become brown.' yhm, '*to burn'; cf. DOrj, 'to be hot'; Hebrew Din, 'brown.' 'to arm oneself.' //?»; denominative from f t , 'weapon.' 'to cast forth, pour.' yM ( 1 ) '*to throw'; cf. "Htf, 'to throw'; "nc*, 'to send' (causative force). i)in, 'to drive, push' y d f , '*to beat'; cf. above. 11J1, 'to walk around, turn around.' ]/iZr, '*to turn'; cf. Arabic j\>, 'to move in a circle'; Hebrew i n , 'circle.' ^31}, 'to become corrupt, bring to ruin.' Y^hn^ '*to confuse, confound'; cf. ^3, 'to be in ruin, decay.' pa'el 'to consume.' Y^h^)3l2> '*to finish, consume'; cf. 'to eat up, consume'; Hebrew 'to finish, consume.' C. SYBIAC. 'to act boyish' l/&r(8); denominative from 'son.' tiU, '*to pierce one's eyes by a money offer, bribe.' j/7wZ, '*to pierce, sharpen'; cf. Aramaic Tin, 'to be sharp'; Hebrew nD^, 'bribe.' 313 311

Similarly 1J?n, 'to explore,' from hq, 'to cut' (p. 78); \br may have also exerted an influence in specializing the r. 312 ykl{2), ' • t o consume,' is a development from kl(i)> '*to embrace, complete' (See Haupt, AJSL., xxii. 205). a« Cf. note 115.

92

R O O T - D E T E R M I N A T I V E S IN SEMITIC S P E E C H

iMA, '*to burn to blackness,' hence, 'to be black.' yhr, '*to '*to become burn'; cf. Aramaic *Y]n, '*to burn'; bleached, white.' u. 'to spread, extend.' yth, '*to spread over'; cf. 'af'el, 'to cover'; Hebrew nils, 'to over-spread, besmear, coat'; 'hemorrhoids'; Hebrew Drills, 'tumours.' ^/l, '*to flow,' hence, 'to go away.' yd, '*to flow'; cf. v j , 'to go up'; V , '*to flow away, be spent, light'; Hebrew 'to be light, worthless'; /It, 'to lavish.' ij|l, '*to compress, collect.' ysr, '*to bind'; 3 1 4 cf. 'to press together, bind'; '*to become constrained, narrow, low, mean'; Hebrew 1JJ?> 'depression, grief.' root of iiAoi, vapor, mist, nothingness.' ybl(i->, '*to confuse, confound'; cf. Hebrew 'vanity'; Va: 'confuse, mix up.' 'to walk around, encompass.' ydr, "*to turn'; cf. /jot', 'circle'; Aramaic Yin, 'to turn around.' D. A S S Y R I A N . tsa^, Sdbatu, 'to beat, strike.' ybt, '*to beat, strike'; cf. •Hebrew 'to beat out'; alt?', 'rod' 3 1 5 t23N4, root of ubbutu,

ebitu, 'oppression, want'; battu, 'weapon.'

pty, Sakanu, '*to put down on the ground, place, set.' ykn (i) '*to bring to the ground,' cf. kanu, '*to be on a level, firm'; )33, kananu, 'to crouch.' ipD, sakapu, '*to topple over, overthrow, upset.' ykp, '*to bend'; cf. ^BD, kapapu, 'to bend, oppress'; HB3, kipu, 'to bow down.' Sapasu, '*to spread out, grow mighty.'

Yps^

3'4 Cf. Martin, op. cit., ii. 16—17. 3>s Cf. Lagarde, op. cit., p. xlvii., Martin, op. cit., ii. 12.

'*to

EXAMPLES OF PRIM. SEMITIC STEM-FORMATIONS

93

spread'; 3 1 6 cf. Hebrew 'to be dispersed'; 'to spread itself, be scattered' (reflexive) (p. 22). rOKi, ajtatu, 'to destroy.' 3 1 7 yht, '*to cut off, sever'; cf. *nro, batatu, 'to bring to an end.' or ciidabu, azdajni, 'to strike down, overthrow.'

Vdp, '*to beat, knock'; cf. duppu, '*something beaten out, a tablet,' hence 'document'; Aramaic KB1, 'board, page'; ^ n , 'to drive, push.' root of edaqqu, 'small.' ydq, '*to grind, make thin'; cf. ppl, daqaqu, 'to crush, make small.' e^zebu, '*to make to go one's way, forsake, let.' Yzb, '*to go, flow';318 cf. zdbu, 'to flow, melt'; Aramaic 'to free, deliver'; 3 1 9 Hebrew 'to forsake, let go free.' E. ARABIC.

