Reconstructing Proto-Afroasiatic (Proto-Afrasian): Vowels, Tone, Consonants, and Vocabulary 0520097998, 9780520097995

Citation preview

THEOLOGY AT CLAREMONT — OFOL SCHO

Mu WOON 4001703910

University of California Publications

- Reconstructing Proto-Afroasiatic (Proto-Afrasian) Vowels, Tone, Consonants,

and Vocabulary

Christopher Ehret

University of California Press

CLAREMONT SCHOOL

OF THEOLOGY

LIBRARY Given by Ernest W. Tune

cy) 1325 North College Avenue Clarement, CA S171

RECONSTRUCTING

PROTO-AFROASIATIC (PROTO-AFRASIAN)

Reconstructing Proto-Afroasiatic (Proto-Afrasian) Vowels, Tone, Consonants,

and Vocabulary

Christopher. Ehret

UNIVERSITY OF CALIFORNIA PRESS Berkeley ¢ Los Angeles ¢ London

Theology Library CLAREMONT SCHOOL OF THEOLO

aremont, CA

GY

UNIVERSITY OF CALIFORNIA PUBLICATIONS IN LINGUISTICS Editorial Board: Leanne Hinton, Larry Hyman, Pamela Munro,

William Shipley, Sandra Thompson Volume

126

UNIVERSITY OF CALIFORNIA PRESS BERKELEY AND LOS ANGELES; CALIFORNIA UNIVERSITY OF CALIFORNIA PRESS, LTD. LONDON, ENGLAND © 1995 BY THE REGENTS OF THE UNIVERSITY OF CALIFORNIA PRINTED IN THE UNITED STATES OF AMERICA

Library of Congress Cataloging-in-Publication Data Ehret, Christopher. Reconstructing Proto-Afroasiatic (Proto-Afrasian): vowels, tone,

consonants, and vocabulary / Christopher Ehret. p. cm.— (University of California publications in linguistics; v. 126)

Includes bibliographical references (p. ) and index. ISBN 0-520-09799-8 (alk. paper) 1. Proto-Afroasiatic language. 2. Reconstruction (Linguistics). I.-Title. II: Series: PJ992.E35° — 1995 492—dc20 95-13152 CIP

The paper used in this publication meets the minimum requirements of American National Standard for Information Sciences—Permanence of Paper for Printed Library Materials, ANSI Z39.48-1984.

Pee

bores ate

asi

a rae catai : , ey oot e

Bo,

oy

o areatte See

NY

-

|

|

ie _

:

obi ae a se SS), — » .

Contents

List of Tables, ix

Acknowledgments, xi Abbreviations,

xiii

Notes on Transcription,

xvi

‘Notes on Alphabetization,

xvii

CHAPTER 1. INTRODUCTION CHAPTER 2. PRELIMINARY CONSIDERATIONS IN AFROASIATIC (AFRASIAN) RECONSTRUCTION Bases for Reconstruction, 7 Chadic Evidence, 8 Omotic Consonant Reconstructions, 9 The Evidential Position of Berber, 12

Reconstructing Proto-Afroasiatic,

13

CHAPTER 3. DERIVATIONAL MORPHOLOGY IN EARLY AFROASIATIC (AFRASIAN) Stem, Word, and Syllable Structure in PAA, Nominal Suffixation in PAA,

15

15

16

Verb Extensions in Early Afroasiatic, 27 Prefixes, Infixes, and Reduplication in PAA,

51

CHAPTER 4. VOWELS AND TONE IN AFROASIATIC (AFRASIAN) RECONSTRUCTION The PAA Vowels, 55 Omotic Vowel Reflexes, 57

Vowels in Chadic and Boreafrasian,

Boreafrasian Vocalic History, 62 Tone in PAA,

67

Vil

60

55

Vili

Contents

CHAPTER 5. PROTO-AFROASIATIC (PROTO-AFRASIAN) CONSONANT RECONSTRUCTIONS Presenting the Data for PAA Reconstruction, 71 Presenting the Arguments and Supporting Evidence, 73 The The The The

71

PAA Labial Obstruents, 77 PAA Dental Obstruents, 120 Velar and Labiovelar Obstruents in PAA, 174 PAA Alveolar and Palatal Qbstruents, 251

The PAA Nasal Consonants, 298 The Laryngeal Consonants of PAA, 338 The Laterals and *r in PAA, 390 The PAA Glides, 452 ; Sources of Prenasalized Stops in Afroasiatic, 479

The PAA Consonant Inventory, 480

CHAPTER 6. THE SUBCLASSIFICATION OF AFROASIATIC (AFRASIAN) The Boreafrasian Subgroup, 483 The North Erythraean Sub-branch, 485 The Erythraean Branch of Afroasiatic, 487

The Cushitic Sub-branch of Erythraean, 488 Overall Subclassificatory Scheme, 489 Appendix 1.

Pre-Proto-Semitic Roots (Additional to those presented in Ehret 1989), 493

Appendix 2.

Proto-Cushitic Roots (Additional to those presented in Ehret 1987), 519

Appendix 3.

Additional Proposed PAA Roots, 523

Appendix 4.

Summary of Numbered Sound Shift Rules, 529

References,

533

Index 1.

Proto-Afroasiatic (Proto-Afrasian) Consonants,

Index 2.

Reconstructed Root Meanings,

543

539

483

Tables

Provisional Attestations Attestations Afroasiatic

Omotic Consonant Reconstructions of Nominal Affixes of Verb Extensions (Afrasian) Vowel Correspondences

Afroasiatic (Afrasian) Vowel Correspondences Revisited Afroasiatic (Afrasian) Dental Correspondences

Afroasiatic (Afrasian) Alveolar-Palatal Correspondences Afroasiatic (Afrasian) Lateral Correspondences Consonants of Proto-Afroasiatic (Proto-Afrasian) pod Subclassification of Afroasiatic (Afrasian) DOWIDNAWNS

1X

10 20 Sb) =| ia ZOO 394 481 490

It

|

Acknowledgments

This work is the culmination of one of several paths my research interests have led me along over the past two and a half decades, and a large number of people have contributed in substantial and important ways to the growth of my knowledge and understanding at different stages of the journey. The first steps on this particular path were my field studies of the Southern Cushitic languages, undertaken at a variety of times in the later 1960's and in the 1970's. Among the various speakers of those languages, Afraim Hayuma, Damas Dinya, and Aloisi Pius Kidara were especially helpful teachers and deserve special thanks for the time and effort they shared with me. I owe a great deal as well to Edward Elderkin, whose knowledge and gracious sharing of his own field‘research on the Dahalo language were essential to my development of an adequate historical-comparative reconstruction of the Southern Cushitic group. At the end of the 1970's and in the early 1980's, the establishment for a time of a close academic relationship between my institution, the University of California at Los Angeles, and the national university in Somalia had the unplanned consequence of turning my attention toward the study of the Soomaali language subgroup. I owe special thanks to Hussein Adam, who helped set up my first work on this topic in 1979; to my colleague Edward A. Alpers, who took time out from his own researches to assist me; to Ali Hersi, then head of the Somali Academy, whose efforts on my behalf provided me a place to stay and a place to work during my major field studies in 1982; and to Adan Maxamed Isaaq, whose unparalleled cultural knowledge and lively participation in the process of field collection, contributed immeasurably to the effectiveness of the work. Above all, I owe a debt of gratitude to Mahamed Nuuh Ali, who on many occasions between 1982 and 1985 discussed and studied with me the materials thus collected, gathered his own additional evi-

dence on the topic, and made extensive use of these data in writing his doctoral dissertation on the history of the eastern Horn of Africa. Over the years a number of scholars have been continuing sources of valued criticism, advice, correction, and apt example for me in pursuing work on Afrasian languages. These include most notably David Appleyard, Dick Hayward, Edward Elderkin, Bernd Heine, Phil Jaggar, Derek Nurse, Russ Schuh, and above all Robert Hetzron. Many others have contributed signifi-

cantly to the growth of my knowledge and understanding along the way,

among them Linda Arvanitis, Giorgio Banti, Lionel Bender, Hal Fleming,

Gene Gragg, Carlton Hodge, Grover Hudson, Marcello Lamberti, Ed

x1

xii

Acknowledgments

Meltzer, Paul Newman, Gerard Philippson, Joseph Pia, Mauro Tosco, and Rainer Voigt. Needless to say, none of these scholars are to blame for any errors that remain in the work. Finally, I owe special thanks to the copy editor of this volume, Orin Gensler, whose editing skills and knowledge of Afrasian and especially Semitic helped me to make this work much better that it could otherwise have been, and to the series editor, Rose Anne White, whose patient efforts and expertise oversaw the publication process from beginning to end. I know that I have surely missed mentioning others who truly deserve inclusion hete, and to themJ apologize most deeply. Several different granting agencies funded different phases of the research that has gone into this historical-comparative reconstruction of this family. My initial field studies of the later 1960's, which included firsthand work on the Southern Cushitic languages, was supported by a Foreign Area Fellowship of the Social Science Research Council. In the 1970's my researches were variously helped along by a Ford Foundation grant and by several smaller Faculty Senate grants of the University of California. The major work in Somalia in 1982 I pursued under a Fulbright research fellowship, while other Faculty Senate grants provided me subsequent support that contributed to the finishing stages of the Afrasian reconstruction.

Abbreviations

A. adj. ampl. and. anim.

Append. Arb. assim. attrib. BA BB BM C (Ge

Cs caus. CCh Ch. comp. conc. concom.

cont. Cush. demons. denom. deverb. dial. diffus. dimin. dissim. dur. EC ECh

Eg. ER

Eryth. ESA

Arabic © adjective amplificative extension andative extension animate Appendix Arbore (South Lowland Eastern Cushitic language) assimilation, assimilated attributive Boreafrican (see Chapter 6 for this subgroup) Bauchi-Bade (branch of West Chadic)

Biu-Mandara (grouping of Chadic languages) consonant Coptic consonant suffix causative extension Central Chadic (of Jungraithmayr and Shimizu) Chadic complementive extension; complement (noun suffix) concisive extension concomitant (prefix) continuant Cushitic demonstrative denominative deverbative dialect, dialectal diffusive extension diminuative dissimilation, dissimilated durative extension Eastern Cushitic Eastern Chadic (of Jungraithmayr and Shimizu) Egyptian East Rift (Southern Cushitic subgroup)

primary branch of Afroasiatic (see Chapter 6) Epigraphic South Arabian

Xili

X1V ext.

extend. fem. fin. foc. fort.

freq. H. HEC

incep. inchoat.

imper. intens. intr. iter. J Ji. LE M. masc. ME modif. MSA eV N NE n.

n. comp. NAgaw Ng. NOm nom. NSom obstr. PAA PBA PC Pen PEC PEth pl. PLEC PO pPS PR precip.

Abbreviations

extension extendative extension feminine finitive extension focative extension fortative extension ©

frequentative Harstsi (Modern South Arabian language) Highland Eastern Cushitic inceptive inchoative extension imperative intensive extension intransitive iterative Jungraithmayr and Shimizu 1981 Jibbali (Modern South Arabian language) Late Egyptian Mehri (Modern South Arabian language) masculine Middle Egyptian modifier Modern South Arabian (subgroup of Semitic)

middle voice extension Newman 1977; Newman and Ma 1966 North Erythraean (see Chapter 6 for this group) noun noun complement deverbative suffix North branch of Agaw (branch of Cushitic)

Ngizim (Chadic language) » North branch of Omotic nominalizing; nominal North Soomaali dialects obstruent proto-A froasiatic (proto-Afrasian) proto-Boreafrasian (see BA above and Chapter 6)

proto-Cushitic

proto-Chadic proto-Eastern Cushitic proto-Ethiopic (South Semitic subgroup) plural proto-Lowland Eastern Cushitic proto-Omotic pre-proto-Semitic proto-Rift (branch of Southern Cushitic) precipitive extension

Abbreviations

pfxd. pron. PS PSC PSom PSom-II PSom-III

redup. ref. S. SC Sem.

sing. SLEC Som. SOm stat. suff. ee

TV We Vie ven. WCh WOT

XV

prefixed pronoun, pronominal proto-Semitic proto-Southern Cushitic proto-Soomaali (South Lowland East Cushitic subgroup) proto-Soomaali-II (second stage of Soomaali differentiation into separate languages) proto-Soomaali-II] (third stage of Soomaali differentiation into separate languages) reduplicated, reduplication referential, reference

Sogotri (Modern South Arabian language) Southern Cushitic Semitic singular South Lowland Eastern Cushitic standard Soomaali South branch of Omotic stative suffix transitive terminal vowel vowel verb venitive extension West Chadic (of Jungraithmayr and Shimizu) Western Omo-Tana (South Lowland Eastern Cushitic sub-

group)

West Rift (Southern Cushitic subgroup) Yaaku-Dullay (branch of Eastern Cushitic)

Notes on Transcription

Voiced labial implosive. (1) In proto-Semitic (PS) reconstruction, denotes a proposed dental affricate [ts];

(2) in Soomaali (Cushitic) orthography, represents the pharyngeal voiced consonant [{]; (3) in Afar (Cushitic) orthography, denotes the pharyngeal frica-

tive [h];

(4) everywhere else, normally reflects a palatal or prepalatal affricate [¢].

Voiced alveolar implosive. In Soomaali orthography, denotes, depending on the particular dialect, either an apical voiced alveolar stop or a voiced alveolar implosive stop (elsewhere given as d’). (1) In Ngizim (Chadic), represents a voiced lateral fricztive [Kk]; (2) elsewhere, denotes a voiced lateral affricate. Voiced velar implosive. Voiced palatal or prepalatal implosive. (1) In Afar orthography, represents the pharyngeal vuiced consonant [7];

sh

(2) in Dullay and Yaaku (Cushitic), denotes a voiceless ingressive uvular stop; (3) elsewhere, has its usual uvular voiceless stop value [q] Voiceless (pre)palatal fricative [$] (used in Soomaali, Sidamo

[Cushitic], and Ngizim orthographies) (1) In proto-Cushitic reconstruction, denotes an ejective lateral. affricate [tl']; (2) in Ngizim, represents the voiceless lateral fricative [4];

(3) elsewhere, reflects a voiceless, non-ejective lateral affricate. (1) In Afar orthography, denotes an apical voiced alveolar stop (equivalent to Northern Soomaali dh );

(2) in Soomaali orthography, represents the pharyngeal voiceless fricative [h]; (3) elsewhere, has its usual voiceless velar fricative value.

Voiced (pre)palatal fricative [2] (in Ngizim orthography)

XVi

Notes on Alphabetization

The alphabetic sequence of consonants followed in this work is as shown:

prema ocd|.Gz, 1.0. 8°. yav6 Doh juk ky ke kw, Hee Pats. SS; it,t,, tl, ts,.wx,-X", y,.Z,\2, T.

m,n, 1.0),

For further commentary on the marking and ordering conventions and on the notations used for the data presented-here, see first two sections of Chapter 4.

XVil

:

hurr

eee

ae

~~

’

a

:

7

7 ‘ | t

os | ee

ee

.

Je

aa

Yo

: ae, i

Pat

7

i

°P

See, |

an ar?

ie tlh

+) lee

*)

heh oa

oe

a) Ye

ba

:

eS

at

:

de

~—

7

eet

pene orm

re Tt ae

es

rast ce

4— a

=.

=

Pe

Pad

,

f

ee

-

ge

= .

yis

am ih

is.

Ri

Mian

om

a ae

cae Bae

he = ay, = :

1”

st

Re seat

Genet 3

7

;

ae

pea: ; ses bir a —

.

:

+

lp Sh

iw ’:

%

eae

se

INTRODUCTION

The Afroasiatic (Afrasian)! language family has a history of scholarly acceptance almost as long as that of Indo-European, despite being a family of much greater internal diversity and historical time depth (Ehret 1979; Fleming 1977). Its comparative study has had many distinguished contributors, most notably in the past half-century Marcel Cohen (1947) and I. M. Diakonov (1965, 1988). But a full, systematic phonological reconstruction of protoAfroasiatic (Proto-Afrasian, PAA), with detailed substantiation and applying the normal canons of the historical-comparative method, has yet to appear. It is that deficiency which the present study seeks to remedy. This work provides therefore a comprehensive, systematic reconstruction of the family, concrete and specific in both evidence and argument. It proposes a full vowel and consonant reconstruction, along with a provisional reckoning of tone, and substantiates in detail each aspect of the reconstruction in an extensive comparative vocabulary of more than 1000 roots based on data from the Semitic, Egyptian, Cushitic, Chadic, and Omotic divisions of the

family. (Data from the sixth recognized division of Afroasiatic, Berber, proved of less immediate usefulness, for reasons discussed in Chapter 2.) Two requirements were rigorously imposed in the building up of this data base: (1) regularity of sound correspondences had to hold without exception for each segment (and suprasegmental, where known) of each root being compared; and (2) only minimal semantic differences, or else meaning differences clearly explainable on substantive morphological grounds, were allowed in roots being considered for possible cognation. A historical-comparative reconstruction of proto-Afroasiatic faces several potentially daunting obstacles. One of these is the great number of triconsonantal roots in Semitic that must be compared with the more commonly found biconsonantal stems of other subgroups of the family. A second is the lack of rigor with which the semantic dimension of reconstruction, in Semitic as much as in wider familial comparisons, has often previously been treated. A 1The author’s own preference is "Afrasian," as proposed and used by I. M. Diakonov and his colleagues (hence the notation in parentheses), and he intends in future publications to use that term exclusively. But in deference to wide Anglophone usage, the term "Afroasiatic" is still applied to the family in most instances in this work. Afrasian has a primary virtue of being a simpler formation. In English this formation has the effect also of deemphasizing the Asian element in the name; apparently in Russian it does this even more effectively. And this point is a very important one, because Afrasian is primarily an African language family, and the great majority of its history has been played out in Africa.

2.

Introduction

third potential source of difficulty is the immense variety and time-depth of the relationships among

the languages that make up the Afroasiatic family.