Complete lists of Arabic causatives of the (K'S), iTB, and types are given by Mez 3 2 0 , although he has made the error, as already noted (p. 68) of considering all these triliterals as secondary denominatives from stems which are alleged to be still in existence. I t would, however, be difficult to show a haf'al or saf'al in Arabic, though both clearly existed in proto-Semitic. 'to fill (water-pipes).'

Ygr, '*to harrow, drag'; cf. ^ J ^ - ,

'to run, trickle, flow'; yL, 'to drag, pull.' 3 1 6 l^P? ( 2 ) ' * l 0 spread,' is a secondary development from "|/jM(i) '*to break, smash' (cf. Assyrian napasu, 'to shatter'). 3 17 abatu and Hebrew 13N, 'to destroy,' are not directly connected as is generally supposed (cf. Muss-Arnolt, Dictionary, p. 11), but \bt and \bd are early differentiated biliterals. 318 For a similar line of development, compare Indo-Germanic leilca, 'to flow' and leiki 'to leave' (Latin liquor, linquo, Greek Xeis-M), (Persson, op. cit., p. 5). 319 «A Babylonism" (cf. Delitzsch, Hebrew Language, p. 69, note). 320 Orientalische Studien, Theodor Noldeke . . . gewidmet, i. 249—254.

94

ROOT-DETERMINATIVES IN SEMITIC SPEECH

J S ' t o let the hair hang' \fdl, '*to lift'; cf. 'to lift up, lower (bucket).' ^Lw, 'to spread out, diffuse.' yth, '*to spread over'; cf. 'to spread, expand'; Isai, 'to spread out.' "Sjh, 'to incise, write', y t r , '*to cut into'; cf. tL, 'to cut.' 'to mix.' ysb, '*to mix' 3 2 1 ; cf. 'to mix, confound'; 'to mix up, to be mistaken for somebody else.' 'to teach, train.' ydb(i), **to toil, labor' 3 2 2 ; cf. i^jb, 'to toil, labor, practise'; ¿ j b , 'custom, habit.' J*.*, 'to forsake, leave off.' 323 \/gr. '*to harrow, drag'; cf. 'to flow, run,' yL, 'to drag, pull.' 'to become aware of something, feel.' VgS m , '*to feel'; cf. 0 ' * t o feel for something, seek out'; 'to feel with the hand.' 'to become sour.' ymz, '*to grow sour'; cf. 'to become sour'; perhaps 'banana-tree.' j l s f , 'to grasp.' V g s ^ '*to feel'; cf. 'to feel,' above. F. E T H I O P I C . ViTVp;, 'to divide, cleave.' 3 2 4 \/tq, '*to cleave, smash to pieces'; c£ *m : , 'to fall asleep.' ydm, '*to lull, quiet'; cf. '*to make quiet, extinguish, obscure'; Hebrew dttl, 'to be silent'; n a n , 'silence'; Q11, 'to slumber.' v Reflexive and Kindred Stems. The direct reflexive in Semitic is formed by prefixing or infixing t to the biliteral base, while the indirect reflexive either prefixes or infixes n to its base. The indirect reflexive generally denotes an action performed for oneself or on oneself in an indirect way, and very frequently becomes mere passive, and is accordingly confused with v"lB, or in a large number of instances. The kindred stems of this conjugation have I, m, and r as a pre- or in-determinative. A. H E B R E W . 2K1J, '*to desire for oneself, long for.' \/'ab, '*to desire'; cf. ¡"DM, 2»;, '*to be in a state of longing, desire.' For a similar development of meaning, cf. Hebrew i>tJ, 'to plunder,' from ygz, '*to cut.1 '26 Cf. Haupt, AJSL. xxii. 205.

96

ROOT-DETERMINATIVES IN SEMITIC SPEECH

*Diiri, ' * t o b e s t i r ' t h e deep.'

oneself, * b e

' * t o r o a r ' ; cf. DOiJ, Din, n » n , 'to r o a r ' ;

yhm,

A s s y r i a n tidmtu, J?J?n, 'to t h r u s t

in c o m m o t i o n , ' w h e n c e Dirw,

'ocean, sea.'

at, clap.'327

yqj5JJ[2, 'imprint, incision.' 1J3JJ, ' * t o

encompass,

overpower,

r o u n d ' ; 3 2 8 cf. '*to '*to

encompass

oneself,'

encompass, surround';

Hebrew *"iro,

'*to

prevail.'

yqf,

'*to

sur-

'to s u r r o u n d ' ; HBipFl, 'circuit.' whence

^fii,

'wall.'

cf. A r a m a i c

«Wra, ' c r o w n ;

' * t o s u r r o u n d , shut up, imprison.'

surround

oneself,' w h e n c e

'*to turn round';

329

cf.

'crown.'

}fkrn)

'to t u r n r o u n d a n d r o u n d ' ;

E t h i o p i c h C h l : , 'to t u r n ( a mill).' 'minxit.'

2)

'*mingere';

cf. D I T i ^

( I I K i n g s xviii.

2 7 ) , ' u r i n a ' ; E t h i o p i c " ¿ 1 : , A s s y r i a n Sanu, yril, ' * t o

p r e s s oneself, b e u r g e n t . '

yhsw,

'to squeeze, p r e s s ' ; A r a b i c

'mingere.'

' * t o p r e s s ' ; cf.

'to p r e s s t o g e t h e r ,

sew.' '*to

t h r o w o n e s e l f over, lift.'

ytl,

'*to cast, throw';

cf. bits, 'to hurl, c a s t ' ; nVob®, 'hurling.' 'to drip.'

ytp,

' * t o drop, t h r o w ' ; cf. A r a m a i c fjito, 'to

drip'; N e o - H e b r e w n ^ t t , 'drop'; **IJ>3, ' * t o b e s t i r 'youth.'

'to overflow.'

oneself, b e full o f m o t i o n , vigor,' w h e n c e ]/"ar ( 2 ) ' * t o s t i r up'; cf. n y , 'to b e a r o u s e d ,

a w a k e ' ; "IJJD, 'to storm, r a g e . ' 327 Later usage has forgotten the original force of this root and the vernaculars have developed forms from a hypothetic biliteral \tq (cf. joiuol, 'to push,' in Neo-Syriac, p. 18). But such a biliteral is not to be found in ±he early language. 328 The roots qf and lcf are early differentiated forms. 329 The roots fcr(i) and kl(i) are early differentiations.

EXAMPLES OF PRIM. SEMITIC STEM-FORMATIONS *f)13, '*to

bend

oneself,

"heave';

97

'wing.' 3 3 0

whence

y k f , '*to b e n d ' ; cf. ^ M , HS3, 'to bend, compel.' Dnb, '*to h e a t o n e s e l f , ' 3 3 1 whence DH^n, '*to be h e a t e d in battle

(reciprocative),

fight,'

b r e a d ' ( A r a b i c ^¿vJ, 'meat'),

and

Dn1?, '* nourishment,

y h m , '*to b u r n , b e h o t ' ;

cf. a o n , Din, 'to b e hot, burn.' 1

l A s y ) * 3 3 2 '*to p r e s s ' ; cf. f r u , 'to

fn ?, 'to squeeze, press.' be urgent'; Arabic

'to p r e s s together, sew'; S y r i a c

'to bind, compress.' pbtf, 'to urge.'

j/'as, '*to press'; 3 3 3 cf.

'to b e pressed,

m a k e haste.' p^S, 'to waste, lay waste.' 3 3 4

ybq,

'*to cleave'; cf. *pi2, '*to

b e cleft, empty'; pj?3, 'to cleave, empty' ( I s a i a h xxiv. 1). •ijbD, '*to lead,' root of go'; cf. ^ n , Arabic

' * l e a d e r , 3 3 5 king.'

'to go'; E t h i o p i c

rtSh'to

\flk,

'*to 336 Of. note 290. 337 The root has infixed the r through the influence exerted by \'pr, '*to cleave.' 338 Assyrian takalu, Arabic (cf. p. 25), are generally held to be secondary denominatives; but inasmuch as the form is common property both in North and South Semitic, it is more likely an original reflexive stem. 339 So also Lagarde, loc. cit. 3io Lagarde, loc. cit. 34i The root nh, 'to breath, rest' (Hebrew raj, Aramaic nr, 'to rest,' Arabic

'to breathe'), has influenced the development of

E X A M P L E S OF P R I M . S E M I T I C S T E M - F O R M A T I O N S turn,

be

hot';

cf. D o n ,

'to

be warm';