With respect to triconsonantal roots in Semitic, a comprehensive and sys-

temic, but certain to be controversial, explanation of the third consonants —

as lexicalized pre-proto-Semitic suffixal morphemes — has now been put forward (Ehret 1989). This scheme, and the linguistic procedures by which it was built up, provide one of the key steps in the Afroasiatic root and affix reconstructions carried out in this book. It has been applied here without apology because, quite simply, it works. The comparative materials from Egyptian, Cushitic, Chadic, and Omotic provide further point-for-point confirma-

tions of the specific meanings and functions of the great majority of the affixes proposed for pre-proto-Semitic. The validity of the scheme is resubstantiated time and again, throughout the long vocabulary of proto-A froasiatic roots by its explanatory power — its ability to account reasonably, economically, and aptly, in more than 2400 separate instances (indexed in Tables 2 and 3 in Chapter 3) for the individual variations of meaning that are found amongst the several thousand cited reflexes of those roots (Chapter 5 and Appendixes 1-3). The procedures involved in this internal reconstruction of third-consonant meanings have consequences for the second problem as well, that of semantic comparison. Specifically, the analytical groupings of roots distinguished by this approach reveal more sharply than ever before the pervasive polysemy and homonomy of a large proportion of Semitic verb roots. From the semantic patternings within these groupings, it becomes clear that in pre-protoSemitic, because of the loss of stem vowel distinctions in verbs, a great many

previously distinct roots had fallen together. A very large number of homonyms, in other words, were created in proto-Semitic by prior developments in verb morphology (Ehret 1989: 186-201). The recurrent polysemy of Semitic verbs can thus be laid to lexical polyphylesis, that is, to the derivation of a very great portion of seemingly single roots in proto-Semitic from two or more pre-proto-Semitic roots. Once this historical factor is recognized, the semantics of Semitic roots become much less confusing — the supposedly figurative development of opposed meanings for single roots, for example, probably received its initial impetus from the falling together of earlier Afroasiatic roots of disparate connotation — and the comparativist is thus able to begin confidently to identify the different clusters of meaning, and thus distinguish the separate source roots, that underlie so many ostensibly single roots. The making of over-imaginative semantic connections among Cushitic, Chadic, Omotic, and Egyptian roots is easily enough avoided through the stricture, applied in this work, of accepting only relatively obvious, unexceptionable meaning resemblances between prospective cognates, and by explicitly identifying, whenever they appear, any meaning-modifying affixes. Maintaining the same degree of rigor with Semitic comparisons requires the

additional step of sorting out the semantic confusions that pre-proto-Semitic root history itself created, quite apart from any later interventions of scholarly

Introduction

3

inventiveness. From a traditionalist viewpoint, it might seem necessary first to work meticulously through all the arrays of Semitic roots that have been proposed over the years. But in fact the task can be accomplished much more economically, and indeed much more effectively, by simply applying the techniques of internal reconstruction developed for the determinations of third-consonant meanings. The reconstructions of pre-proto-Semitic biconsonantal roots produced in Ehret 1989 form a sizeable body of stems already sorted out semantically. By the same procedures, a large additional body of such roots can and has been identified in the present work (for all of which see the many Semitic entries in the 1000-plus vocabulary of proto-Afroasiatic roots in Chapter 5, and also the additional demonstrations in Appendix 1 of this monograph). Even more important, this approach allows us to go a step farther back in time in generating our comparative data: by this means we recover earlier, pre-proto-Semitic root shapes, closer in time and appearance to their protoAfroasiatic etymons. Seen in-this light, a detailed reconsideration of past Semitic etymological work takes on a different aspect; instead of a prerequisite it becomes a subsequent task, more effectively pursued once there exists the external evaluatory apparatus that only the systematic reconstruction of Afroasiatic phonology can provide. It not only appears an unnecessary sidetrack, given the existence of a procedure fully capable on its own of revealing root forms and meanings of a stage preceding proto-Semitic, but it would commit us to data less ancient and less archaic of aspect than best suit our historicalcomparative purposes. Quite deliberately, then, this work selects its Semitic attestations principally (although not entirely) from either reconstructed preproto-Semitic forms or particular Arabic or Modern South Arabian root attestations from which such forms can explicitly be derived. These latter considerations bring us to the final problem, that of sure and effective historical linguistic reconstruction in a family considerably more ancient than Indo-European. The key elements for success are reasonably early historical-comparative reconstructions of at least some of the branches of such a family and, where this may be lacking, solid collections of data from phonologically conservative individual languages. Contrary to received opinion, detailed reconstruction of every subgroup of a family is not at all necessary to overall reconstruction, provided that the family has numerous member languages, some of them very accurately and extensively recorded, and that there exist among these languages a number of major branchings and that for at least a few of these branchings good reconstructions are available. The resources at hand for Afroasiatic reconstruction more than meet these requirements. The proto-Semitic consonant correspondences have, of course, long been known. Detailed reconstructions with extensive supporting data now exist for the Southern and Eastern branches of Cushitic (Ehret 1980 for Southern Cushitic; Sasse 1979, Arvanites 1990, and Ehret 1991 for Eastern Cushitic) and for Cushitic as a whole (Ehret 1987), while a less extensive re-

construction has been made of the Agaw subbranch of Cushitic (Appleyard 1984 and n.d.). A considerable body of Chadic roots also exists (Newman

4

Introduction

1977; Jungraithmayr and Shimizu 1981; Jungraithmayr’s expected revision of Chadic was not available at press time)

and can, with careful reanalysis, be

put to effective use in overall Afroasiatic reconstruction. In the case of Egyptian, which may always have been a single language, although at times dialectally diverse, it is important to use a consistent but sufficiently large body of relatively early occurrence, and for that reason attestations are taken mostly from Middle Egyptian (Faulkner 1964) and only rarely from Late Egyptian (Lesko 1982-90). Finally, a provisional reconstruction of Omotic consonants and vowels is offered in the present work (Chapters 2 and 4, q.v.; cf. also Bender 1988).

Whenever possible, the particular attestations for a proto-

Afroasiatic root (in Chapter 5) are drawn from these various reconstructions. Of course, in the case of Semitic, for the reasons noted above, an important

source is reconstructions of the pre-proto-Semitic forms. Berber data figures in the outlining of proto-Afroasiatic vowel history, but otherwise is little drawn on here (for reasons, as noted above, that are described in Chapter 2). In addition, from each major branching of Afroasiatic one or more very well recorded languages, usually ones with important archaic features in their phonologies, have been chosen for special representation among the data cited for their branch. From Semitic were selected Arabic, which retains all the standardly accepted proto-Semitic consonant distinctions except one (though often not with their original pronunciations) and is exceedingly thoroughly recorded, and the Modern South Arabian languages, usually Jibbali and Harstsi, which are particularly important for their retention of lateral values for the old Semitic (and Afroasiatic) lateral obstruents, and for a number of

other phonetic archaisms. From Cushitic a variety of individual languages provided valuable information, especially Soomaali, for which a very large and reliable dictionary exists. As for the Chadic branch, Ngizim data proves especially useful because of its preservation in many cases of the probable proto-Chadic vowel system. The reconstructed roots of proto-Southern Cushitic (Ehret 1980) and lexemes from Ngizim in Chadic, Awngi in Agaw, Arbore in Eastern Cushitic, and Mocha, Bench, and Yem in Omotic jointly offer crucial evidence for the provisional reconstruction of proto-Afroasiatic tone. Other individual Cushitic and Omotic languages are cited from time to time in instances where they may happen to preserve old roots as yet unrecorded for other languages of their branch. There is, finally, another obstacle of a different kind facing those who at-

tempt the historical-comparative reconstruction of Afroasiatic, potentially much more daunting than all of the practical obstacles put together, and that is the immense wealth of previous literature on the family and the enormous accumulation of specialized knowledge to be found among the thousands of publications on the family, and above all on its Semitic branch. So particular and so varied are the expertises and specialities involved that almost nothing can be said that will not be opposed, often with great vigor and warmth of feeling, by someone in this vast field. This reservoir of accumulated

knowledge, moreover, has over the decades generated all sorts of historical

hypotheses, some widely held and others idiosyncratic, some explicit and

Introduction

5

consciously adhered to and others implicit and perhaps not even recognized as such by their adherents. The scholar who reconstructs proto-Afroasiatic must thus be prepared to face, along with much substantive and justified criticism, the magnification of individual mistakes into fundamental flaws, the assertion of a priori views as though they were factual counterevidence, and sweeping dismissals of the work based on agendas sometimes hidden even from those who hold them. The central point to be made here is that we must not allow previous views to impose themselves on the structuring of the data or on our responses to the conclusions that emerge from the analysis. That a particular Semitic root may have been proposed in a past work (for example, in Cohen 1947) as cognate with a particular word found in some Cushitic languages does not mean its automatic acceptance as a proto-Afroasiatic root today. Now that we actually possess relatively full reconstructions of proto-Cushitic and of the Eastern and Southern branches of Cushitic, and partial reconstructions of the

Chadic branch — major additional bodies of evidence not available even twenty years ago — we must ask all over again, "Do these formerly accepted items actually show fully regular sound correspondences? Can their points of dissimilarity be explained by defensible morphological arguments? Do the semantics still seem plausible?” Inevitably, many of the lexical connections posited in the past and often widely accepted today prove wanting in the face of these criteria, while others not previously considered or not known at all turn out to work quite well. Nor can we afford to let the great imbalance of interest among Afroasiaticists in favor of Semitic, and secondarily toward Egyptian, preshape our historical views.

The Cushitic, Chadic, and Omotic divisions of the family

are not pale reflections of Semitic.

Each is an equally important, separate

source of comparative evidence, and their contributions, combined with those

of Semitic and Egyptian, depict a proto-Afroasiatic language only distantly similar in structure to its Semitic daughter tongues or to Egyptian. It must be reiterated that the reconstruction of Afroasiatic proposed here follows throughout the established techniques and approaches of the historical-comparative method. It is systematic, comprehensive, and rigorous, with unyielding phonological and demanding semantic requirements for cognation. It is detailed and thorough in its etymological analysis, providing at the minimum a parenthetical explanation of almost every morphological modification of each cognate form and an explicit justification of any less-than-transparent semantic connection. Its citations from individual languages, where they appear in the data, are also given parenthetical phonological derivations from their proximate proto-language whenever that is needed; and such derivation in each case rests on established reconstructions of the correspondences involved. The reconstruction may bypass areas and resources of concern to other scholars, but only when its own routes of inference and substantiation seem more direct and effective than those not taken, and every bit as sound. It uses a body of evidence which is, for Semitic, still controversial, but does

so precisely because the validity of that evidence appears strongly corrobo-

6

Introduction

rated by the wider Afroasiatic comparisons presented here. Its conclusions may often surprise, and some of its consonant correspondences may be unexpected. But its arguments and conclusions rest on sound method and solid and detailed evidence.

2

PRELIMINARY CONSIDERATIONS IN AFROASIATIC (AFRASIAN) RECONSTRUCTION

BASES FOR RECONSTRUCTION Two recent works (both mentioned in Chapter 1) contribute especially to making the systematic reconstruction of proto-Afroasiatic (Proto-Afrasian; PAA) an eminently feasible enterprise. The first is a comparative-historical reconstruction of proto-Cushitic (PC), and the second, an internal reconstruction of the origins of the third consonants (C3) of Semitic triliteral roots (Ehret 1987, 1989). The earlier of these studies (1987) permits the establishment, on a reliable basis, of regular sound correspondences between proto-Semitic (PS), proto-Cushitic (PC), and the third well studied division of Afroasiatic, Egyptian (Eg.). The comparison among these three lies at the heart of the present effort. The second study (1989) allows for the first time an effective morphological analysis of Semitic triconsonantal roots and in consequence the identification of a great number of underlying pre-Semitic biliteral stems with which to compare the biliteral roots of Cushitic and Egyptian. The internal reconstruction developed for Semitic C3 meanings turns out, moreover, to explain equally well the third consonants of the less numerous Cushitic and Egyptian trilieral roots. The comparative evidence thus confirms conclusions reached previously from Semitic evidence alone, and demonstrates that the morphological processes from which the third consonants derived were productive at very early stages in the history of the Afroasiatic languages (see the text and tabulation of comparative data in Chapter 5 for extensive exemplification of these features). Prefixation, in contrast, was not a common derivational process in very early Afroasiatic; it became predominant only later, in the Boreafrasian subgroup (for which see the subclassification of Afroasiatic presented in Chapter 6). A point already brought out in relation to Semitic — that, more often than not, the particular PS root shapes tend to appear polysemic by reason of their polyphyletic origins — requires iteration here. Because of the Semitic verbal system's capture of the vowel realizations in PS verb roots, a great many instances were created of formal homonymy between what were once separate PAA roots. Identical in their PAA consonants, such roots were distinguished originally by different stem vowels and prosody — features that have since been lost. A second and lesser, but still important contribution to Semitic homonymy has come from the falling together in PS of several pairs of PAA consonants. The failure to recognize the pervasive polyphylesis of PS roots

8

Preliminary Considerations in Afroasiatic Reconstruction

has led to much imaginative but specious explanation of divergent meanings for root shapes having identical PS consonants but distinct PAA provenances. The West Semitic root *4{r "barley" need no longer, for example, be explained by assuming barley to be the "hairy grain" (e.g., Diakonov 1981); it and homonymous proto-Semitic *4{r "hair," as the evidence for PAA roots #887 and 889 in Chapter 5 shows, can be derived much more straightforwardly from distinct Afroasiatic roots. The polyphylesis of such roots has not contributed merely to scholarly confusions. Undoubtedly it has also significantly influenced the actual semantic evolution of Semitic languages*— by allowing or encouraging the emergence of semantic usages which combine or blur originally distinct meanings, by building metaphorical linkages which might not otherwise have been chanced upon; and by the development of opposing meanings for a seemingly unitary verb root. (The origin of such oppositions in distinct underlying roots is apparent in a number of the examples given in Ehret 1989, although not explicitly commented upon there.) Once the multiple origins of PS roots and the morphological provenience of third root-consonants are recognized, a great many clearly comparable Afroasiatic roots emerge for the first time into view. Besides Cushitic, Egyptian, and Semitic, two other major divisions of the Afroasiatic family, Chadic and Omotic, also contribute in significant ways to the overall reconstruction of proto-Afroasiatic (PAA) — among others, to our understanding of PAA tone and vowels. Berber, the sixth recognized division of the family, plays a lesser, but not entirely insignificant, role in the work undertaken here. Because these three language groups tend to be less well studied, they offer data that, in several respects, require further interpre-

tation and explanation before the details of PAA reconstruction can be laid out.

CHADIC EVIDENCE Proto-Chadic (PCh) root reconstructions have been offered by P. Newman (1977, and earlier, Newman and Ma 1966) and, in considerably larger

numbers, by H. Jungraithmayr and K. Shimizu (1981). Newman (1977) provides both consonants and vowels in his reconstructed PCh roots; Jungraithmayr and Shimizu frequently offer only consonants. Newman's consonants are certainly too few to account for the full range of Chadic sound correspondences, notably those involving the laterals and sibilants. His reconstruction of the PCh vowels, however, seems very probably to be correct in most cases, since they appear to produce regular correspondences (somewhat obscured at times by apparent ablaut effects) with the proto-Cushitic vowels. Jungraithmayr and Shimizu's reconstructions account for the range of PCh consonants far better than Newman's, as the comparative data presented

in Chapter 5 show, although they embody a major problem of their own. Certain kinds of distinct PCh consonants have fallen together with other con-

Chadic Evidence

9

sonants in so many of the modern Chadic languages that the diagnostic evidence for distinguishing among them has been maintained in relatively few languages. From the comparative PAA evidence, it would appear that, instead of explicitly noting alternative possible reconstructions when relevant diagnostic examples are lacking, Jungraithmayr and Shimizu have simply chosen to group such cases with legitimate examples of one of the possible alternative consonants. For example, they postulate on a quite solid basis three PCh laterals in addition to *1, denominating them as *; (or simply *4), *4, and *k. The latter two each show a consistent and distinct four-way set

of correspondences with particular Egyptian, PS, and PC consonants.

Their

*¢ (*4,), in contrast, seems to be the catch-all member of the lateral set. It

normally participates in a third four-way match-up with particular Egyptian, PS, and PC consonants, confirming its status as a valid additional PCh lat-

eral. But *4 is also capable, in the Jungraithmayr and Shimizu, lists of having the same extra-Chadic correspondences as their *42— such instances presumably being cases of actual *42 in-which evidence to distinguish *4 and *42 was lacking in some way (for particular cases see section on PAA laterals in Chapter 5). Similarly, their *s (*s,) seems to be the catch-all representative of a sibilant set composed of PC€h *s, *s>, *s3, and *s' (see sections on PAA dental

and alveolar-palatals in Chapter 5). It should be noted here that the distinctive outcome on which a separate *sz is based appears in only one section of one sub-branch of one branch of Chadic, and its occurrence even there may have

an extremely restricted environment — seemingly found word-initial only where the second consonant is a voiced labial (*b or *w), and non-initial only

after a preceding *g. Its manifestations are thus probably the regular reflexes of PCh *s or *s3 in a restricted pair of environments in one particular subgrouping of Chadic, and no distinctive *s2 should therefore be reconstructed for the PCh language.

OMOTIC CONSONANT RECONSTRUCTIONS The Omotic languages emerge from the available comparative data as definitely Afroasiatic. The demonstrations in Fleming (1969, 1974) and Bender (1975) that Omotic forms a division of the family quite distinct from Cushitic seem fully convincing (contra Zaborski 1986), and the evidence discussed in

Chapter 5 below further validates its distinctiveness not only from Cushitic but from the other subdivisions of the family. A provisional partial reconstruction of Omotic is formulated here, in order

to facilitate the inclusion of Omotic evidence among the comparative data. The observed consonant correspondence patterns across the group are noted in Table 1 following. For a majority of the known languages of the group, the correspondences are those that are overtly apparent in the comparative wordlists of Bender 1971. But in several instances these data can be confirmed or expanded upon by other materials (inter alia, Mocha evidence from

10

Preliminary Considerations in Afroasiatic Reconstruction

Leslau 1959, Koyra collections in Hayward

1982, other Ometo data from

Hayward 1987, and especially high quality sets of data for Yem, Bench, and Ari in Hayward 1990). Bender (1988) provides additional comparative Omotic material independently supporting a reconstruction of PO consonants differing only slightly from that offered here. There remain many gaps in the correspondence sets, and some patterns are as yet known from only a single kind of word environment. If no environment is specified in Table 1, then the reflexes cited here for a PO consonant can be considered to be general ongs. In the table the proposed PO consonant reconstructions and supporting data are listed in the same order as the order of presentation of the PAA consonants from which they derive in the comparative Afroasiatic materials of Chapter 5. Here and in subsequent tables the sign @ represents a zero-reflex of a proto-consonant. Blanks in the table indicate a lack as yet of relevant comparanda from the language group in question. Three languages, Wolayta, Male, and Zayse were chosen to represent the Ometo subgroup, and two languages, Kafa and Mocha, to exemplify the Gonga subgroup. These subgroups together with Yem, Bench, and Maji constitute the North Omotic branch. Dime and Ari represent the two subgroups of South Omotic. Table 1.

PO

Provisional Omotic Consonant Reconstructions

Wolayta Male Zayse Yem

*bh ft /V_V

ep p ft t_ iN

Bes

Kafa Mocha

—-b

i

b

bab

w

w

Ww

w

b,

ie

p fi

p p

6 f

DMM

Del

iper

Leb

f

w

gh Dp. op f

p

Bench

Maji

Dime Ari

b

brlipag

ep

b

b

b

p Pp

PAO NE foot

DS DIN erpaa aaa

p'~b

IV_

tdi IN-=? *

oc

eats IV_

IN YS

ied ete

d

coh steeds ae dvantdan camed

d d

oo

co

oc

o

Z

Gz

oc

fons

d

y

Z d

ts

(SUN

zy

y

ons

y y

Or

dx ay ad o

Z Z

+

o

ase

zZ

d

S

S

dztdisz

(oe *S

S

s

S

S

S

Ge

GLRS

Omotic Consonant Reconstructions

Table 1. PO.

*ts'

i

ts'

Se

*g

g

g

g

k

k

*k

pris.

ck ty ke

es

k

ee

ih

6-4

h

k/_-#

*i'

/#_

k'

Kafa Mocha

Ce

k

k

ees

wk)

c

{S:SiaeeZ

g9

9g

g

k

k

k

wake

(NS

h

§ O-

k

G6

k

k

Ge

Ge

k

@

k

k'

q(k']

k'

z

¥

§

%2/e

S

c

s/_-#

Z

k

k'

g

c

ts

ts

c/o

esi

ly

c

c

ts

*ts /V_

S

S

ts

S

§

cS s/_-#

ts

LOS

x

&(?)

c 5

c

§

s

S

§&

s

5 §

g

re

s

SHES

TCs

§

5

§

§

§

*So/V_

§

§

C

g

SRS

roo S

suc

SAS

é

Z is

c

ts

q

Z

c

ts

g’

ets

c

tae a ING

g

Go

fol Ai

's

kitietk

k/

/n_ k'

tS:

q

Z

/V_V [V_-#

ts”

ak

IN=

=tsts/ Vt

Dime Ari

k

g/V _#, ~

Pe

Bench Maji

*y [Ht

#54 /#t

ieae

Provisional Omotic Consonant Reconstructions (continued)

Wolayta Male Zayse Yem

*x, /#_

!

ts

(e

S

§

§

§

aac

‘e

‘oats

ays

ch shies tsoft

ts' ts'

ish tS aw tsAre

ts

Cuisite'

rest

(eR

‘oat

got

or

atoly

oe

Ce

(oh

*m

m

m

m

m

m

m

m

m

no

m

*n

n

n

n

n

n

n

n

n

n

n

€/_u

Preliminary Considerations in Afroasiatic Reconstruction

12

Table 1.