'*to

99 be

burnt, brown.' 'to b o r e out, hollow.'

y q b , ' * t o b o r e ' ; cf. X n ^ p ,

'room';

K n a p , ' i n s i d e w a l l s of t h e s t o m a c h ' ; H e b r e w Dp, ' p a i l , measure.' D p i , ' " t o a r i s e f o r oneself, r e v e n g e . '

]fqm,

'*to s t a n d ,

arise';

cf. Dip, D^p, ' t o s t a n d , a r i s e . ' JfiS,

'to

twitter,

chirp.'

ysp{2),

'*to

chirp';

cf.

fpi'J,

'to

t w i t t e r , c h i r p ' ; H e b r e w ^SSS, ' t o c h i r p . ' Virh, ' t o s p e a k i n a w h i s p e r . '

y h S ^ ) , ' * t o w h i s p e r ' ; cf. " ^ n ,

' w h i s p e r i n g s ' ; H e b r e w t!?nj, ' * h i s s i n g o n e , s e r p e n t ' ; B^nj, oracle).343

' t o f o r e t e l l ' (of a n

tsVa, ' * t o s l i p a w a y , e s c a p e . ' '

]fmt, '*to

slip';

cf. tJifc,

'to

slip.'

«BO, ' t o r e a c h . '

y f a , ' * t o e x t e n d , t u r n ' ; cf.

J?"]», ' * t o b e b r o k e n u p , sick.'

'to reach.'

y>'a(1), '*to shatter,

break';

cf. y j n , ' t o b r e a k . ' 10tf, 'to

destroy.'

343

ysd,

throw'; Hebrew

'*to

throw

over';

cf.

'to

' t o d e v a s t a t e ' ; 11"^, ' v i o l e n c e , r u i n . '

nn*l, ' * t o f e e l h e a t , b e i m p a s s i o n e d , love.'

yhm,

'*to burn,

b e h o t ' ; cf. Dion, ' t o b e h o t . ' tihn, 'to b e c o m e dumb.'

yhS(3),

' * t o w h i s p e r ' ; cf. ^ n b , ' t o

w h i s p e r ' ; "B'n, ' w h i s p e r i n g s ' ; H e b r e w

'serpent.'

the connotations of this root. Hebrew nrun, accordingly, means 'to feel sorry' (Cf. Haupt, BA., vi. 2, note 22). The kindred form tarn, 'to love' (Hebrew, 'to comfort'), proves conclusively that etymologically Dni is to be joined to yhmThe two roots are also connected by Baudissin, Studien zur semitischen Religionsgeschichte, Leipzig 1876, i. 281,287. See, however, Lagarde, K. Gesellschaft der Wissenschaften zu Gottingen, xxxv. 50. 3« The infixed m is due to the influence of 2) (Hebrew D»ti>, 'to lay waste'). 7*

100

ROOT-DETERMINATIVES IN SEMITIC SPEECH

C. SYRIAC. 'to rout, repel.'

\fhs,

' * t o c u t ' ; cf. « w ,

'to cut off';

•rriTi, ' t o slay.' ouVl, l * t o c u t o n e s e l f off, s p l i t . ' y l h , ' * t o c u t ' ; cf. U X , c u t off, b l o t o u t ' ; ' t o c l e a v e , divide.'

'*to

'board, tablet.' Y r l a m , ' * t o s h a t t e r , b r e a k ' ; cf. y j n ,

'to break.' iKu, '*to k i n d l e oneself, b e aroused.'

|/7;r, ' * t o b u r n ' ;

cf.

A r a m a i c * 1 i n , ' * t o b u r n ' ; -J^M, ' * t o b u r n t o b l a c k n e s s . ' o)^», ' t o f o r c e a p a r t , b r e a k , s h a t t e r . ' ' c o m e f o r t h ' ; 3 4 4 cf.

Vpq,

'*to force apart,'

'*to force oneself out, go away,'

AS, ' t o f o r c e o u t t h e b r e a t h , sneeze.' iXa, ' * t o m a k e a s o u n d t h r o u g h t h e n o s t r i l s , n e i g h , ' n a t i v e f r o m I'^UJ, ' n o s e . ' 'hollow';

Assyrian

mn,

haru,

'to

'nostril'; A r a m a i c K T W , 'nostril.' K&i, ' t o go down.'