Provisional Omotic Consonant Reconstructions (continued) Wolayta Male Zayse Yem

PO

n

n

q

2

©

©

. »

h,®

h,@

h

ow

Ww

©

h

d n

d n

d

]

]

l

l

]

l

1

]

t

d'

n

n

n

ey IN

2?

2?

QQ)

*hi /#_

h

h

*ho /#_

oh

Ww

Ww

w

*]/#_VC /#_Vb

1

l

l

n

/#_VN

n

]

n

n

]

l

]

[V_

Dime Ari

n

n

an

h

Kafa Mocha Bench Maji

h,®

n

2?

7 4)

et"

d'

d'

t

t

Be

€:

a

ral

d'

r

t'

te:

t'

3 a AY

r

r

r

yenhhe

Se

r

r

c

ac

r

*w [Ht

Ww

w

Ww

Ww

Ww

Ww

Ww

Ww

w

ew

ay)

y

y

y

ye

wy,

y

y

Yi

OV

aaa,

ed [V_

d'

THE EVIDENTIAL POSITION OF BERBER

Evidence drawn from the Berber subgroup of Afroasiatic plays a much more minor part in the present work. It can be and is used here to further our understanding of vowel history in the family (see Chapter 4). Certain lexical innovations of the Berber group can already be recognized also, as developments shared with certain other divisions of Afroasiatic and therefore of value in the subclassification of the family (as noted in Chapter 6). For two interconnected reasons, however, Berber attestations have been left out of the direct comparative tabulations of PAA roots in Chapter 5. In the first place, proto-Berber greatly reduced its range of consonant distinctions via its apparent merging of many of the original PAA consonants and its resulting loss of key Afroasiatic phonological distinctions, most notably in the lateral, pharyngeal, and ejective categories. Such a contraction of the inherited consonant system is bound to have added considerably to the difficulties of identifying the particular Afroasiatic sources of many Berber roots. On top of this, the kind of extensive reconstruction of proto-Berber lexicon that might help in sorting through alternative possible etymologies is not yet available.

Evidential Position of Berber

3

Reliance on Berber consonant data is thus at this stage in the development of PAA reconstruction as likely to obscure as to clarify the process of discovery. In contrast, Cushitic and Semitic and also Middle Egyptian, except for its collapsing of the PAA sibilants and certain ejectives, have each retained a significant proportion of the key complexities of PAA consonantism. ProtoChadic and proto-Omotic were clearly of similar rententiveness, although the reconstructions of these two are less advanced than those of PC and PS. Each provides the range and variety of consonantal distinctions from which an effective discrimination of valid from invalid etymological connections can usually quickly be made and on which an overall PAA phonological reconstruction can effectively be built. RECONSTRUCTING PROTO-AFROASIATIC In the next three chapters the reconstruction of PAA is set forth. The effort differs from other works of similar intent in presenting systematically and in detail, here and now, the specific comparative evidence for each of its phonological and root reconstructions. From these extensive data, contained principally in the very long Chapter 5, it proceeds to establish vowels and prosody — both actually argued for in Chapter 4 — as well as the consonants of PAA, which form the organizing focus of Chapter 5. The evidence in that chapter shows that many previous conjectures regarding the consonant system of PAA are indeed solid postulations, e.g., the distinct sets of velars and labial velars, the four distinct oral labials, and the several laterals. Diakonov's

(1988) list of "Common Afrasian" consonants is largely congruent with the PAA inventory required by the data presented here, although he also includes at least four for which the writer can find no basis: *x and *x~ plus four nonglottalic dental affricates where only two can be solidly demonstrated. The Omotic data do offer one consonant, however, that might someday be found to fit in with one of Diakanov's additional consonants. The correspondence set represented as *h in Table 1 just might support *hY, since its reflexes, known so far only in word-initial position, are variously /h/ and /w/ across the Omotic branch. In the four roots noted in Appendix 3 (#10121015) for which an Omotic reflex containing *hz can be cited, this sound cor-

responds to *h in most of the rest of the Afroasiatic, but possibly to *w in Chadic, again best favoring an original PAA pronunciation [hw]. But only those four cases can currently be given tentative identification, and so for now the existence of a PAA *h¥ must remain a conjecture for future investigation. At the same time, the data presented here confute various other conjectures and proposals about PAA. Notably, the word and stem ("root") structures posited by Diakonov (1988) may be valid as secondary developments within the Boreafrasian subgroup of the family (see Chapter 5 below on subclassification), but cannot be projected back to PAA (see Chapter 3 for the simpler stem structures which can in fact be postulated for PAA). Syllabic

14

Preliminary Considerations in Afroasiatic Reconstruction

"sonants" similarly may have emerged in conjunction with the development of the Boreafrasian stage, and independently in a couple of instances elsewhere in the family, but have no place in the comparative picture for earlier PAA (again see Chapter 3). Finally, while a considerable degree of congruence may obtain between the consonant inventory established here and previously published conjectural or comparatively-based inventories, the actual reflexes of the particular consonants can often be quite different from those previously suggested, as for instance in Bomhard's (1988) work and, one may suspect, sometimes even in Diakonov. The most striking such case is that of the Egyptian palatals dg, ¢, and §, which can be shown consistently and resound-

ingly to derive not from the PAA palatals, or normally even from palatalized dentals or velars, but from the lateral obstruents of PAA. Chapter 5 presents more than 130 PAA roots containing these three laterals and cites around 2530 Egyptian examples of each of the three correspondences. A historical connection between lateral and palatal obstruents is neither surprising nor unprecedented in the world's languages. Such a link is clearly attested even in Semitic, notably by Arabic /§/ from PS *4 But the lack of previous recognition of the correct Egyptian correspondences in this instance does underscore a point which comes forth very strongly from the data of the succeeding chapters. Bending of sound correspondence rules, assumptions of "unstable" consonants, over-imaginative stretching of semantic resemblances, and other ad hoc measures need not and should not be invoked any longer in PAA reconstruction. The comparative-historical method, rigorously applied, works very well indeed for Afroasiatic once sufficient data and some of the key intermediate levels of reconstruction are available to the comparativist.

DERIVATIONAL MORPHOLOGY IN AFROASIATIC (AFRASIAN) RECONSTRUCTION

STEM, WORD, AND SYLLABLE STRUCTURE IN PAA The laying out of the comparative Afroasiatic data, undertaken in Chapter 5, shows that just two fundamental stem shapes can be reconstructed for proto-Afroasiatic, CVC and C(V), the latter having the possible alternative shape VC in verb roots. To the stem could be added any of a number of nominalizing suffixes of the form:-(V)C- or any of a great variety of verb extensions of the shape -(V)C-. The evidence makes it probable that the under-

lying form of such suffixes was usually -C-, with the surfacing of a preceding vowel depending on, and its particular realization in different Afroasiatic subgroups predictable from, the syllable structure rules of the particular groups. (The particular outcomes of such processes will not be further argued here, but will be left to future studies.) Afroasiatic roots containing such suffixes are therefore given in Chapter 5 in the form *C;VC2Cs, where Cs represents the suffix. Two exceptions would have been the nominal suffixes in *w and *y, which probably did have fixed vowel accompaniments and -VC shapes (see suffixes #1 and 2 below). The Omotic, Cushitic, and Chadic evidence conjoin in requiring the existence in PAA of an additional element in word formation, a terminal vowel (TV) in nouns and modifiers, the original function and meaning of which re-

main obscure. TVs have been subjected to comparative-historical investigation in only two groups of Afroasiatic languages. In Omotic they have no reconstructible function beyond their necessary attachment to singular noun stems in semantically unpredictable fashion. With the exception of Kafa, in which two TVs, -o and -e , have been grammaticalized respectively as mas-

culine and feminine markers, they carry no grammatical or recognizable semantic load (Hayward 1987). In proto-Southern Cushitic, pairs of TVs formed a variety of singular-plural markers. Particular paired sets tended to go with either masculine or feminine nouns, but an individual TV on a singular noun generally gave no indication of the grammatical gender of that noun (Ehret 1980: 49-50). From these indicators it seems reasonable to conclude that TVs are fossils of a nominal morphology productive in pre-proto-Afroasiatic and predating the rise of grammatical gender in the family. Having lost their original grammatical function, they have been reanalyzed as markers of the singular or sometimes, as in the case of Southern Cushitic, of the plural in nominals. In the Boreafrasian subgroup (Semitic, Egyptian, and Berber: see Chapter 6 for

|)

16

Derivational Morphology in Early Afroasiatic

this classification), the TVs have generally been dropped entirely, leaving most nouns and adjectives as consonant-final words. The existence of TVs at early stages of Afroasiatic evolution obviates the need to reconstruct any syllabic consonants for PAA. The usual word structure of nouns and adjectives would have been *C;(VC2)(Cs)Vty, in which the only possible syllable structures are CVC and CV and never just C. The presence of syllabic C in Boreafrasian languages can be understood as the nat-

ural outcome of vowel loss, whether word-internal or word-final, within that

particular subgroup (as is also separately the case in a few modern Omotic languages, notably Bench and Maji, where the same kind of sound change has independently been at work). NOMINAL SUFFIXATION IN PAA

At least ten and possibly thirteen nominalizing suffixes can be reconstructed for very early Afroasiatic. They occur in Semitic, Egyptian, and Cushitic with the same or easily relatable meanings and functions, and most of them can also be observed in the Chadic and Omotic data presented in Chapter 5 below.

1. *w (*-aw-) deverbative complement and result suffix.

This suffix, found in all divisions of Afroasiatic, seems most often to form deverbal result, complement, and attribute nouns. It is not unreasonable to consider *w as having originally been a deverbative complement and result marker, and its attributive application to have arisen secondarily. An original shape *-aw- for this suffix is indicated by its attestations in Cushitic, Chadic, and Omotic. 2. *y (*-ay-, *-iy-) attributive deverbative and attributive noun suffix. 2a. *y (*-ay-, *-iy-) adjective-forming suffix. This suffix can operate as a noun-forming deverbative in Semitic, Egyptian, Chadic, and Cushitic instances, but is also often added to nominals to form

attributives — names of things having the attribute(s) of, or associated by location or resemblance with, the item named by the stem to which *y is affixed. It appears as an adjective-forming suffix in all the divisions of the family (see citations listed in Table 3, item 2a). Evidence from Chadic and

Semitic as well as Cushitic requires *-ay- as one shape of this affix, while an alternative form *-iy- can be found in Omotic and Cushitic reflexes and in Semitic cases (i.e., Arabic nisba -iy-). In West Semitic a relatively rare prefixed version, *y-, of this affixal pair is also present (Moscati 1964: 80). 3. *t associative noun suffix.

Nominal Suffixation in PAA

ee

A noun formative in *t is well attested all across the Afroasiatic family. Its original function may no longer be clearly specifiable. It appears to have been especially productive in Egyptian, as a variety of entries in the comparative lists show, forming noun instruments, attributives, and complements from verbs as well as deriving nouns from other nouns of related or associatable meanings. This latter function has also been observed in Cushitic derivations (Ehret 1980: 53-54), and hence the formative is called an "associative" marker

here. The suffix apparently remains productive in many Afroasiatic languages today. Semitic *-tt/*-it forming abstract nouns is presumably a direct reflex of it, and it is the probable source, although shifted to prefixed position, of

the Semitic *t- verbal noun marker. It is not to be confused with the widely found *t feminine marker of Afroasiatic. A distinct Afroasiatic suffix in *t, a nominal plural marker, may be reflected in the Egyptian cases where *t indicates a collectivity, such as a flock ora crew (e.g., #322, 367, and 470), and in the *-Vt- plurals of Southern Cushitic (Ehret 1980: 54-55) and Eastern Cushitic (e.g., Afar entry in #579).

4. *t adjective suffix. An adjective suffix in *t, often still productive, is prominent in Cushitic (Ehret 1980, 1987) and is more weakly attested in Egyptian, Semitic, and apparently Omotic.

5. *m attributive noun suffix. 5a. *m adjective-forming suffix. A deverbative suffix in *m, forming attribute and result nouns, must also be

posited. It is common in Semitic in the C3 position and is well attested also for Egyptian, Cushitic, Chadic, and Omotic. In a few instances in Cushitic (see #635 in Appendix 2) and Chadic (see #30, 228, and 295) and in at least one Semitic case (Ehret 1989: Table 34a, Arabic "sword"), this suffix ac-

quired an instrumental connotation; but the bulk of the examples are clearly interpretable as attribute, or sometimes possibly result or complement, formations. The use of *m as an adjective formative in a number of cases in Semitic, Omotic, and Chadic argues for it having had an originally attributive connotation. The *mV- instrument-agent prefix of Semitic, Egyptian, and Chadic is argued below (this chapter) to have had an origin quite distinct from that of this suffixed *m deverbative. 6.

*n attributive (7) noun suffix.

6a. *n adjective suffix. Another nasal, *n, also functioned at times as an attributive suffix, but its

scope more closely paralleled that of the *y attributive (q.v.). Like *y, it appears frequently to have produced adjectives from verbs or nouns. It is

Derivational Morphology in Early Afroasiatic

18

known from all branches of the family. been *-an.

Its Semitic reflex appears to have

7. *1 attributive and complement deverbative suffix. 7a. *1 adjective suffix. A noun-deriving suffix in *] turns up widely in Afroasiatic with a variety of effects. In pre-proto-Semitic (pPS) it can be proposed to have been a nounpatient and noun-complement formative (Ehret 1989: Table 13a). Examples of the suffix in Egyptian seem often to go with attributive nouns or nouncomplements, while a similar function may existed in Chadic (e.g., Chadic entries in roots #654, 864, and 909) and in Cushitic.

In Cushitic, *1 became

especially prominent as a suffix in animal names, probably because such names not infrequently derive from roots descriptive of the animals’ attributes, i.e., their appearance or behavior. Like *y and *n, *] became important as an adjective-forming suffix, with a variety of proposed examples of this usage attributable to Semitic, Egyptian, and Cushitic (see Table 3, item 7a below). At least one example of *1 as an adjective formative also appears in the Omotic evidence (root #546). 8. *r instrument and complement deverbative suffix.

Noun suffixes in *r are common in the comparative data, but their reconstructed meaning is unclear. For pPS, *r has been suggested to have been a noun-complement marker (Ehret 1989: Table 11a); but in Egyptian PAA *r (> Eg. 3 /_#) seems fairly commonly to have marked instrument or means (e.g., roots #170, 546, and 713). In still other cases *r appears to have carried little

or no semantic load even when quite clearly suffixal in origin (e.g., entry for Yem [Omotic] in root #11). A possible early noun singular marker in *r known from Cushitic (Ehret 1987: 9) may have been at times conflated with

the *r deverbative and thus contributed to its semantic opacity. 9. *r adjective suffix.

In Cushitic, Egyptian, and pPS (Ehret 1989: Table 11b), a probably distinct *r suffix (realized in PSC as *-ar-) forming modifiers, usually from verbs, can also be identified (see Table 3, item 9). A couple of instances of this suf-

fix se Chadic (#268 and #497) and one from Omotic (#262) can also be

cited.

10. *b animate noun and deverbative noun suffix.

A suffix in *b, used to derive from verb roots nouns that name animals and

parts of the body, is strongly attested in Cushitic (Ehret 1987) and in Semitic (cf. Diakanoff [Diakonov]

Egyptian and from Omotic.

1965) and is evident in several instances from

Nominal Suffixation in PAA

19

Three further nominal derivational suffixes can be given a more provisional reconstruction. 11. *s deverbative complement suffix.

The case for this suffix seems relatively strong, and is supported by a number of plausible occurrences in the comparative data of Chapter 5 and also Appendix 1. The known examples are limited to Chadic, Semitic, and Egyptian, especially the latter two (Table 1, item 11 lists the proposed instances). The two other possible formatives — 12. *y attributive noun suffix; and 13. *? adjective deverbative suffix —

are both best known from Cushitic examples of derived adjectivals or adverbials. The first of these two suffixes, in *9, was productive in Southern Cushitic (Ehret 1980: 59-60); it also appears in at least one Chadic citation in the comparative data (root #641), where it forms an attribute noun, and is tentatively proposed in one early Afroasiatic root (#916). The second suffix, in *?, also can be identified in Southern Cushitic and possibly elsewhere in the Cushitic branch of the family (Ehret 1980: 60; Ehret 1987: 9). Several

probable cases of *? as an adjective formative appear to be present in the Semitic evidence as well (cf. also possible examples in Ehret 1989, e.g., in Table 38, no. 44), and several Egyptian adjectivals may also contain it (see Table 3, item 13 below for particular citations).

With the exception of *w and *y, these suffixes cannot as yet be assigned an original vowel component, and in fact are probably best considered to have been composed simply of a consonant, with any vowel accompaniments arising later in different branches of the family, as a consequence of languagespecific syllable structure rules. Analogy effects may then account for the particular vowel components chosen to satisfy such rules, as for instance can be seen in the development in Southern Cushitic of adjective suffixes in *-at/*-it- and *-ar-/*-ir- (Ehret 1980: 54-58), which follow the pattern of the PAA adjectival marking *-ay-/*-iy-. Table 2 (beginning on the next page) lists the more than 800 occurrences of these thirteen nominal-deriving suffixes that are cited in the extensive etymological dictionary of Chapter 5, thus enabling the reader quickly to locate for critical examination the attestations of each particular suffix. (The table also includes citations for two additional nominalizing processes, via prefixation and by stem-internal change, which are dealt with later in this chapter.)