\'rht,

velled, below'; 'to blow.'

denomi-

] f h r , ' * t o d i g ' ; cf. H e b r e w i n , 345

dig';

Hebrew

THS,

.

'*to u n d e r m i n e , l e v e l ' ; cf. K-ul, ^ l e 'to u n d e r m i n e ,

destroy.'

V p h , ' * t o b l o w ' ; cf. o i s , ' t o b l o w u p . '

f a j , '*to s h a k e oneself up, quiver.'

'*to

shake,346

t r e m b l e ' ; cf. i » , ' t o q u i v e r w i t h e m o t i o n ' ; f s ^ s , ' t o s h a k e , tremble'; n i l l s n n

11b, ' i t is s p l i t u p ' ( I s a i a h xxiv. 19).

uuij, ' * t o p u s h o n e s e l f f o r t h , a r i s e , s h i n e ' (of t h e

sun)

347

.

344 The Arabic always uses this root in the special sense of 'to come forth high' (cf. J j i j , p. 82). The iafixation of t in .ok.* may be due largely to the influence of \pt (Hebrew nns, 'to crumble'). 345 Hebrew in:, Arabic 'to snort' are denominatives from the word for 'nostril.' 346 ypr(2) '*to shake' is developed from }/pr(i) '*to cleave, break.' 347 This is generally held to be a dissimilated form of Hebrew nil, 'to shine,' but proves its connection with ydh. The root nit is from a biliteral \zh, '*to shine,' the differentiated forms of which are \zk (cf. Hebrew iji, 'pure') and \zh (cf. inj, 'to shine,' a.lt, 'gold').

EXAMPLES OF PRIM. SEMITIC STEM-FORMATIONS 101

ydh, '*to push'; cf. U», 'to push'; 'to break forth, arise (of the sun).' '*to bestir oneself,' whence pa'el 'to arouse, stir up.' ybt, '*to beat, poke'; cf. -L^j^», 'spark'; Aramaic BBS, 'to flame up'; Hebrew B3n, 'to poke'; Aramaic B^a, '*to push forth, project.' 348 uA-j, 'to arouse, stir up.' ydh, '*to push'; cf. JJ.;, 'to push'; oil), 'to break forth, shine.' ifs», '*to change, become corrupt.' \'dr, yzr, '*to turn, twist, change'; cf. Hebrew *K7], 'to become corrupt'; "lit, 'to be loathsome'; Neo-Hebrew It», 'to twist, be loathsome, spoiled' (secondary denominative) (p. 23). 349 'to swell up.' yph, '*to blow'; cf. 'to blow, diffuse.' «¡¡a, '*to break forth (of light), lighten, flash.' ybq, '*to cleave'; cf. JLA», '*to break into, explore'; p p , '*to break into, explore'; Aramaic jjjja, 'to cleave.' D. ASSYRIAN. iN3D, ti'ou, 'to make an invasion, rise against.' yb'a, '*to enter'; cf. K13, ba'u, 'to come.' "^30, tabaku, 'to pour, spill.' ybk, ypk, '*to pour, shed'; cf. 033, baku, 'to shed tears'; Sapaku, 'to pour out.' prD, bataqu, 'to cut through.' 350 ybq, '*to cleave'; cf. Dp3,

baqamu, 'to cut off.' 348

The word cannot be, as Levy ( N H b C W B ) maintains, from the Greek f3\aBJ, natalu,

sahu,

ydn,

'*to use power, judge.'

*bnt,

'to b e unlucky, inauspicious.' 'to conceal.' 'to b e

firm.'

yicm, '*to cover'; cf. 3ss

]/lb-,

cf. SJ,

'to cover.'

'to have h e a r t , u n d e r -

s t a n d i n g , courage.' js?,

'*to b e cut off, perish.'

/ g z , '*to c u t ' ; cf. J*., 'to c u t ' ;

tUi, '*to b e cut off from, leave behind.' J J J , 'to become low, base.' \fdl,

'*to b e low, m e a n ' ; cf. J > ,

'to b e low, b a s e ' ; J \ > , 'to b e low, on t h e ground.' J a M , '*to b e c u t off, diminish.'

Yqs,

'*to cut'; cf. J*», 'to

cut'; p a » , 'to b e c o m e short.' 'to t u r n aside, deviate.'

\rlcb, '*to bend, t u r n ' ; 3 5 6 cf.