SUOIIEISANY JO [BUIWON SaxIJV

098 SETYSTD “Lege

“C8S ‘PLS “SES “09F

Ele

xygus

eATIeqIOAEp 686 ‘89€ =—»s-PATINuH ‘THT = ‘SLO ‘“LTOB ‘96S ‘OWI = “ISS “EOS “LOW «=== Pue unou ‘cSH © ‘€67 ‘6€7 ‘BITH ©PATINGUMIE Ay °7 €001

oC. qu], °7

"89L “909

LOY ‘€0r ‘96€ “CLS “EIS ‘SOE “LSE ‘ZZE L96‘7S9 ‘obh ‘S9E “HSE “BET «= ‘909 18S ‘LOh © “067 ‘197 ‘Ghz ‘SII ‘ZII ‘98 “SOE ‘O87 ‘VOTH ‘6IZ ‘OPH

:

‘056 668 ‘LES ‘88L Eli ‘ZOL “€H9 “79S “69S *S6S "C79

== “LET “ELT “6L7 "887 “SEB *LLEB “699 Sar ‘CoS vv9 pur ‘688 “LOE “OEE “OSE “LSE xIPucddy “TIE#*1 ‘SLO xygns xe My = |

‘916 7E6 “798 “O6L SLL ‘OTL “9P9

‘€Or ‘SOb ‘LO’ ‘T0P i178 © =PATRqIOADP USMUA|dUIOD

(986 6£6 ‘8L8 “L18 LLL ‘“SvL *SL9

NCLEavce

‘766 ‘TH6 ‘6L8 ‘9E8 OLi ‘OSL ‘989

“80P “97 ‘VEh ‘OEP

| CIS

"GCS “SES “EOS “09

KOCHOLE

‘€6‘LIZ ‘BE‘861 “LT uendasy‘O91 ‘OT “S#‘L6 == “OTT“69€ “SOT“E67 SHTLIES “EOTH “EvT

€£6 = —‘O9T“67h TPETD “C6#“ELZ

916 69S ‘OPl#

vol8001 SOLAS, ‘Lev# 908 O7¢

unou

“SOL “ESL “OTL “1L9 “rr9 “819 “€6S “LSS “OTS “S6r “IP “Z6E “OSE “OEE “1672 “HHTZ ‘901 “TCH

‘16# “SIT ‘OSE DANRIOOSSE UVECMOLE SESE xljjns “66€ “669 ‘LEB L98

“86L ‘O8L ‘6LL ‘TOL “OSL “VSL ‘CSL SPL “byl ‘60L *689 “SL9 ‘OL9 *799 “LS9 ‘EV9 “LTO “619 ‘IT9 “€09 “966 "88S “18S “O8S “CGS “LYS “TES ‘LIS “10S “96h ‘C8 “PLY “VEY ‘yTy “COV “66E “CLE “PLE “S9E ‘TSE “OSE “LEE ‘OEE “LOE “767 ‘067 ‘987 “6v7 ‘SIZ “LEI “OVI "C8 “EL “6E “OE

“€ 1,

“ELT ‘OEv

‘CSL “8Lb ‘18 l#

“60E# “96€ ‘C69 L69

“S8L ‘Obl “97S “69P “SETH‘OIT “L81# ‘IST

“LOS “68S

7S9 “LE vH

"6LS89L

vO8#

€00I ‘SL6 °966 ‘L78 “OSL “ESL “OS9 ‘19 “66S 96S ‘66S

:7 ajqv

oanoalpe xIJjns

976 6SL ‘Ov9 "88h 986

6S86 “V16 ‘L96 ‘LE8 “V08

Tendasg (panuyuos) saxiffy jounuoy fo suoynisany

wae

“SOT# “ISI ILE

SHIIG

‘66€ “LOE

“LEv#

xtpuaddy :7 LL6#

SATS

€86 SHOW)

“7ST7 ‘66€

sipey)

"eZ A,

‘90S

|G

xyjns

“ESe# “ple T€8

DSd) :-18-y -O1d (aayonp

“07S

SOSH :] XIP -usddy °986 ‘€98 TLE ‘T8E# ‘979 “OSS *L6 “E9#

oansalpe

“OC#*

sanoofpe

x1jjns unou AQNGUIIe Wy, °C

“Y 1x

"$69

CEO# :Z xIpuaddy

pE9 :7 xipuod

LO9# -T XIP

‘vLe#

-dy ‘21001 ‘I9L

‘Srl

“SOT

“SEc# OSS “P88 ‘8LYy ‘18h 796 887‘OLY ‘C67 ‘ILb

SIPEUD

LSTOL# *€ xtpued -dy ‘S19 ‘ply ‘SIV ‘C87 ‘L6L# :1 XID -usddy ‘396 “796 ‘806 ‘SLE ‘SHB ‘TIL ‘699 ‘Ips ‘C8 ‘6Cr ‘LTP 9001 ‘PLE ‘SE “PHT "1G. VEC SELF

‘866 ‘L789 ‘ILb ‘ECE “67S “LLOGH

endAsy

‘IE ‘78 -uoddy

SHTISS

‘60L

fo jpvumon saxiffy (panuyuod) suoypisauy ajqvL :7

“ESO “TES “18H “STE “C67 “ETL#

OATySTD

“108 “208 “P08 ‘808 “LE8 “SS8 “P98 "8L8 “088 988 (068 ‘768 856 ‘V16 976 ‘16 “€S6 ‘856 ‘796 ‘IL6 ‘SL6 ‘9L6 ‘986 ‘L86 “EOOT “900T 8001

_ ‘SOTH “697 9SS ‘e9¢ ‘9¢¢ dy xpuad :€ 61O1# ‘OOl# "CLT “iLtY 'S6b ELL

‘O9T# ‘097 “8TZE “ESP “OzS ‘OVS ‘OOL -€16 ‘¢16 -uaddy XIp-€ CIOI#

DHOUIC)

Wy xyjns “eC

xyje

CC

xIpuaddy *1 “ESST#

‘6Cv

“Cvl# “ILP

“60L‘O18‘SHB “OTC#

86 éSOS

“COL# ‘6€r VIL

‘“vSr “LOC# €68

‘6EV

“ZOL

“6T#

‘O9T “p6# “L061 ‘6ZE “OVE “VS9 ‘798 ‘606 LL6

“89L‘T78‘L76

“éSVL ‘LL6

“LEO*€96

“i798 “ETED

‘OL#H ‘686 ‘LSZ “6E# “CEL “COT ‘097 “E97 ‘p97 ‘Ive “197 “BE “ELE ‘IZS ‘8SE ‘9 ‘69r ‘SLY ‘STS ‘p6L ‘LL8 ‘¢¢6 -uoddy XIP21 ‘OST#H‘PEE ‘9Er “LOS ‘Ors ‘€89 ‘vSL ‘66L

sanoalpe

‘O7I# “El “E8LP “E8b “TIS "87S “iES9 “VS9 "C86 ‘6LI‘C8h

-9{du109 pue SADNGUHG [x “L

xIJjns

“6S7Z “HTP "E8h ‘89S “86S ‘O8L °L68 LOOT éTO0l

“eO uy

‘LZ “LES ‘Ors ‘866 °666 -usddy

‘SEZ ‘687 ‘9b ‘66S “vOS “S19 “C8b “€8P

XIP:T T# “OSS “VSL “L17886

‘ZOL ‘€9L ‘608

‘PST ‘787 “Pr ‘OZS “L766 76

"€96 ‘€S9 ‘OOL ‘T1# “880 ‘Lev

oxTgns

‘OTH ‘6SZ ‘66 ‘EIS

ONTYST)

“ET ‘OZI ‘BE “7#

xa

“9 U, SsAQNInIe unou (7)

suounisay fo jouluonsaxiffy ‘71 SHOU ‘Cyl

60001

Tendasg

SHTOOS

(panuyuos)

SIPETD

ajqvy ‘7

Lec ‘6L1 ‘O8r‘69S‘L76 ‘p9# “ISZ “Eps “S16 “L@r “OSL I8L ‘COT"80S “ODTH

‘186 OOOT

“86 “869 “éSOS 9S7

‘£706 186

SAVE -QI2Aep uduXIJJNS

$6 ‘2606 “6918 “E8SL “8PL “ES9T#

‘EOL ‘6bL “CHL ‘6L9 “679 “LOS “L6V “99€ “SOSH

E16 ‘IS ‘ESL ‘IZL ‘O87 ‘E9I “CCL ‘68 ‘S# , where C2 was a consonant of the pharyngeal-glottal set widely lost in

28

Derivational Morphology in Early Afroasiatic

Chadic) with the addition of an extension — long vowels not normally being present in Ngizim except in such cases (e.g., Ngizim entries in #157 and 299 in Chapter 5, among others). Another kind of indicator is the coexistence in Ngizim of several differently extended forms of the same root (e.g., #594, which attests to the recent or current productivity of six different extensions). Cushitic holds a position intermediate between Semitic-Egyptian and Chadic in its retention of productivity in these old extensions. The two inchoative/denominatives in *y and *w have continued widely to be productive

in the Cushitic branch, along with *m stative/intransitive, *s causative, and *t

durative (see below extensions #34, 37, and 25 respectively).

In addition,

Eastern Cushitic preserves active use of the *dl (> PEC *d') middle voice ex-

tension (#1 below), while some Agaw languages until recently probably had a productive reflex of the PAA *n non-finitive (#17 below). About twenty of the extensions can be identified in the available Chadic data, and at least 28 in Cushitic roots. Omotic attestations of the extensions are harder to discover. The paucity of examples is surely in part an artifact of the lesser availability of Omotic data, but may also reflect the actual historical absence of many of the extensions. The general Afroasiatic causative in *s is well attested in the Omotic branch, as are *t durative and *dl (PO *t') middle

voice, but thus far less than fifteen of the 37 proposed extensions are even tentatively inferable for Omotic. The Omotic grouping, it is argued in Chapter 6, may have split off from the ancestor language of the rest of Afroasiatic while this extensive body of verb extensions was still being elaborated and, as a result, may never have included many of these affixes. The widespread lexicalization of the proposed extensions, above all in Semitic, has a particular consequence for the analysis of the extensions’ meanings: the original connotation of a given extension cannot be expected still to be overtly apparent in each individual case. Once an affix has become lexicalized, the subsequent semantic history of the root to which it has become attached is no longer governed by the former meaning of the affix, and future meaning shifts thus might well take directions that would tend to obscure that meaning. Most such semantic shifts retain something of the characteristics of the action previously denoted by the word concerned. "To Cut into pieces" easily yields "to slice," still an action involving repetitive movement; if the original sense "to cut into pieces" had been derived via the addition of a no longer productive iterative extension, its iterative implication would remain clearly discernable.

Still later, however, the sense "to slice" for

the verb might evolve to "to slash." The fact that the affix originally connoted repeated action would now be obscured. But while some of the individual lexicalizations of any particular extension may become analytically opaque in this way, the overall ensemble of cases normally brings out with fair clarity the original implications of the extension. The verb extensions identifiable for early Afroasiatic, some no longer productive anywhere, have been numbered here according to the sequence established in Ehret 1989. For the consonantal sound correspondences involved, the reader must look ahead to Chapter 5.

Verb Extensions in Early Afroasiatic i dismiddle voice: pPS *¢=Eg.'d

;

29

=PC *dl =PCh*k = PO *d'.

The evidence is consistent across the board in indicating *dl to have been a middle voice extension. Throughout, its occurrences tend to involve the re-

flexive/benefactive nuances of classical "middle voice" constructions or the factitive/inchoative and reflexive/autobenefactive senses visible in its modernday productive Eastern Cushitic reflexes (Hayward 1977). It appears proba-

bly to have been productive in early Chadic, and may possibly still be so in some Chadic groups today. It is also clearly productive today (or has been in the recent past) in some Omotic languages, such as the Gonga subgroup (which includes Mocha, a language repeatedly cited in the data of Chapter 5).

2. *h-amplificative: pPS *h=Eg.h

= PC *h.

- This extension in *h indicated, apparently, an action carried out to a fuller extent than the action of the verb root to which it was attached.

3. *w inchoative/denominative: pPS *w =Eg. w =PC *-Vw- = PCh *-aw- = PO *w; and

4. *y inchoative/denominative: pPS *y = Eg. i = PC *-ey-, *-ay= PCh *-iy-, *-ay-. In pPS both the *w and *y extensions often appeared in verbs that can be explained as subsequently having been shifted from intransitive to transitive application, in which case the presence of *w or *y came to have a pluractional effect on the verb meaning (Ehret 1989). The same effect can be noted in Egyptian in a large number of the instances of its i extension ( *[u] /#C_C[+labial]: */o/->*[u], */e/ -> [i] /#Q)_C; 0/-> 10); */e/ => [1] /#k, xk C (C= Po “ttt, *s°): */o/-> *[u], */e/-->.[i] /#()_C- (but not /#(2)_CC-); and

*/o/ -> *[u], */e/ -> [i] /#C[+labial]_C2 (C2 = ts, ¢’).

In at least one case of stem-final CW, preceding PAA *i also yielded NO *u (#170); whether this was another allophonic outcome of shift #7 or was the consequence of an earlier rule shifting pre-NO *i > *u in some, as yet undefined set of CW environments. In addition, some PAA *a went to NO *o before a [+labial] consonant, at

least in a restricted and as yet only partially defined environment of a PO dental or palatal consonant (see #169, 193, and 463):

#8. *a > *o /#C;_C[+labial]- (observed for C; = PAA *t, *z, or *)).

Three subsequent sound shifts in the evolution of proto-North-Omotic then rephonemicized these originally allophonic distinctions: #9. *CW >C

(deleting all remaining labiovelars);

#10. *e(:) > i(:) /HS_C (S = sibilant; C = *d, *n, or *r); and #11. *V: > V /#C[+velar, -voice]_C, where root had PAA falling word-tone (see discussion of tone below, this chapter);

along with a fourth shift apparently restricted to the Ometo subgroup,

#12. *u > *o, *i > *e /#(2)_C/-continuant, -voice]North Omotic data also indicate an apparent stem-vowel lengthening in one restricted environment (see roots #4, 148, 273, and 842): #13. *u > *u: /#C_P (C = voiced oral stop, P = voiceless labial).

Vowels and Tone in Afroasiatic Reconstruction

60

Another sound change, probably morphophonemic in inspiration and possibly still synchronically productive in some languages, may well be of wide Omotic occurrence, although the examples so far encountered come principally from Gonga and Ometo: #14. *V: > V /#C_C- + -(V)C- suffix.

This rule is in fact probably of ancient projo-Afroasiatic provenance since it can be observed in early Afroasiatic formations (e.g., the root pair #709, 710)

and is postulated for proto-Southern Cushitic (Ehret 1980: 70, 71).

VOWELS IN CHADIC AND BOREAFRASIAN The proto-Chadic vowel correspondences can be explained by a brief sequence of shifts that yield Newman's four-vowel system of PCh, *a, *a, *1, and *u. Note that the first three of these shifts are identified as North Erythraean (PNE) rules, the reasons for which will become clear shortly. A sound shift, represented formally as PNE #1. *Vy, *Vw > *VV /#C__C, is proposed to have intitiated this set of changes (see section on the glides in Chapter 5 for more on this sound shift). Its particular outcomes remain to be fully sorted out. PAA *ay in such environments, at least, apparently became *ee, and PAA *ey probably became *ii, and so both were subsequently further affected by the PNE shifts #2 and 3. The first of these,

#2. PAA *ee > pre-PNE *i, *oo > pre-PNE *u, raised and shortened long mid vowels; it was followed by

#3. pre-PNE *o, *e, *a (< PAA *o, *e, *a) > PNE *4,

pre-PNE *u, *i (created by #2) > PNE *a.

centralizing all short vowels and, in consequence, resituating the formerly distinctive long vowels *1i and *uu in the slots in the vocalic system vacated by short *i and *u. A fourth sound shift, PCh #1. PNE *a [aa], *4 > PCh *a,

La eliminated the last remaining environment of distinctive vowel length in RCh:

What has been treated to this point as simply Chadic vowel history was in all probability, however, a history of wider inclusiveness than Chadic alone. Sound shifts #1-3 appear to have been the common heritage of a North

Vowels in Chadic and Boreafrasian

61

Erythraean branch of the family, consisting of Chadic, Berber, Semitic, and

Egyptian (for which see Chapter 6) — hence the denoting of rules #1-3 as "PNE." The Berber and Semitic vowel systems can each be seen to have passed through a stage in which the PAA system was contracted in a fashion parallel to that described for Chadic — into two vocalic tiers (to use a term coined in Pia 1984, although somewhat differently applied here). The central tier comprised one or both of two central vowels, high central *a (realized probably as [+] or [A]) and low central *4; the peripheral tier consisted of three vowels, *i, *u, and *a, or, if *a had fallen together with *4, of two vowels,

*i and *u. The vowels of the central set would have been articulated with shorter length than those of the peripheral set, but vowel quantity as a formal phonemic distinction would have been lost. Much the same history seems implied for the Egyptian vowels. In the case of Berber, Prasse (1975) reconstructs on solid grounds a

proto-Berber system composed of the full two tiers, five vowels in all — *a and *4a, and *i, *u, and *a — the first two being shorter in quantity than the

latter three. He goes on to conjecture that a pre-proto-Berber system with three vowels *1i, *u, and *a, occurring both long and short, lies behind the proto-Berber array. But that conjecture is not directly implied by the Berber material itself, and reflects the influence of the usual attribution by scholars of such a three-vowel system to proto-Semitic. Ironically, from the work of Diakonov (1970, 1975), it appears that Semitic vocalic history itself requires revision. Diakonov shows that just two vowels, *a and *a, are enough to account for most stem vowel realizations in

Semitic nominal roots of *CVC(C) shape. Apparently, PS #1. PNE *i;*u, *o > PS *a; PNE *a, *4 > PS *a /#C_C(C). The two vowels, *u and *i, remained distinct only in verbal and other mor-

phological operators and in word-final position in adverbials of the shape CV (e.g., roots #209, 310, and 803). Vowel length in Semitic is not required at the phonemic level in this PS reconstruction, and its latter-day occurrence in Semitic languages was in all probability generated anew by morphophonemic and word-structure requirements in PS and its various descendant languages. Diakonov's arguments thus leave us with a four-vowel array in PS, *a, *a, *i, and *u.

The earliest Egyptian vowel system is normally represented as comprising *a, *i, and *u. The original Egyptian vocalism showed, in other words, reductions in the vowel-quality distinctions similar to those in the PCh, Berber, and PS and, as well, the same loss of phonemic vowel quantity (phonemic vowel length did, however, re-evolve in the language, through processes of consonantal merger and loss, by or before the Late Egyptian period). The indications therefore are that the proto-history of the Egyptian vowel system should fit quite well within the broad scheme of North Erythraean vocalic history outlined here.

Vowels and Tone in Afroasiatic Reconstruction

62

BOREAFRASIAN VOCALIC HISTORY

Like all good hypotheses, this proposition provides a set of testable predictions. In the case of Semitic, the posited PNE sound shifts #2-3, together with PS shift #1, predict that in nominals PAA short *i and *u and long *ii, *yu, *ee, and *oo should all be realized as Diakonov's PS *a, and PAA short *o, *e, and *a and long *aa, as his PS *a. In Arabic in particular (following

Diakonov 1975), PS *a should surface as.u in the environment /_C/+labial] and often in the environment /C[+velar]_ but usually as 1 elsewhere, while

PS *a should appear as i /C_CVC, where V is long, and as a in other /C_C(C) environments.

An examination of the Semitic (especially Arabic) vowel reflexes found in old Afroasiatic nouns or adjective roots and in very early derived nominals, as presented in Chapter 5, consistently bears out these expectations. PS *a can be seen in these data to derive specifically from PAA *1 (roots #25, 230, 247, 255, 364, 445, 505, 513, 555, 556, 563, 564, 618, 621, 726, 809, 827, 836, 917, 935, and 985); PAA *u (#133, 146, 366, 669, and 707); PAA *u or *i (#220);

PAA *ii (#59 and 641); PAA *uu (#282, 323, 501, and 515); PAA *ay (> *a /#C_C by PNE rules #1-3: roots #833 (2nd entry) and 984); PAA *ee (#254);

PAA *oo (#422); and PAA *oo or *ee (#28).

PS *a, for its part, is amply attested as deriving variously from

PAA *a (#47, 102, 140, 144, 154, 181, 201, 257, 259, 260, 262, 263, 264, 285, 324, 341, 374, 358, 383, 388, 399, 447, 477, 503, 520, 663, 706, 718, 724, 725, 727, 749, 864, 877, 891, 1000, and 1002);

PAA PAA PAA PAA PAA

*aa (#26, 792, 958, and 974); *a or *aa (#887); *o (roots #29, 105, 243, 244, 318, 327, 550, 806, and 888); *e (#82, 247, 289, 529, and 558); and *o, *e, *a, or *aa (889).

Three modifications of Diakonov's specifications can be proposed from

these examples.

First off, in a restricted environment additional to those he

cites — /#C_{n,r}, where C is a voiceless dental obstruent in Arabic (see

#254, 255, and 501) — PS *a yields Arabic u rather than i. Secondly, PS *a gives Arabic i in /#C_NC_r# word shapes (as in #47). Thirdly, the realization of PS *a as *[u] (Arabic u ) or *[i] (Arabic 7 ) has no consistent rela-

Boreafrasian Vocalic History

63

tion to whether or not that velar derives from a PAA labiovelar or simple velar (cf. #317, 364, 366, and 422 versus 396 and 445). What does govern the appearance of these two alternative outcomes remains unclear (the two South Semitic instances cited here of [i] for *a, in #396 and 445, both are followed

in the same word by a voiceless fricative). These reflexes indicate that the labiovelar distinction was already lost in PS and must therefore have been recreated in Ethiopic Semitic by processes particular to that subgroup. It appears also that two of the noun roots for which Diakonov reconstructs PS stem *a are better posited, on the basis of wider Afroasiatic comparisons and their specific Arabic vowel reflexes, as containing PS *a or, alternatively,

as preserving earlier *a/*a stem-vowel alternances (#827 and 935). This latter suggestion points up a further consideration for PS vowel reconstruction — that seemingly irregular stem-vowel correspondences in nominals may not always trace back to intra-Semitic borrowing, but may at times reflect the PS retention of older Afroasiatic stem-vowel alternances.

Some instances in a

single language of alternative vowel realizations for the same noun may also derive from such phenomena (e.g., the two Arabic vowel reflexes in #121, which correspond to the respective alternate vowels reconstructed for the PAA etymon in question). This reconstruction can similarly be tested against the Egyptian evidence as preserved in the Coptic reflexes of PAA roots. If, as is commonly presumed, Old Egyptian indeed had three vowels /a/, /i/, and /u/, the PNE set of sound shifts #2-3 predict that Eg. /a/ should have merged PAA *a, *aa, *e, and *o, and that Eg. /i/ should derive from PAA *ii and Eg. /u/ from PAA *yu. The hypothesis does not overtly predict the outcome for the posited North Erythraean high central vowel *a, derived from the merger of PAA *1, *u, *oo, and *ee; but in fact, as will emerge below, it clearly fell together

with Eg. /i/. Between Old Egyptian and Coptic times, a complex history of vowel sound change took shape. Early Eg. /a/ evolved usually into Coptic o in closed syllables and 6 in open syllables. Following a nasal consonant in an open syllable, however, Eg. /a/ yielded Coptic 0. Eg. /u/, like /a/, produced o in closed syllables, but its reflex in open syllables was long é. More rarely, Eg. /u/ could also have a Coptic reflex transcribed as i. Eg. /i/, for its part, developed into Coptic a in closed syllables, into i in most open syllables, and into é in open syllables where the following consonants were tn or ty. A fourth direction of change, the deletion of Eg. /i/ altogether, took place in the environment /#C_C>, where C2 was a voiced continuant. Coptic short central ¢ also generally derives from earlier Eg. /i/. (See Vycichl 1990 for these various patterns; the data used here is drawn from the Sahidic dialect of

Coptic, unless otherwise noted). The Coptic vocalic system thus reflects a very long history of complex vocalic change. Nevertheless, even with the complexity of this history, the Coptic and the earlier Egyptian reflexes visibly confirm the predictions already generated by the examination of Chadic and Semitic.

64

Vowels and Tone in Afroasiatic Reconstruction

Firstly, early Egyptian /a/ — yielding Coptic o in closed syllables, 6 in open syllables, and @ following a nasal — combines, as the hypothesis predicts and the roots in Chapter 5 attest, the following reconstructed PAA vowels:

PAA *a (roots #48, 50, 69, 93, 110, 139, 145, 164, 167, 170 (v.), 198, 205, 229, 249, 279, 294, 297, 316, 330, 336, 352, 361, 367, 384, 381, 398, 401, 408» 442, 448, 460, 483, 498, 539, 552, 569, 587, 589, 590 (first alternant), 591, 621, 679, 685, 691, 703, 725, 734, 741, 747-749, 751 (first alternant), 754, 786, 788, 804, 810, 822, 837, 853, 864, 884, 885, 924, 926, 927, 933, 952, 955, 956, 965, 970, 982, 1002, 1004, and 1008); PAA *aa (roots #78, 267, 379, 478, 491, 521, 600, 627 (second alternant), 637, 831, 862, 958, 962, and 1003); PAA “*a or *aa (roots #814 and 989); PAA *e (roots #250, 793, and 890 (first alternant));

PAA *e or *a (root #42); PAA *o (roots #435, 619, 629, 631, and 850); and PAA *o or *e (roots #168 and 178)

In the rarer instances where PAA *a was word-initial and not preceded by any

other consonant, even *?, it yielded the reduced Coptic vowel ¢€ (roots #23, 85, 616, and 924) The second early Egyptian vowel phoneme, /u/— realized in Coptic as o in closed syllables and as é, or in some instances by the grapheme for 7, in open syllables — derives generally from PAA *uu (as shown by roots #282, 323, 546, 553, 595, 697, 713, and 930). Interestingly, in one restricted environment in Coptic, /#{h,x,p}_(y,t)#, Old Eg. /a/ and /i/ also regularly produced é (#51, 374, 382, 399, and 757).

The third vowel, early Egyptian /i/ — changed in Coptic to a in closed syllables, to 7 in most open syllables, to € in open syllables before tn or ty, to © preceding a voiced continuant, and to a reduced central vowel ¢ in certain other contexts — combined the rest of the reconstructed PAA vowels: PAA *i (roots #23, 85, 106, 187, 2407, 280, 301, 396, 405, 470, 479, 485, 538, 581, 611, 621, 683, 689, 698, 720, 725, 751 (second alternant), 808, 827, 918, 939, and 969); PAA *u (roots #36, 146, 162, 240?, 265, 319, 3307, 366, 484, 546; 590 (second alternant), 810, and 866); PAA *i or *u (roots #220, 419, and possibly 1015); PAA *ii (roots #59, 268, 369, 382, 553, 627 (first alternant), 709, 873, and 944); PAA *i. *ee, or *ii (root #616);

Pie

(> *a /#C_C by PNE rules #1-3: roots #260, 309, 983, and ?

Boreafrasian Vocalic History

65

PAA *ey (probably > *i by PNE rules #1-3: root #433, first entry, and root #565); / PAA *ee (roots #95, 537, 778 (second alternant), and 950); and PAA *oo (roots #122, 355, 393, 500, 599, and 799).

Overall, these data show us that the derivation of the presumed Old Egyptian vowels /a/, /i/, and /u/ from the five vowels posited for the North

Erythraean stage can be explained by a single rule change. This shift collapsed the high and low central vowels with equivalent members of the peripheral tier:

Eg. #1. PNE *a > Old Eg. /i/,

PNE *4 > Old Eg. /a/.

Curiously, a reversal of thé usual Coptic reflexes of Old Eg. /a/ and /i/ turns up in two kinds of consonantal environments. Most notably, Old Eg. /i/, when it occurs before a reconstructed laryngeal in the same word, often yields the usual Coptic reflexes of Old Eg. /a/ (notably, roots #275, 433 (second entry), 701, 800, 819, 834, 870, and 922).

An opposite shift affected

Old Eg. /a/, with such /a/ — again in the environment of a following laryngeal — producing the Coptic reflexes expected of Old Eg. /i/ (roots #357, 372, 373, 626, 659, 660, 817, 839, 858, 887, 888, 890 (second entry), 922, 948,

961, 990, and 1007, and possibly 627). The full environment of this dual sound shift has not been definitively established. In general, however, it appears to have switched vowel realizations of Old Eg. /a/ and /i/ before a laryngeal in C2 or C3 position, as long as the preceding word-initial consonant was not a PAA labial, dental, or alveolar-palatal obstruent. This vowel switch also seems not to have operated in triconsonantal words of the LE shape C,_C[+laryngeal]C, in which C; was w, y, or a velar obstruent (see roots

#297, 336, 398, 435, 873, 962, 989, and 1008). The switching of the reflexes of /a/ with those of /i/ appears, too, in two other, rather different and more restricted contexts.

The first of these is the

Coptic environment /#[-contin./+velar]_r; examples include the Coptic reflexes of roots #285, 301, 329, 330, and 419). The second kind of environment in which the vowel reflexes are reversed consists of both /#(_)Ca(_)Cp

and /#Cp_C,, where C, was a dental obstruent (cases of ME s ord are known) and Cp was a non-dental voiceless stop, Old Eg. /a/ again produced

the expected reflexes of /i/, while Old Eg. /i/ and /u/ yielded the usual reflexes of /a/ (see roots #53, 251, 412, 426, 428, 620, and 705). The shared Chadic and Boreafrasian vocalic history, proposed here initially on the basis of the Chadic data, thus proves out very well indeed for Semitic and for Egyptian as well for Chadic. Its applicability to Berber re-

mains to be similarly tested, but the close structural fit of the proto-Berber

vowel system with the expectations of this history gives us every reason to think that the Berber data too will turn out be well accounted for by this scheme.

66

Vowels and Tone in Afroasiatic Reconstruction

To sum up, the systemic distribution of vowels in Chadic, Berber, and Semitic, and Egyptian indicate their having shared in a common proto-period, characterized by the contraction of the PAA system — which had five vowels

occurring both long and short, in effect, a ten-vowel system — into a differ-

ently distributed set of five vowels, consisting of a peripheral tier (*i, *u, and *a) and a central tier (*a and *4). Seen in this light, the sequence of vowel sound shifts #1-3 emerge as rules of a wider proto-North Erythraean provenance (for this subgroup, see Chapter 6),,able to explain the initial stages in the evolution of the vowel systems of proto-Berber, proto-Semitic, and Egyptian, as well as that of proto-Chadic. In PCh an additional sound-shift (PCh #1) would have collapsed *4 with

*a to produce a four-vowel system, composed ‘of two peripheral vowels, *1 and *u, and two central vowels, *a and *a; alternatively the proto-Chadic vowels could be considered to consist of three peripheral vowels, *i, *u, and *a, and one central vowel, *a.

In the proto-history of Semitic, a sweeping reduction of vowel distinctions apparently collapsed North Erythraean *u and *i with *a, and *4 with *z (as per PS rule #1 above). This merger presumably took place in all CVC(C) stems, but its consequences in verb stems are no longer apparent — stem vowels in verbs having been deleted by the developments of PS verb conjugational history. In morphological contexts, i.e., outside of C_C(C) stem environments, *i and *u apparently did not fall together with *a and so remained distinct PS vowels. In the evolution of Egyptian, a less extensive merger of the North Erythraean vowels can be proposed from the evidence. The reflexes, it seems, of two members of the central tier, *a and *4, fell together respectively with those of the peripheral vowels *i and *a, creating a three-vowel array in Old Egyptian. The proto-Berber system revealed in Prasse (1975) suggests that there, in contrast, the five vowels created by PNE rules #1-3 simply persisted, with PAA short *a being realized as a reduced central vowel *4, and PAA *aa, in

the guise of proto-Berber *a, filling the low-vowel slot of the peripheral tier. The assumption that a still earlier three-vowel array underlies that system would seem to be an unexamined transference to Berber of earlier ideas about proto-Semitic vowels, themselves in apparent need of revision. The investigation of the Berber correspondences of the reconstructed PAA vowels will thus provide a further testing ground for the propositions presented here. We can now expand the charting of Afroasiatic vowel sound correspondences, presented above in Table 4, by integrating into a revised Table 5 the findings on North Erythraean vocalic history that emerge from the Semitic and Egyptian data:

Boreafrasian Vocalic History Table 5: PAA

67

Afroasiatic (Afrasian) Vowel Correspondences Revisited PS (in nouns)

PC

PNE

PCh

Pad

Wide

*a

*a

a

*q

a

Sc

se

*g

*q

a

*a

*e (> *i in several environments)

mec mee ce

«9

3

*]

*5

se:

Go

Fee Vin several environments)

pre-Eg.

NOm

*j

*]

*S

Bae)

+41

*ii

*j

*]

*j

x5

we

*o

*o

#4

*q

*a

*a

*o (> *u

at

/_[+labial]) *900

*oo

*5

*5

*j

*9

*0:

au

ru

*a

eS

Si

use

solu

*ygu

*uu

*y

*y

*y

*9

*u:

/_[+labial])

TONE IN PAA Phonemic tone is a widespread feature of Afroasiatic, appearing regularly in the languages of the Omotic, Chadic, and Southern and Eastern Cushitic

divisions of the family. Only the Boreafrasian subgroup (Semitic, Berber, and Egyptian; see Chapter 6 for Afroasiatic subclassification) has entirely deleted tone. Tonal reconstructions have been produced previously for only one division of the Afroasiatic family, Southern Cushitic; proto-Southern Cushitic (PSC) had what has been depicted as three phonemic word-tones: falling, rising, and level (Ehret 1980). A provisional reconstruction of the PAA system can nevertheless be proposed by comparing the PSC patterns with those recorded for the Chadic language Ngizim and for the Omotic languages Mocha, Yem, and Bench, data from which are used extensively in

Vowels and Tone in Afroasiatic Reconstruction

68

Chapter 5 to supplement the relatively sparse comparative evidence from their respective branches.

Ngizim has just two tones, high (/’/ and low (/*/), as do many of the

Chadic languages; and if PCh had the same system, then Ngizim may well largely conserve the tonal situation of PCh. An alternative possibility, however, is that PCh had three tonemes.

Mocha, on the other hand, is given five separate prosodic markings by

Leslau (1959).

One, described as "secondary stress" ([,]), occurs only in a

few polysyllabic words and participates in no minimal pairs in the available data. It surely represents a conditioned development of original low tone ([ °]) and can safely be ignored here. Leslau's "main stress" ([']) and "low level tone" ([_]) are in complementary distribution; the first appearing only on short-vowel syllables and the second only with long vowels. His high tone ([°]) appears nearly always on nouns; and in fact internal reconstruction in

Mocha supports the proposition that at least some [“] were in origin the realization of the complementary pair, ['] and [_], in simple noun derivations

from CVC- verb stems (e.g., root #842 in the comparative data of Chapter 5, and tu:c’- "to take a handful" versus tu:c’c’o "handful," among other exam-

ples in Leslau 1959). Still, by no means all cases of [“] can be explained by such secondary developments. This high tone sign also occurs on terminal vowels where the CVC- stem to which it is attached is unmarked; in these cases the comparative evidence suggests the unmarked stem to have had original low tone. One is thus left with certainly two, and probably three, underlying Mocha prosodic phonemes: a.

f/ f-> l/V ©main’ stress), -> [_]/V: (‘low level tone"), and

b. //-> [], along with

c. // (sometimes deriving, however, < ['] ~[_]). Yem and Bench each have prosodic systems that probably reflect an earlier three-tone system, of high, mid, and low. In Yem the reconstructible root tone was usually retained in words of underlying *CVCC, *CCVC, and *VC

root shapes. In CVCV words built on *CVC root shapes, however, a number of tone shifts, not yet fully understood, took place. Generally speaking, in such roots a reconstructible low tone changed in Yem to mid tone and, except where the second consonant (C2) was a sibilant, the high and mid tones

became low tone. But in words having a sibilant in C2 position, it seems rather that the reconstructible mid and low tones both went to high. In Bench the three-tone system has been expanded to one consisting of five level tones and one rising tone, largely through the reduction of bisyllabic stems (Wedekind 1985, 1990). In simple *CVC verb roots the original three tones — high, mid, and low — are, in general, maintained. They are real-

ized, however, as respectively mid, low-mid, and low tones (coexisting with

Tone in PAA

69

the derived high and very-high tones), and are represented in the published Bench materials by the respective superscript numbers 3, 2, and ! (Wedekind 1990; Breeze 1990). The high and very-high tones of Bench, marked by 4 and 5, and the rising tone, denoted by 2-3, normally reflect the former attachment of segmental morphemes, sometimes still present on the surface (e.g., Bench sums‘ "to name, call out," which adds the PAA causative in *s as a denominative to sum! "name"), but often no longer visible at the segmental level (e.g., Bench k’og 3 "cold" versus k’og 4 "to be cold"). In verbs the high tone (4) may derive from original low, mid, or high tone, but in nouns it seems generally to reflect reconstructed high tone. The very-high tone of Bench (5) is quite rare. In nouns of surface CVC shape, at least, its occurrence can signal the loss of a former medial consonant in a word that originally carried PO high tone (e.g., root #235); it also can occur in CVC nominals derived from verbs (see root #101), again possibly but not certainly where the PO tone was high. ~The sliding tone (2-3) in Bench occurs almost entirely in CVC nominals for which an earlier CVCV shape, with high or sometimes mid tone on the first syllable, can be posited. The usual tonal correspondences among the three North Omotic (NOm) languages, Mocha, Yem, and Bench, appear to be the following:

(1) NOm *’ (high) = Mocha ['] ~ [_] = Yem/*/= Bench 3 (/“/); (2) NOm ** (low) = Mocha /*/= Yem/7/= Bench

1(/°/).

(3) NOm *- (mid) = Mocha /*/ = Yem/*/= Bench 2 (7-/); As noted, Yem has /“/ for NOm ** and * ~, where C) is a sibilant, and apparently preserves the North Omotic values in underlying triconsonantal stems. The reconstructed North Omotic tones in turn show regular correspondences with the PSC and Ngizim (Ng.) tones: (1) PSC * 4 (falling word-tone) = Ng. /“/ = NOm ** (high); (2) PSC *~ (rising word-tone) = Ng. /°/ = NOm *~* (low); and (3) PSC *~ (level word-tone) = Ng. /*/ = NOm *- (mid). (In Ehret 1980, it should be noted, a different representations, /+/,

/“/, and

/~/, were used for the respective PSC falling, level, and rising word-tones.) One notable, and seemingly regular although uncommon, exception to these patterns has been encountered in the Mocha data in the comparative lists of Chapter 5 (#304, 335, and 339): Mocha ['] /#{g,k}oC-# (C =s or t) = NOm *° = PSC ** =Ng./7/. Ngizim, for its part, regularly shifts high tone to low upon addition of an extension of the shape -C or -a to a CVC- verb stem; and some nouns may also be derived by tone shift from verbs (cf. root #880 for an example of noun derivation by tone-raising). The language appears also to reverse tone

Vowels and Tone in Afroasiatic Reconstruction

70

in transitive verbs derived from CVC- intransitive stems by substitution of /a/ for the stem vowel /a/ (e.g., roots #163 and 932, among others). But other-

wise Ngizim consistently shows the correspondences noted. — On the basis of the three correspondence patterns, PAA is proposed here to have had definitely two reconstructible tone phonemes, 1. PAA ** = PSC *+=Ng./7/=NOm Il, and 2. PAA ** = PSC *" =Ng./°/=NOm/7/,

and probably a third,

3. PAA *- = PSC *- = Ng. /°/=NOm/*/. The reconstruction of the PAA etymons of the three correspondence sets as respectively falling, rising, and level word-tone in PAA would yield a neat phonetic accounting for the third pattern, but would also make the PAA re-

construction perhaps suspiciously parallel to that postulated by the writer for PSC (Ehret 1980). An alternative possibility is a system of high, low, and mid tones for the respective PAA markings *4, * and *”, with the syllable rather than the word as the tonemic environment (as can be argued to be the case in Omotic).