'to i n v e r t ' ; Cs, 'to fall p r o s t r a t e . ' *

**to b e hollow,'

whence

'*to c u t out, hollow'; cf.

'flank,

side.'357 Si J

'to c u t o u t ' ;

'well'; ¿ J ^ - , 'to m a k e a hole, perforate.' 'to blow over, s t r i k e gently.'

y f h , '*to blow'; cf.

stem. But the word for 'twin' is a common heirloom and must go back to proto-Semitic. 353 As has already been shown (p. 26) most Arabic stems with prefixed t are secondary denominatives. «»I Cf. Mez, foe. cit. »55 Of. note 293. »56 \Kband '*to bend', are differentiations. 357 'to break the side,' or 'lead by the side,' are secondary meanings.

104

KOOT-DETEEMINATIVES IN SEMITIC SPEECH

'to blow, hiss (of a viper)'; ^li, 'to be diffused'; 'to blow gently.' 'to be fortunate, become glorious.' Vgd, '*to cut, destine'; cf. 'to be good, excellent'; j^», 'fortune, luck.' J>«, 'to become low, base, vile.' ydl, '*to be low, mean'; cf. J>, 'to be low, base'; J J o , 'to become low, base.' '* to be strong in belief, follow a right course.' Vsd, '*to be hard, firm'; cf. wX-ii, 'to be hard, strong, persistent'; 'to become strong, mature.' F. E T H I O P I C . "ThA:, 'to fix, fasten.' ]fkl i3 ), '*to make firm, fasten'; cf. Arabic J ^ , 'to rely on oneself'; Aramaic 'to confide in.' * T h £ ; , '*to bend oneself,' whence " » t h ^ t : , 'shoulder.' 9 fx y k f , '*to bend'; cf. Arabic ui-Xi, *)ri|, 'shoulder'; »li 'palm.' * f 7 £ : , '*to draw, drag,' whence o v ' T l C ' r o p e . ' \rgr, '*to harrow, drag'; cf. Arabic y*-, 'to drag, pull.' itO:, 3 5 8 'to flee.' \rn'a, '*to move'; cf. f 0 ® ' t o run down 'to shake, game'; Arabic ¿li, 'to be moved'; Hebrew wave.' it£ii, 'to cover.' yknp), '*to cover'; 359 cf. * h i i ' * t o cover,' whence ""hrt:, 'hidden, unknown'; Arabic 'to conceal.' i f t : , 'to destroy, demolish.' yst, '*to tear down'; c f . v + £ : , 'to lacerate, massacre'; Arabic 'to break up, disorganize.' 358 Praetorius ( B A . i. 25) thinks that the two following roots are secondary denominatives. 359 \ h n and \lcm, '*to cover,' are differentiations.

E X A M P L E S OF PRIM. SEMITIC STEM-FORMATIONS 1 0 5

i»*?:, 'to close up, obstruct.' y&g, '*to close up^; cf. '*to close up, capture^ in a net'; Arabic " f ^ i 'to complicate, confuse'; ¿yc^; 'to become obstructed (of the throat), be choked.' * } 7 r t ' * t o be uncovered, exposed,' whence ffD'}7¿V:, 'opposite, towards.' y~gl, '*to roll'; cf. 7rt©:, 'to uncover, cover'; *7rt?:, '*to uncover, make plain.' 'to run quickly.' \frdw, '*to run'; 36° cf. CA;, 'to run swiftly, assault'; ¿ & h ' t o run to the assistance of someone.' ¿7H:, 'to cut the throat, slay.' i f g z , '•to cut'; cf. *7HH:, 'to divide off, wall around'; 7 HO:, 'to divide up'; 7H1B, '*to break forth, sprout'; CHE, 'to tear, rend'; fcHE, 'to spread out'; D")E, 'to break in two'; "V]E, 'to divide': and pIB, 'to tear apart.' 361 The root gz, l *to cut, shear,' appears with various sufformatives in the following verbs: 'to cut off'; *DU, 'to crop'; 'to cut'; and ^>15, 'to plunder.' In Aramaic the root '*to hew, cut down,' yields 3Sn, 'to hew down'; fjsn, '*to cut deeply, 3«o yrd, yrt, and \rs (Hebrew pn) are differentiated biliterala. 861 The q may have been due to the influence of \ t out, hollow' 103. gb(3), '*to be high' 79. gd, l*to cut' 83, 104. gh, '*to face' 82. gl,