A third alternative, that distinctions of stress underlie the

tonal correspondences presented above, is possible but — in view of the weight of the evidence from all over the family, viz. Omotic, Cushitic, and Chadic (see discussion of subclassification in Chapter 6) — much less probable than some kind of tonal system in PAA. Diagnostic evidence for tone in Cushitic generally comes from Southern Cushitic, but in a number of cases the Eastern Cushitic language Arbore (Hayward 1984a) and the Agaw language Awngi (Hetzron 1978) provide the necessary indications. The following regular correspondences of Arbore (Arb.) and Awngi (Aw.) to Southern Cushitic word-tone have been noted:

a. PSC ** = Arb. CVCv, CVVC = Aw. CVC¥, VCV, CVCCVCVCY¥; b. PSC *” = Arb. CVCV, CwwC = Aw. CvCW(C), CVC, VC¥; and probably

c. PSC *" = Aw. CVCVC, where sequences of C and V (or v) represent Arbore and Awngi word and

verb-stem shapes. The Arbore and Awngi tone patterns seen in (a), (b), and

(c) thus correspond to PAA (1) **, (2) **, and (3) *~ respectively.

The data drawn from Southern Cushitic, along with Arbore and Awngi,

from Ngizim, and from Mocha, Yem, and Bench have made it possible for a

considerable fraction of the Afroasiatic roots presented in Chapter 5 to be marked for tone. For many others, however, such marking cannot vere ventured, and these have been left with tone not indicated.

5 PROTO-AFROASIATIC (PROTO-AFRASIAN) CONSONANT RECONSTRUCTIONS

PRESENTING THE DATA FOR PAA RECONSTRUCTION About forty consonant phonemes can be firmly attributed to the protoAfroasiatic (proto-Afrasian) language. And it is the presentation of these consonants that brings this study around finally to the main body of primary evidence: the comparative listing of Afroasiatic roots and their reflexes, referred to repeatedly in previous chapters, on which PAA reconstruction as a whole rests. The consonants come to the fore at this point because the fundamental data — the more than 1000 reconstructed roots presented in this chapter — are grouped for discussion according to their first stem consonants. The order of presentation will follow two patterns. The obstruents, with the exception of the three laterals, are treated in sets defined by shared point of articulation. The comparative data for such consonants are therefore grouped into sets of roots beginning respectively in labial, dental, alveolarpalatal, velar, labiovelar, and pharyngeal-glottal obstruents. Within each of these sets, the listing starts with voiced-stop-initial roots and moves on to roots beginning in, successively, the equivalent voiceless stop, voiced fricative (where present), voiceless fricative, and ejective. The pharyngeal-glottal set differs in having PAA *h rather than an ejective as its glottalic member. The sonorants, and also the non-sonorant laterals, are grouped instead by manner of articulation: there are thus three further sets of roots beginning respectively in nasals, liquids (the laterals and *r), and the glides *w and *y. Within the nasal set, roots are subgrouped by point of articulation, beginning with the labial *m and moving back through the mouth. The liquid set, in contrast, is arranged much like the obstruents, with the voiced member *] opening the set and the voiced lateral affricate, the voiceless lateral fricative,

and the equivalent ejective lateral following; PAA *r rounds out the set. This method of grouping the evidence for the Afroasiatic consonants is heuristically important because a great many of the consonant sound shifts in the family have taken place within such sets, and the majority of the numerous consonant co-occurrence constraints have operated within the same confines. The actual order in which the eight consonant

sets (the velars and

labiovelars are combined as one for heuristic purposes) are discussed is somewhat different from the order in which they have just been defined. The PAA labial, dental, velar, and labiovelar consonants form natural groups of clearly

parallel composition, with equivalent stops, fricatives, and ejectives occurring in each. The alveolar-palatal set, in contrast, may eventually come to be seen as a grouping of convenience. Its voiced and voiceless non-continuants seem

Vol

Tee

Proto-Afroasiatic Consonant Reconstructions

better reconstructible as affricates than as stops. The alveolar-palatals will thus be discussed immediately following the four well-defined obstruent sets, but before the presentation of the set of nasals, because each of the points of

obstruent articulation (including the palatal) can be matched up with either

definitely or provisionally with a PAA nasal stop. The presentation proceeds thereafter to the pharyngeal-glottal (laryngeal), liquid, and glide sets, each of a which lacks a corresponding nasal. Within each main group of roots —.those beginning with the same PAA consonant — the reconstructions are ordéred according to their second root consonant, alphabetized as follows: *b, *c, *c', *d, *dl, *dz, *f, *g, *g¥, *y, *yW

*h, *h, a

*k, *kW

Hr Re mio RS. ee EME

id

*kw"

*]

*4

Mtg, hw, Moe eK

*m,

ene *n; =a

Py zt

ey My ia3) Aas

ane en

awe

roots have identical consonants, they are ordered according to the usual alphabetic sequence of their vowels. If they differ only in vowel quantity, the root with a long vowel is put after the root containing the short one. Because of the overwhelming evidence that Egyptian and Semitic (along with Berber) form a single subgroup of Afroasiatic, called here NorthAfroasiatic (see Chapter 6 below), a cognate was normally allowed into the comparative data only if the root could be found in Cushitic, Chadic, or

Omotic, as well as in Egyptian and/or Semitic. A very, very few items having just Semitic and Egyptian attestations do appear, however, because they provide essential illustration of a rare correspondence or a point in need of clarification (e.g., the actually quite restricted distribution of a root for "two" that has commonly been presumed to be PAA: see root #503 versus #505 and 877). Roots so far known only from Cushitic and Omotic have also only rarely been used here, since the ancient long-term geographical proximity of these two divisions of the family increases the possibility of undetected wordborrowing in such cases. The primary aim in selecting among the welter of possible sources of data was to obtain high-quality evidence from as far back in time as possible for each major subgroup or, failing that, for individual languages that have been well recorded and are especially conservative in phonology — thus the importance of the works by Jungraithmayr and Shimizu (1981) and Newman (1977) on Chadic reconstruction; by Sasse (1979) on proto-Eastern Cushitic, as revised and added to by Arvanites (1990) and Ehret (1991); by Appleyard (1984; n.d.) on Agaw reconstruction (as supplemented in Ehret 1987); and by the writer on Southern Cushitic and proto-Cushitic reconstruction (Ehret 1980, 1987). The provisional and partial reconstructions of Omotic consonants arrived at in Chapter 2 (Table 1), and of vowels in Chapter 4, above similarly provide a comparative perspective on the data cited from languages of that branch. The longest timeline for Semitic comes from the pre-proto-Semitic biconsonantal roots reconstructed in Ehret (1989), supplemented by those to be

found in Appendix 1 of the present work. Many of the classically accepted PS triliteral roots are thus only implicitly present in the comparative data that follow, because they are themselves subsumable as derivatives of the pPS bi-

Presenting the Data for PAA Reconstruction

72

consonantals that are directly cited here or indirectly implied in the citations of Arabic words. Of major importance as a second and direct source of Semitic data is Arabic, which has maintained all but one of the PS consonant distinc-

tions, although changing many of the individual consonant articulations. Definitions of Arabic words are commonly cited here from Steingass 1884, not because his is a particularly good dictionary but because his meticulously detailed glosses are especially useful in the comparative semantic analysis. A third source of value here, because of the conservatism of its rendering of a

number of key PS phonemes, is Modern South Arabian (especially Johnstone 19771981). The views of Alice Faber (1985) on PS sibilant realizations are followed in the representation here of the PS and pPS roots. The only two significant departures from convention are the substitution of *c ([ts]) for traditional PS *s, and of *s for traditional PS *§. The validity of using *s in place *S is nicely backed up by the wider Afroasiatic comparisons, from which it is clear that. PS'*s (erstwhile *§) is 1 the regular PS reflex of PAA *s. In addition, the now standard recognition that the conventional PS *s was a lateral fricative is given overt expression here by representing it as PS *¢. Egyptian materials. are drawn principally from the dictionary of R. O. Faulkner (1964), supplemented by data from Late Egyptian (Lesko 1982-90). The Coptic evidence, useful in establishing vowel correspondences, is taken in all but one instance from the Sahidic dialect (Vycichl 1983; Czerny 1976).

PRESENTING THE ARGUMENTS AND SUPPORTING EVIDENCE In order to encompass within a single volume the amount of argumentation and supporting information required in a reconstruction of this scale, it was necesssary to develop a shorthand format for presenting the historicalcomparative evidence. This format includes a regular sequence of laying out the different kinds of data and supporting materials, and the use of a variety of signs and abbreviations to convey concisely the particular bits of evidence and argument needed for interpreting each data set. A first set of conventions used here governs the transcription of data:

1. Reconstructed Afroasiatic roots normally appear in either of two forms in the tabulations of this chapter: (a) in the case of verb stems, as *-CVC(C)-, *-VC-, or *-CV-, with the initial and terminal hyphens indicating the points of attachment of conjugational and derivational affixes; and (b) as *CVC(C)- or

sometimes *VC- or *C- in nominals, where the hyphen denotes the point of attachment of the PAA suffixed TVs (terminal vowels) as proposed in Chapter 3 above. A few pronominals and markers of position (notably roots #1, 2, 118, 181, 209, 210, 309-311, 470, 482, 568, 571, 608, 609, 803, 924, and 949) manifest the shapes *C-, *-C-, or *CV, presumably reflecting their different syntactical or morphological positioning from those of the verbs and the

74

Proto-Afroasiatic Consonant Reconstructions

more common nominals, as well as their sometimes variable vowel components. 1.1. Two subsidiary conventions followed in the root reconstructions require mention here. One has to do with interpreting the instances where alternate shapes are given for a particular root. If such alternants are linked by the conjunction "or," the evidence is considered insufficient to decide which one is correct. If, on the other hand, the alternant forms are separated by a slash (/), then they represent alternate progunciations that may have existed in the proto-language. 1.2. The second convention involves tonal representations. In general, tone when known is marked on the first vowel of a reconstructed root; in

the case of long vowels represented as VV, the tone mark will reside above the first V. 2. For each comparative-historical data set, the reconstructed root and its meaning are presented in the lefthand column, with, successively, the Semitic, Egyptian, Cushitic, Chadic, and Omotic attestations of the root in the next five columns. Reconstructed roots are usually shown in the Cushitic and Omotic branches as stems of the shapes *CVC(V)C)-, *C-, and *VC-, and

also, in the case of verbs, sometimes as CV-, where the hyphen denotes the stem-final point of attachment of any applicable suffixes. Verb stems cited from particular Cushitic, Chadic, and Omotic languages are given the same kind of representation. 2.1. A few verbs in their Cushitic reconstructed forms appear with hyphens both preceding and following the stem; in these instances, the roots are known, unlike most Cushitic verbs, to have been prefix- rather suffixconjugating. For one particular Eastern Cushitic language, Yaaku, all verbs have conventionally been represented in this fashion and appear so in the data cited from it. 2.2. A rather different use of the hyphen appears in the Chadic reconstructions of Jungraithmayr and Shimizu (1981). In their Chadic root representations, such marking forms part of the root itself, in general denoting a consonant of undetermined reconstruction. Usually such consonants can be derived from one of the PAA laryngeals, *{, *?, *h, or *h, but in some rootinitial instances they indicate the presence of the Afroasiatic attributive prefixes *a-/*i- (as in the their reconstruction of PCh "four" in #94).

The morphological composition of an attestation and its semantic explanation, if not self-evident, are laid out parenthetically, directly following the citation of the reflex (or root) and its meaning.

Where further information on

the phonological derivation of an entry is needed, the parentheses. Cushitic correspondences are Appleyard (1984; n.d.), and Ehret (1980, 1987, spondence patterns, see Table 1, Chapter 2. If no proffered, the particular cited form, whether an structed root, should be understood to preserve

this too is presented within taken from Sasse (1979), 1991); for Omotic corresuch further information is attested word or a reconthe same realization of its

Presenting the Arguments and Supporting Evidence

eS

segments as that reconstructed for its PC, PCh, PS, or PO etymon (as des-

cribed in Chapter 4 for the vowels and in Chapter 5 for the consonants). A second set of conventions is followed in communicating the etymological information about particular reflexes and their roots:

3. Within the parentheses that follow the root or word and its gloss, the first item of information offered — provided more than a simple stem is involved — is an abbreviated morphological analysis of that root or word, following the format:

prefix (if any) + stem + suffix(es). The nominal affixes

and verb extensions are given in their reconstructed PAA shapes, for the patterns of which see Chapter 3 above, and the name of each affix is abbreviated as indicated in the list of Abbreviations. For example, the finitive fortative verb extension in *p' is represented as "*p' fin. fort.," and the non-finitive extension in *n as "*n non-fin." To cite a particular case, proto-Eastern Cushitic (PEC) *bakkal- "morning star" in root #12 is followed by the nota-

tion "(stem + *1 n. suff...)," indicating that the PEC root is composed of a PAA stem, in this case *-bak-, plus a named nominal suffix, here of the

shape *-1- (gemination in this instance is a secondary morphophonemic consequence). The lack of any further identification of this suffix means that its function and form are as discussed in Chapter 3. 3.1. In a few instances (nearly all from Egyptian), where a very common suffix is self-evidently present, the morphological explanation has been left out as superfluous. These cases typically involve occurrences of stem plus PAA *y inchoative/denominative (Egyptian i or y), PAA *t noun suffix, PAA *w deverbative, or Egyptian w plural, wy dual, or ¢ feminine markers.

3.2. Nominals derived from verbs by stem-vowel lengthening are depile depicted differently, by the expression, "n. (or adj.) < v. by V > VV (or V:)."

Also, in a scattering of cases, verbs seem to have been derived from

nominals directly, without overt extensional marking, by simply applying the conjugational system to the stem of the word. Outside of Semitic, examples of this kind are relatively rare. They are identified by the notation "(v. *b /#bV_).

Where two non-identi-

cal labials other than *w compose the first two consonants of a word in Cushitic, Egyptian, Chadic, or Semitic, one must thus postulate the loss of a former intermediate non-labial consonant, the addition of a labial-consonant prefix (e.g., the instrument-agent prefix in *m) to an already labial-initial

Proto-Afroasiatic Consonant Reconstructions

78

stem, or a borrowed or onomatopoetic origin for the word. As part of a more

general development, proto-Semitic extended its labial co-occurrence constraint to forbid even instances of identical labials in the first two consonants of a root. In its more general form, this pPS or PS co-occurence constraint apparently forbade any and all sequences of the same consonant in C, and C2 position in PS (noted as pPS/PS rule #8 in Appendix 4). Only two simple verbs with identical C; and C2 appear in Steingass's dictionary of Arabic, hahh "to stammer" and zazz "to touch lightly on the neck,” the first transparently onomatopoetic and the second perhaps arguably of such an origin; and only a handful of nouns have this shape. What became in Semitic of PAA *C; VC2-

sequences where C; = C2 is a matter for future investigation. Among the five Afroasiatic roots here reconstructed with identical C; and Cz (#354, 355, 489, 708, and 911), just two have identified Semitic reflexes (#708 and 911),

each showing a different kind of dissimilation of the consecutive root consonants. A rather different constraint in Egyptian greatly reduced the frequency of [f] in the language by blocking its occurrence as an initial consonant when the following stem consonant was Eg. r or one of the voiceless fricatives s, h,

h, or f. (Curiously, however, a new f apparently deriving from earlier *p appears in the case of the alternate Egyptian forms fsi ~psi "to cook," root #53.) In such instances PAA *f was replaced with [p] (roots #74, 76, 85, 86, and 89). That this shift is specifically counter-indicated for Eg. § (root #72) is further demonstration of the originally voiced pronunciation of § (and its derivation from PAA *y and *y, as shown below in the section on velar consonants): Eg. #9. PAA

*f > Eg. Dp /#_VC[+continuant, -voice] and /#_Vr.

(Egyptian sound shift rules #2 and 4-8 are given below in the section on PAA dental obstruents below, while rule #3 appears in the section on the velar and labiovelar obstruents.) A further shift collapsing PAA *p' with *p,

#10. PAA *p'> Eg. p, may have preceded a sound change,

#11. *p>Eg.b

/#dl_ (=Eg. /#d_).

But as yet the only known examples of the shift to b involve PAA *p' and not *p (roots #853 and 854)

One notable but as yet rarely attested labial sound shift marks off the Cushitic division of the family. In one environment PAA *b did not yield PC *b, but rather *m:

PAA Labial Obstruents

PCH

PAA

79

b> PC *m /#_ Vn.

The rule thus appears to be a PC phonological innovation (see roots #18 and 19). The same assimilatory sound shift developed independently in certain Chadic languages and apparently in the Gonga subgroup of Omotic (see Mocha entry in root #32), but is not a general feature of either of those divisions of the family.

(w19}s + “WOdUOS -N x f) ,J89q 0},

Bou

YOO ‘YOM

;

wy

0},

-Aod, 10 -A0qy

-GOQx

JIduue,,

Avy JOU O},, ~BQx NQ,

,punols,,

jo UOTsAL,, ImMqq ‘,,SOUOgIeTIOD,, KMqq

JOU ST o19q} uq HI

4noyIIM “30U,,

(i, MeQx)

JAnesou “q Mq

orld,

NVILId ADH ‘SJUANAISGQE [01qD'] WVd

OLLTASNO Gx

,Jeoly}

(pareorput st

“mA /10U ‘Ay/

(,JOAq],,

a

(N) Joxzeur

eq

odAQesou

JIGVHO

Jopudg) ,,0U,,

JILOWO

eqx WON

(29x ‘8861

-d:ng, WON

He) 38249,

askeZ (WON

-¥ng

:poure;dxoun st Ys Ny < Oy) w4SoTy} 0},

08

zAeqg Vv

(ou -UBU JO *Sud} -UI Z + W19}S) WPF “Te2q 01,

-Gx VSW

wAq “UIA “UL,

JLLINYS

wy 4, -hOQ-x “YV

,Jsayo Joddn,, -dog, 10 -Joqx “€

‘T

bQx TC

2204} oq 30u,,

NQx

,ooeyd,,

Gx VVd

“Y2q-x 0}, apeyq ,Jutod 10

“9

3199S

ndJI0A,,

-Yerg,

OFUIS

“1

Apap

“HE 0}, 08

uyq

“UM, (jgns

W9}s) + U,, “UTyse

OSd

-Yeeq. “§NOYs,, ‘AID

-Yeqs

[Teh(CU)

OSd

-uOU (‘woUSp

SMoyyTaq —_,,aJTAOT

‘yIeq ‘Aeq

sys pup -duad 3180 (ysnom) uyq 0},

“ sorpreo(-{pe

-OHH OWePIS ~“[eyxzeq 0}, SUI099q MOTTA ,, WI9}S) + Atqeq Lx oid

[pe -T]Ied

“TeHALQs

“SNnBd

OSd ‘(dx

Dd Sy

oreyeqd -gaayyeg ,,d[PU Q)Ty

OLLT ~ ASNO

-SIUANAISGO x

JYSLq —_, OAd

({'fpe

W9)s) + Zy “jJjyns “A >

0}, 9q

AT E194

‘(‘Jgns

£49 -OS 94}, MOLIO , U Wd)s) + ‘UI, 0},

VWd

0},

,UTeI

:0}9UIO eIAOY -ynq

ROSNe

a6eg (

Ui,

orp

“einy-eq Was +

Jes, Woz)

WON “eqs

JILOWO

£8

v8

SWON

,Sosusug,, ¢ [#9 yous

JILOWO

(bx TA< Fx YOd ‘Opry > Inj ‘a"T

‘WIda}s +°WIOOU0D 4Qurx Gd

-Nix “f) ey,

” (/e/ °3N < xu “T) -D-y YM “INP 14 + WIS)

-queg 3N

,Aq ssed 01,

BENQx OSd

vepry “UTYS,,

“T29x

NVILdADF

urer3 yo rea, —,(¢) ym “(Z) dn, 4q XT [110 x

wlPdIq O},,

CINP J

+ WId}S) ,,dSnoy

SAOUL 0}, “yeTeQx OSd

-TeQx Od

JOM,

19x

01,

OLLIHSNO

J TUIOA

(j; np < ‘yeoyo -UI My + WId)S ‘- mnueg “SN) (1)

JIGVHO

qx -StUANAISGO [0140'T VVd

te

w4SAF,,

(qyns “UI,

a ws)

(UIJ-UOU sy

+ W9}s) ,,(°919

‘s31y) yonyd 0},

({ pueddy)

,,SUuOTe SAOUL ‘paaoold 0},

19x Sdd

2pfuuds

‘JOM O1,,

1T#q Vv

OLLINYS

219%

-THQy

dn,

-$NGx

UTAS,,

Jom

-T8Q-»

01,

-[@Q-%

QAOUI 0},

uC-DUD)

uC) Yeolq 0}, “T9Q-

ub

‘Cainp

(‘sned Sy + wia}s) mod 01, ¢ SI@q yousg :WON

JLLOWO

aed “pring 0}, -UFQ sud

mbx + W218) ,uoseas AuTel,, 16.143q UY :WOS

¢8

(N) ,,POOTG,, CIE Q

“UlQs

(f) .PrINq 01, Ud. “(N) ,2snoy,,

(f) OTe],

UCOy

YOO “YOM JICVHO

lx he UW9}s) an

(JOJO JO “SUDIUT

>,,UTeL,,

[jam ‘suds 0},, -44N9% Od

>, POOTd,,

24mg vlog

-91) ,,UIel SUT]

Jeg, Me3V “((y] =92 SIO Yy + “Susy -ul se ways ‘dnp -[ej jo punos, oov see Jeg JEV -OH yoeatd O},, -uew elag -,9SN0Y,, “URUL,/-U FW»

ued

3x

(Yimig aAIs > “ol

~“Seuew, OSd

a W19}S) Aqeq,,

JLLIHSNO

“Jpns “U My

(Q¢# 90S :u0 -nquyye sAneU -Iayje aqissod

+ ‘yeoyour A, + wia}s) ,duep,, MIEq

uuq

,3039q O},,

NVIId ADH

qx *SJUaNASGO 101GUT VWd

({ pueddy) STOW 3q O},, IQ Sdd

UUgs

uPTAg 0},

uqi Vv

,UOS,,

qUTQ

Jorysnep,,

*(Uegx)

OLLINYS

ontysny [euor}

(-3ng-», pue -JTQ-» SJUBUIOI[E

-Ippe) ,,P201q ‘doas ‘ye9] 01, “JCQ-% “OC

“UTQ-x “6T

,2Snoy £ajeo19 “PIN 0},

,3939q 0},

-U2Q-4/-UIQ-» “ST

“IT

uP[OY

“J3Q-%

01,

“SUdIUT JO

(JdaJJO

1001)7Z#+ lx

uiGSepy

NVILdADH

101GQUT

O1,,

SUTALL *A

‘wousp-Opar (Qsuds

(,,Yseqy,,) T: Ky

‘quue Was

329.193 ‘ -qoid

“ouTYs-H]8 wYseyy‘D)

31g -B, ‘u ‘Joid +

44q

0}, ,SUTUIYST] “30)

“uaIysty

FT 639 01, 2q “IYST] 01, 3y3Tydn ,Y4suG joryA) wajs) + 6, uly (“LOy

Sdd JQx

Q-4/-TITG% “IR 449,

ung “TQ-x 0}, wAT “TZ

O},,

-1y3Lq

“ET

OLLINAS

VWd

JLLIHSNO

-SJUANAISG Gx OC

JIG

0},

01,

>a

(14.

©

‘[pe)

eysoy d

=A‘DA Q4oys

dx
Aq A < “CA yousg SJeq6-7

OJ.TQ

98

eyseUlys

JILOWO

‘WON

US], 4TNOWD+

.,2A18(N) 42q°8N) (- “f)

JIGVHO eJegy

01,,

~FITQ4/-

wGSPTJ

ONS

Teewoog JTq

UpputJTq SUTUIYST,, -

AT 4459 JFOQx “Vv wseg 0}, ‘XT yeu 0}, ‘I9JJO-d1d 0}, ,dSPIS UL ,, W9}S) + UaS ,OAIOS waIs) + Jy (ANP Wy, (‘pud}xd

(6861

,ULYS,, >

W9}S) + I, "U ‘yynsDS -W9S :sonueAjzeo Speoq o19M speur Jo (T[9ys

‘ThYy,, [94S ul

ILTQs

“SEQ

,UIsaq

“SEQx

01, 1135 “JJ

,JUOIJ UI O38 ‘UOY, UL 9q 0}, ~JeRQ-%

“SC

“VC

odiun,,ny

UOjo (Apogq

"OLoj0S

Be I

(TI9yss3o

¥8

(QA0ge

wISIZ

“I@RQx eed, 99S) 77H

0}, 08

OSd 10

youn

b/ “#0 -UBUIOS :son ‘Apuaroue speaq WO -SO

Ix
2(9q)

por

0},

wax -6nJ

0},

Aeiing[Om

aqeqoid ueoy

:wos]uy 6nd JOpPpeyq,, :

JILOIWO

col

LoL

CL

‘OL

,(qnour

0}, MOTQ YiIM) -Yeejsy, OSd

-do ysnory}) poaisem oq

(,a7enOuUAd 0},

AT) .Jeads 01,

(Jgns “U1,

+ SUIMOT} do}s < :soqueulas

yd

,

,Jida1sep pue WUITZUT 9q 0}, YALL -V

UOTTTRIS,,

[ges

"V) ,Stued,,

(jo ino

uv} Joye OfU1

M14 + WHOIS)

,dn pasn aq

‘poystury 9q 01,

“pus ue 0} 9W05

,oeouad 0}, “UN

«(YINOU 9Y)

“OL

O},, MTG YIM)

“Pl

VSIN YJx

,,o10us ‘dnpos) ‘asIBOY 0}, 3q (‘SUd}UT Wd)S
Od

~LPYO,9x

&

JILTASNO

/S1UaNSGO dy C

DAd

NVILdADH

1D1QUT

:

YeoIq JJO

‘t1q

Jig

youtd JoB

0},

‘3N -peg

JIGVHO

#

,Poom dx

Od

>

Dx

(JgNS “UI, + w9}s) Soong, yd AT ‘(nyed *D) —_—,,SJoyenbpury,, Amyd

(dy

,2PIs, -Y IQ, 10 -Y Td,

NVLIdADa dy /SJUansSqC [VIQUT VWd

OLLINSNO

CA) ,sups Aq ye2Iq 01, -¥%,Q% CA ‘OF

-4%,Qx OAd

ue ns 0}, -Tue dd) ,A1ds 01,,

-(@) Atg 3N

Curly

(D429

[x + Wla}s ‘f) eo 0}, :

JIGVHO

CSNJJIP I,

+ UWI9}S) ,dAvOTO

01, ‘ajds

bbeq ‘Vv

sbeq

“ads 01,

(UDA Fy,

+ Wd}s) ,[oured B Jo Joy} dU} UISOO] 0}, SYeq ¢,sodotd OUT

JILLINAS

“WV YY!Q 01, ‘1ed} yea

Cll

uC) yds 01,

-41,d-4/-,42,d-%

iC)

“SOT

-¥¥,d-% “LOT

dn yeaiq 0},

UPL,

1d, ‘OOT -Y

.SUTITY

-4ed-, ye2Iq Aq

OT]

01,

-Ind, ulI@yS,,

“601

yy

W9}S) +

“SUSJU JOT (jdaqyo

yonS ,

0}, 3e9q YIM B

bseq'y

JLLINAS

0}, aq

*(sJuRye SdTIS q

Jo 6g#UO

SMOYS

QOUSNTJUL

(OLT#

PIOM > °68# IN }S] 9SUAS SMOYS

souapuo -[esIp d SMOT SUIALIO ‘D p

“S309

‘oprarp‘ads

-UIJ-UOU

‘sod 3god

,,(pearq)W91S)

bseyy = ([3]

Iejy ‘q

>

827g(N) -T1929 O01,°3N) 7

»9STO (f) LIO}

JIGVHO

339 :, ‘dnpai se ‘bay -uewlas :son -duiosap 9ay)JO $,JOTYD SOTO 38 -yedIq 3m ino Jo (339

yeyINO JO

-bnyeqegyCF

‘paddiys01,,

Od (dx

‘Og reyy

dx>

a}eredas-W109) dy,
‘u (‘A

Wy, ‘pud} < xa (‘WO0F 4NQ) Ww

Asd

0}, [OJ ,(2U0

“40,d%

JILTTHSNO

/S1Uan idx4SqO

OSd

NVILdXDH

1DIQUT

‘sosepurg,, 0}, oyeur ‘odor ‘sdins ,2dor ,SULYS Wd)s) + Ay,

*SONUBUIS

S[SUIS uondasap SUIOSUO (ssaoo1d

“Vv wsJeg AY 0}, ISIA\} & odol

aT]ST & ‘108 SI UR

SUTT

VV 4RbqQ 01, “OT wT] W9)S) + ale

OLLINGS:

Vd



Afroasiatic (Afrasian) Dental Correspondences

PAA

PS

Eg.

RC

PCh

PO

*q

*q

d

*q

*q

*q

*{

*¢

t

*t

*t

*t

£7.

*7,

Z

*7

¥*q

*7

*5

*s

S

*s

J: *s3/s N: *g

¥s

ad

aCe)

S

ts

J: *s'/s N: *s

Sts,

_ A consonant co-occurrence constraint of far-reaching scope affected the distribution of the dentals in Semitic and Egyptian. Arabic (per Steingass

PAA Dental Obstruents

123

1978 [1884]) shows no root sequences of *td, *tt, *ts (< PAA *tVs, *tVs',

and *tV§), *tz, *tz (< *tVc’), *td (< *tVj), *ts (< *tVtl'), or *td (< *tVdl). The regularly corresponding equivalent sequences in Egyptian td, tt, tz (= Arabic tz, t0), ts (= Arabic ts,tz), tt (= Arabic ts ), and td (= Arabic td )

— are likewise missing from that language (as attested in Faulkner 1964). This constraint seems probably to have affected Berber as well. But several such sequences, forbidden in Egyptian and Semitic, certainly are attested in both Cushitic and Chadic, notably *tVt, *tVs, *tVz, *tVj, and *tVtl' (Ehret

1980 and 1987; Jungraithmayr and Shimizu 1981). The evidence is not yet sufficient to determine fully the situation of PO, but at least some of the se-

quences are clearly present in Omotic, too. The constraint against the co-occurrence of PAA *t with a range of other dental obstruents, including the two

laterals *tl' and *dl, thus shows itself to have been a shared innovation of Egyptian and Semitic and presumably also Berber, but was not present in the prehistory of Cushitic, Chadic, or Qmotic. Because the specific effects of this constraint are not yet known, it can be given only a generalized formulation: PBA #3. PAA *t > ?/#_VC (C = dental obstruent). A second fundamental constraint disallowed virtually all sibilant sequences in Semitic and, from all appearances, ruled out the equivalent pairings of consonants in Egyptian. From Arabic and other Semitic evidence, it can be concluded that it was forbidden in proto-Semitic roots to have both C; and C> selected from the set {9, z, s, c, 3, 9}, i.e., for C)C2 to both be frica-

tives (or affricates) derived from the PAA dental and alveolar-palatal serices. Specifically, the sequences *66, *6z, *Os, *Oc, *05, *69, *sO, *sz, *ss, *sc, * Px OSd) ‘SSOUYOIU},,

JIGVHO

NVILdADA ‘“SluandsqCQ [DIUIG VVd

Its

0},

({ pusddy)

,UMOp

( sqip'y) (‘dutoo

wia}s) ,,Aduoy,,

‘(Jgns ‘UI, + WI9}S) ,,S99q jO UURMS,,

Aqip ‘4gep “VY

JILINAS

0},

“9BP-% “ECT

i MOTF

“ICT

dt YONS 0}, -qOop-x “@ZI

Aquoy dary 0}, -Q IP-x/-G2P-%

SqPx

‘

+ WaIS) ,,[e},, BUeG-OP » euueg ‘wos

JILOWO

Px

Sdd OPx

"QIDAOP Sy +

-UI ‘"JyNs “U Uy

9Cl

qP

“OZT

“SZ

“PZT

(199JJO

0},

yea“

-mbep-~

,n0ge

0},

,3U0Te

qe

ffep-y w19}S) me Ix

Te -JIp

A[qeqoid $1, (‘sngjip

+ “snj Wa}s

OF,

‘¢6p sbp

,9pry pnBnp

Teewloos

(AA

Dd)Px< Px ~#/

0}, 3q

OSd

OSd

.3UI03

-m¥ePx

S-uiogCA/

-HEP,, OETA WdIs) + |, ‘(pe syns > Od --PIMPs Ax
“"U‘A + “WI09U09 -UT,)

TRRWOOS “3N

JILIHSNO

-SJUaNdSgqOC Px

Jrip Jhnp

:q

NVILdAD8

101U2G

0}, dup , ‘dnpos)wia}s se (‘boxy

JPJp

-Bnp-» Sad 6px 16p 0}, ‘QIndsqo 0}, ‘I9A09 ind 0}, YSOTQ WOry Ul uSSOUxTep 10149) (6861 WANOTA

“sud}JO

WHOIS) + Yq. “UL

/-JeP-x/-FIP “vy -x byep -jnp-, 0}, nod ‘Ino 01, ‘yeay ‘daas aq painod ‘Ino UNI MOTZ ,20AO

JLLINAS

VWWd

,Jeoddestp (poure;dxoun

yisuay-A)

“Bop

PUTTQ,, >,

eyoojyoBop

JLLOWO

Lo!

oAeD

0},

“Yep

(‘ulj-uou

Uy + WIS)

,AdAUOD 0}, uyp HT]

-AATP NVIId ADA

(inp mb + ‘JO SB Wy + Wa}s) , Sdays

1oYs ystA AT

-1JIMS WRM 0}, fewyep ‘(Oj se

"pua}xd Wy + was) ,,Ajparsed -X9UN 3UI09 01,

myep WY

.SUTY} B JO

J» + WIS)

(sont

‘dnpal) ,,oqis quney SATQIST] -JoyuTuN SuTY} B JIS Ol, AvepAuep “Vv

(‘SUdJUT SB UA}S

Yyonu

O4PI

‘KEMP

“Px

-mAIPx

ul [93 O1,,

,AeMP PUSS

‘QAOUIAI O},, OILTASNO

O},, 348}

:

( -zp *3N) (N) ,,03 01,

JIGVHO

OLLINAS

jkep

-STyep ,,dSBYd WI9}S) O},, + Sy ("‘sned

UV -WOS

JILOWO

Dy ‘SJuanusqoc [D1Uaq VWWd

Vv

8cl

uu)

QAOUI 0}, “YePp-x “6CI

O},,

-mATp-» “871

ull}

APMP O¥P} 0},

-Atp-x “LL

“CEL

“TET

“OT

“YepAUS wu

+

De

(*90J

punossUtM ay) 1997 WId}$)

Vv syep 0}, Jv2q9]

“ISOYO

Oy}

yoouysuo prey SuIy) ysulese

uyep YOU UO

0},

‘3ZI “O[QUINIO yeaq JeTy +,

,9U0Id puaddy (| )

ADH

0}, 9q

AT yp

“VV 47Zep 0}, ‘PpuULIS ‘punod -Jaatnd

0}, dT

-Yogp-,/-Yaa Sdd YPx p-,

“ATP=% 0}, ,punod

01, 1] ‘ITZ9q MO]

0},

OLLINYS

WWd

MOT

NVILd

JolUuag

‘[Yy] -was :Sonue-MO]

TReWOOs ,,asqnd, JO),

(pooyTq

-OA Bip

SULA] (vore

x) =

JIGVHO

“JEP 0},, punod ur) ,,(@euo (N)ur

4 ‘ured

Teewoog

xoop oO ‘AaqTeA,,

:Oq

‘

“Yeps wIOUZ

DSd)Px > Od Px C#/

OSd 0},

JILTIHSNO

‘SJuana Px sqgc

Uu,,

(‘ulj-uou

-WON yousgq Uyap¢‘z 0}, ITY W9}S) +

0}, IIS Pur .JSOI was) + -egn{uod [euon (Jyns

‘Ip nnp

0, ,,[[EJ W191) + Sy. “('Sn’d WON WX

WOS -S:0P,

JILOWO

6ZI

mm;)) Px Jd > Px

OSd) ,S2001d 0}

4p AT

dn yqoey 01,

,surpunod

&q dn yeorq

‘punod 0}, HIP x

Suny

Aq saoard

0} Yeazq 04,

- m{NP-» IO “ANP-»

“JJOSUT 0},

,OJUL YOUS

‘ ANOTZ,, 4p

[PACD 0},

(‘UIJ-UOU S$, + wid}s) ,,punois oY) OUT oye. -guod ‘ysno1y}

udyOIq 3q 0}, “ANP* OSd

(#/Px OSd< Px Dd) 20"|d

‘pued (1

dey, -aTNNP x

“CET

-FOP-x “PET

UL x1 “Jueyd 0},

-dy) ,dn deoy Oo} “TTY MOT, yynp “V) dey,

snbnp “vy

(‘WOy “puaixe Zp, + WId}s) , (ed

JLLINAS

~4OP% OSd

-[NA) yong o},,

NVILdADd

[oIuaq

‘sluansqo Px

OILLIHSNO

JIGVHO

-ZBep “3N

WVd

,

(Px Ud < Px OSd < Px Od) udesy ysiqqns,,

ui9}s) ,,punod

(‘pus}xo Wy +

Px OSd * #/

Pk Od > Px Sd) ,uossiod JOIUIS “I9pP{9,, -T TP» IO -T#Px OSd JILTHSNO

(duos

,puus ‘punod 0}, Ssep “V

(dn {nd ‘-a't

“mp < ‘yeoys

,J82q 01, “FIP-4/-FNP-~% “SET

dae] MOIS ‘QSBIIDUT 0}, “T3P-% 9CT

({xIgUr “ -n-, -UI My + W9)S) ,oAOWdI

‘dn I9A09 0},

‘VST ~XBP-

-XED- CST

0},

| Joplours “MO] UNG O},,

wes

“EST “MED x

“8ST

LL

O},,

0},

lB

“SUP-%

ytOSUT

TOMhEee

,,0}

WWd

£P 0},

WId}S) + My ‘Jeoyout < “CInp

Mj5ep 0}, [Ted “pnoye {yes ‘uodn[feo

yous ,,oyut -dy) pued ([

0}, ind ‘oyur

Sad $P%

OLLINAS

ep BPs 5 0}, [[B90} 2q uo s,du0 ,prens ‘dnpas)ways se

V

(‘sudj}ur

x

0},

SUP ~

OSU

OIGVHO ~~

+ Sy

Sx < (SBN

“UOTOd

‘3N “Seep 0}, dojs‘dn 3njd,dn ways)

u(ZATINJODI09)

yeads

|

JILIHSNO

‘Sluanasqg Px

x “SRP ,aejndos 0}, {19}U9 0}

NVILdADH

jDIUag

OLLO ~

| ba

JOY

Vv qvigei

-QI4-% JO -QTEI-» 0}, ‘OSI MOIS ,

0}, 2q

SUIOIS

(inp

£1

+

"JEOYSUj1T

dA0QP) Joy,




0},

DUG, 70] ISABZ, o[-eZ

YOIeIO < S “OILIM 0, 10

OS UM -FOPx = wOHIM 0}, AAD DSd = :SoryUBUAS) YM)O%>

2105 YOM(Sx

AOD YOM SPx

0}, ,PULS “f) Wd Od OSd) + My ‘JeOYSUI < —#/Zxy snp‘2D (os

MOT] ‘IQAO ,uonepunut WIS) + ‘UI,

ITUS ULI *,

Xxez-

[(-SOZ,, 10

~SOZ-% -UBZ-» “SOC JO

9U1099q

,Peyousrp

AXDZ

01,

Vy

JUTeNSUOD OUST ISUTeSB -JaSUOd syueyIqis sAnN Ul sues (JOO

O},,

+

[Px

-UBUI9S)

“AU

:,,WIMSUWId}S

O10WQ'S] -,PPxLPx

uP[O9,,

:010WQ oJayseg yeZz .

dq

-XBBZ-»

“was

-INd90-09

Su0s YOM

poloodxa) :,jur0d,,

“POT

JO asned

*37

(AA < A 4q

(Aud

,UTeyUNOUL,, “(f)

ueoy

[Da > uROT

+

"A >u ‘[s] Ys

oe AULT “OFTUY,, Mord

99S “y

2UOS YOM -SPx

>

“Idi< Sanp‘)

-Zpeb-%

0}, O03 ynd

-pnnb-,/-p1B-» — Ipefy

OLLINAS

ADIaApuv

“ZLT

0}, O38

,UMOP

jjel-yv

ploy ULI

MOT, ‘PUNOIS ured ‘(‘dnpaz) Ajel O},, AMOIY} ,UMOp W9}S) +

jeljel-y

1sy1e3 uv daoy ,JoyIas30} -O1) ‘dnpse (‘boaj

0},

Jelzef

0}, 199710pu 9 OIL} ,ABAB :,

Feb-4/-JN6-%

=e, 1, ‘ploy 2481 ployJO

du090q Ind (Mo]

0}, UIN} ,20A0 WI9}S) + My, -UI ‘yeoyo < (‘inp

VSI M6.

30 Moy
Od -Ax “ot9Y

ory Aq

O},

BYOO|“Soy,

S07

JILOWO

m Ax

‘OVE

Jeu 0}, @ ,Qstou pnoy

(10g “uty (dy + wajs) ,,ATpeq

JNO QUWOD ‘T[aM

-gee my “3N

,AOUPLY,,

Nx HN < NAx VVd ‘Sistys Je] -n3al a]qeqoid ‘C6# L861 yIYY $991 -109)

x“Tem que Ul x /-Tem

01,

“Tex

,punod

(CAA

({'sua}

({susqut

01,

_ se ways ‘dnpod) sU099q

Yyqy

uP[O

-YeeyAx OSd

"Uu) ,SSoupely,,

NVILdADY

(DPx

“ Aoupry,,

(a0ejins

SULYLYS MOI -reu A[OWaDX9 sey diy :son -UBUWas S'dU09 Lx + Was)

cley V

diy Feoq 01,

"A SINS “U [x

wiajs) ,a[doad oy) Wey pue d[HAIJUI 9q 0} ‘rea O[DIDJUL,,

Max “WWd JO SJUuansqC ADJaAO1qVT puv svjaA

JILIHSNO

(u>

[onrysnD UId}Se_ > SULOT ayeredas :,,Aou “PP, 2137 UV

JILOWO

vIAOY 019]

TWIOS $ a77Bley

907

OLLINAS

Vv yey

,AOUPTY,,

-YeCm-x “Te

“CVE

“TeaUlx “EVE

oF

-[PmA-x

,punod

JNO UWIOM 9uI099q 01},

IO -Y@mA-x

“PPE I-x “FOC 0}, O38 ACME

“SVE

Vv

AD9MAx ‘ysnq,,

gsey

(‘jyns

,UTe]UNOUL,,

WIA.

(O%E#

‘dnp (wiajs

Ysniq,({) -o1)

x

‘puaddy (p

poresjoun,

poreopoun ,Pukey] 01,

JIGVHO

voeuUS f)

,odOl

-X9

® Gy, STA} +

0},

W9}s)

‘ysniq

IIux

0}, STA} JUL) uC “OVE

(‘puar

Was + py

(‘INP

wia}s + -qeqold AT Ty “U SJjns ayeura}ye:adeys

,,

~YTITMAx 20s YOM ~YTTIT Dx PLIx/I Dut) uC

Wa}s) + U» (‘jdwie

un}

OSd IO Ax

‘pury ysng

“T9987

OSd MAx “FEE 0}, ‘03 p[OARD

sas

YOLyM

JILTAS

‘jyns ‘uU

0},

Jper ‘Vv

JLLINAS

‘IQA09

"u I, + Ws)

y[T9A

[VVd {0 SIUaNAISGO AD]AAOIQVT puDd ADIaAA

0},,

-JeA-~ “ESE

JO JOY UI IO dAoge 03 ‘jo do} uo 115 01,

Nd

-qgA-,/-qnA-% “TSE

AIABD [e0,,

0¥Y4:04 BYO||

[Da > : UTS,

06:06 wax]

uvo|

JILOWO

JICVHO

1B e6 3N

‘u Ay + W9}S) ,punois 3} ‘THOS “YLea,,

(jyns

Ay

(‘ulj-uoU Uy + Wd}s) ,, puBISIOp -un ‘MOU 01, -uyerd,

W9}S) ,, SUS -dINseoul,,

aes “(FD saNp

w9}s) ,,QUoT} -ed) sutuexe ‘aINSBOUI O1,,

1¢9 “Tue JO) 9pry,, meh

(TU “yer,

*“yoou,,

(xex *D)

‘UDYS,,

-hok,

(4) SHBPULSP,, 99 AT ‘pul “ysny,,

({ pusddy)

OLLINAS

46, Sad

, SUTUIB Xo O},,

SWWd 0 S1uaNnsISGO AD]IAOIQUT Pud ADI9A

NVILd ADH

BOY)

[aoInos SHOWIC peynusprun >

,Apoq ‘ysoyy ‘ULYS,,

(jd ‘3q+ wm

‘yeoyout Ay 6 — -Ipow) sadreyo = (TB, s-Jyns “u “SIP, U7 EY [4 + Wels) ,3eq Ad) .MOTZ 01, -ye, 6,

/-4n,b, Odd

JLLIHSNO

,ouly-quinyd,,

:

JIGCVHO

AvIJ2Apuv

lx

WON < !Xx Od

se eto

“+ WON < "Xx Od) AIP,

coe

JILLOWO

(‘qIOA GUC

“ZOE

“19

-WIA-%

-ped-,

,[QjzaMod

CAU

“HA-»

35,.

(ued

+ Waals) ,, Geis

IP + WIS) ,ayenoued 0},

(‘ulj-uou

pawns “y

Uy, + WId}s)

ty Feaq 0},

“(Q99]J9

US eh

bgek -v

OLLINYS

3q 0}, ‘ZUONS

(Ax 0d >

JO “IaqeM) ayenoued 0},

9 uy =—_--X9. q,, +_ W198)

= (soy ‘dD Spual

By Od

(f) ,(poom) = Bx DAd) ,,O1r Nd ‘dOI]S 0}, dAIBO 0},

WB, YQ eos =~

(ay

0},

989 AT

+Wd)s) ‘UI,

NVILdADH

€ jug

—_,({) souedoL,,

(Cygnus

s9sd aT

‘(boy = ‘(qdayJa Jo *suay -ul 4, + ways) sewajs‘dnp -O1) ,21qqQnI,, _,,18aq ‘punod 01,

,ovepanu

poyeiodo ine] -qe dy ox Ix plo

Ta < Fa #/ Te

wyiguens,

JILIHSNO

-J1OX410 -JJPA»

~JEYx/-TEAx MEBY

“Mesy < Ay Od) «> ,S2d0td 0} 1e9} 10 yeolg “yeolg 0},

JIGVHO

yeaq

‘JOT{S 0}, yN9OUT

01, yeoq ‘dn uC)

“O9E

+ sy 0},

LLC

uu9}S)

(‘sneo tds

-SBIeY

wax :WON

JLILOWO

Ay /WWd fo SJuansisqQ 40] ano1quy] pun aja,

uede

“¢g¢

“p9¢

NVILdADd OLLIHSNO

snjaao1qvq Sluanu fo sqg /VW¥dAx

0},

,SUBTOISNUL, (jd

surpunod (ureis

,2eWOuJO}) Ur,,

suWIOs UD_q ub,

JIGVHO

(,,208J

pug fund DAd6, “ut “WOM YOO 0}, pusq AIM, ‘sooey pub, 0}, 1ST} ‘9dOl) -(poom) ISI] SsovuNI ,, sW3}S) ‘SuLys (019 0}, 30 .(2d0I) + mb, SInp >U Sad)bx> Od ‘f) Uld}S 1943930)-MOJ) + Z» pouinsaiJoljres d 19 ,,(SUId}SWI9}s) ‘Ay Od Oy< *SUudqU JO T -UBUL + Py (INP °A 0}, JSIM} OddUx/ (# (10u

“GtA-, vy

STM} uC)

-UIA-, 10

urajs) + m

Mug

VSN -un&-, ug Jo /-uvd, Auk, Dadby /-ue -Gnd-,/-Ged-,, 0},, SUIS ,, -W9S) o},, dejoJo) 6, -un 01, JY “3q) ‘sou BIA -IOJUI 01, IY .OHDS “Wag -BAOUUI 31eIPOUI SUIULOUI SunaSl0} Dad)6»> :U01} -MOLIU IJTYS UdaS Osye Od ‘Ax Od Gy *(O1snu SurJo rug SuluesulUl Sq Wa}s + < OddUx/ (#0} Suyeagq ou A, ‘yeoyoul < 0}, Avyd ,o1snur “inpse Ur puc 0} (OIsnu UWI9IS) + Ay "1x9 ‘34 (Aqua se ‘(np

OLLINAS

ADJApup

JILLOWO

VE

“99¢

“COE

,JTey W9}s) +

Jegued

01, aaey yor—-

"Vv

(yyns

J-eqgnpund UdTJOMS,,SPURS Jo ,,JBOIy}Wa}S) + q,y. ‘wtue-oid pouns SUIDIOA “WIsse UI Y10q ‘satnua Aq yrys ‘A

(‘snj -JIp 1, + Ways

(,189q

(z'puaddy)

MOT

36x Sad

({ pueddy) ,dn yord 0},

‘336,)

-[9q ‘MO 0},

‘U Ix F

WId}S) ,,d8RI 0},

Igely (@g°D

24

JO) 3SU 0}, T5Q ‘TEU.

+ W9}s) (uns

< ‘yeoyout Ky

SSN Ul Se “QUI

“jyns

Jo

,AdAU0D 0},

“Seva,

,2stou pno]

BOYCUI O},,

Px Fl

=-gnnA,/-yevd,

(np

-peeb 3N

Ways) eq 0},

*(nuUb
iO

,UMOp

+ Wa}s :, daojse

e

pup0} (dn Aeme dn Suryord

0},

-deeyr-»

puog wed) .(Apoq

Med

0}, 2q UO 3] yurod Jo Suey ‘doojse Treg

“Vv O10F So UU, Fy UWIdIS) +“UDA SOURS -UN Mpdff

wos) ,yooung,,

uPusq O},, -deemar,

+ Py SIND -W9S :sonue WO -x9 pepus} SuIpusq Jo syoorng uf Sunits 10 -yenbs

‘( gdog

“6LE




jo oytd,,

Jey}

goo6b

YOO dyx ‘UDA oW0s imAx

(T8c#

YOM,Wx>

OFLU, JOO *

jo

DAU

Cur)

punos s)JIys YCEC#H (aaoqe

UZOSUT

Joyita Od By

[Py

‘ety, asneooq JO

>

+

0}, wyds ‘-1,dey,) wia}s

ey -4,d Byoo|

,Pou

*c) SIeopoun

-devma-y $, UWIdIS) +“USA “IR}IUIIS,,, SOBs -UN(rego ,oxe-apeq

sdq

01, IIs uC)

NVILdADH

sdq ‘UOWIUOITAUS

TTeeW0og Dy

OLLIASND ~

-dyma-x IO

dub,

“10}) (2A0ge P7E# 998

,deads,,

o4,, ULA sty}

‘WON W2X Sey ,,PIIG,, -URUIRS)

JILOWO

ECC

Woz ‘py Zey

»

puyy

(f) ,2024},

,]eus,,

[aw3y, JO

WY» “—D pue

plnom “uIsse

pwy , AW PIetA VVd [eursu0

uy < Wy Od ‘Yd WM “Wy

SUTUINSSE UAD ‘QouR]quuas -91 souRYyd JUS -redde :,,201U},,

(au wos

‘D) 19] [(#Y A “Wog
==—_--op Ay, + Was) ueo]) , ep e, —_,Uesuu ‘ayquuny

STY

~——-Heewioog :q]

Sx

YOO ‘YOM

JLLTHS NO

Xx /WW¥d fo SUaNndSqC ADIaAOIqUT PuD ADIAA

JIGVHO

(‘puslxe
“SUdJUt

(IE Wx odeys was WWd ‘Inp J, + Was) uP TGQtu O},, - IN Ax OH ‘OW ‘Jasuny,, JB wx OSd

JLLIHSf0

Csngztp

Iy + WIS) ,,2]

-youy ‘sdosp ut MOTJ JO T[PF 0}, 33¢b ‘(boy se ways ‘dnp

jebieb vy

-O1) ,,UTeI O},,

(6861 19248)

+ LE pH 1001)

(‘SUdJUT “INP I~

- 88 mA-x

,UMOP 19M O},,

dig & Oye} 0},

JO wI-4/-FL ae

aig & O¥eI 0},

,MOT[EAS IO

I4x Sad

JO MOTTEMS JO

EXEF AT

OLLINYS

(‘bay se ‘dnpal) ,,J89 01,

NVILLd ADA

“WWd {0 Sluanssqo 40]aA01qvy] pun ADA

“ISP

“OST

0},

“ESY Ax

De

YOO],J®

“AACR

(‘90}

,SULYULIp

‘punol

Csny

WIdIs + 1, -JIp

194M“T¥§)

-KIQA9

Yoo]

Jo1a¥4

W215) + Ja. (ION

‘Sued

90Ud3|NPUT,, UI

:SonURUaS -[TPMS *({dn jseb

youanb ,JSity Wd)s) + 3, ‘ued

YOO]

0},

YH

MOTEMS

gseb-y

SOOM “MOYO

Moy

,MOTTRAS 0}

1001) p# LE +

0},

“TSP mI-* “DER

OLLINAS

ADIJIAAPud

YOO,J@

\*

MA

“AVY.mIx

JILTHS

OSd (VVy

Dd) Ave,
*dz, PAA *ts, *c > *ts.

The reflexes of these consonants in all but Semitic favor an alveolar rather palatal result from the merger — thus the formulation of the rule given here. (The notation "Eryth." refers to a proposed primary branch of Afroasiatic, Erythraean, composed of Semitic, Egyptian, Berber, Chadic, and Cushitic, coordinate with a second branch comprising Omotic alone, for which see Chapter 6.) These arguments require a distinctly different phonological prehistory for the PAA affricates in Omotic, possibly accounted for by a sequence of three sound shifts:

i. PAA *dz > pre-PO *2; ii. PAA *j > pre-PO *¢, te PO-*s3'((5 2). iil. pre-PO-*c = PO *3; *ts > PO *s /#_. (Small Roman numerals are used here to distinguish these suggested shifts,

which probably operated before many of the the rest of the PO and North Omotic sound change rules, discussed in Chapter 4, took effect.) The determination of whether PAA *dz or *j or PAA *ts or *c is to be re-

voa

Proto-Afroasiatic Consonant Reconstructions

constructed depends entirely on attestations from Omotic, the least well

known of the divisions of Afroasiatic used in this study. Lacking Omotic attestations or, in some instances, having only non-diagnostic Omotic reflexes available as yet for comparison, a significant portion of the roots containing these consonants have had to be given two alternative reconstructed forms, with both PAA *dz and *j, or both PAA *ts and *c, as the possible protophonemes.

For this reason, a few additional reconstructed PAA roots, often

known at present from just one division of Afroasiatic other than Omotic, but unambiguously attesting either PAA *j or PAA *dz, have been provided in Appendix 3.

NVILdADH

(JOOI SUIS 3U} UT SJURTIGIS OAT} -ND9SUOD SUIMOT -[RSIp JUTeSUOD Vdd [e19ue3 Jo aouanbasuod se

‘OA #/ bx Sd
) 6.7895 01,

eee!

(LP] x

qeyy < ylog Px

‘TP Od ‘S eFV

‘uoT]0UunX9,,

— —alrevered)>/

CSC

ego

- J(eeZp-,

‘Ajdud dyeul O},,

JO -[PeeZzp-x “OCP

~9BZP-~% “CCV

To} 0},

ZPpx VVd

“6S

“SSL

“LSP

~-WRZP-x

MOLY} ,UMOP

JIZp-»

,snonur

O},, MOTJ YIM

“VW

wueg

:sonueu -oid ias) pauins [euIsLIO JNUdIAF 0] OI uonena Aq ued oy} (stuod

ili

Akeoy

O1, dq

-,YPZp-,/-

,2dratd

-ABYZP-,

O}, 199 JOM

O},,

0},

JILINAS

WWd


U [Ids

sulos YD_ §=pas,

‘

,

JIGVHO

0}, ,JNod “f) WI9}S + Zy -UI

“ques,

JLLIHSNO

&x Ud ‘Sx Ud < Slx Dd) ,,1S9I 01,

,peay JO UMOID,, -besi,

-Besy Ud ‘DS

J6s, YA

JILTASNO

“QAOU JOU 0},

-bevo-, JO -bevsi-, 10 10 -Besi-y “86P

Vv -Bvo-, ulee

,Ostl 0},

-Beo-, 10 -Bvst-» “L6

< OSLI [TOMS

OLLINAS

ISAl 3q 18

(,,pozATered 0q 01, 20S ‘2D sfpe

Joqred pouuns -o1d > *A S"Jyns ‘(pe 1, + Wd}s)

-UI My + Wd}S) wJUZTIS Oq O},,

“Jgns “(pe 1, + W19}S) ,, FOI} pue peoig,,

:SOMUPUAS

‘(OZIS UT

(-ygns [pe Tx + wiaIs) ,Aqqeyy pur yj 9q 01, jefee

,OLSIRYI3]

‘pidio} 9q 01, c6s AT

NVILdADH

Ix 1O Sly /SJUINSGC [DIDID_ pud AvJOaAI[Y VWWd

SAIS f) ,,SUOT,,

(ee

JICVHO

g[qeuajun ue jo asnesaq JbZs,

Sty} JO Suryury

_ ST O10} 1JOO1 JUSIOJFIp & O}

‘YO ur %s,, OU

[pe 1, + Ws,

|[¥1<


(Auvg,)

WaisA (Lx

YIM

MBUS

9}eUIAITe JOOI ayeUIOITe

OMI,*))

Mus

sorURUaS -UN

,,(490u)wra1s)

-Gesi, 10 ~Gys, 01, 4Tds W090 ,. ede

(Dy.

unno,

WOSd

OLLIHSNO OIGVHO

Iaqiog [-ues},

YOryM 99S COS# ‘MOTAqSo0p jou puodsa1109 UI A Jouul puz 019 “Was“Sq

(-)Jes,]

O41, *(N) JOJ

‘UM YOO US» 01,, ,ULYS(f)

,2TeY,, (f)

‘YOM YUDD US,

/SJUANAISG Sly 4OC I~

o1gnd,,ley Dd) S}x< WOgd

‘OH

NVILdADA

AvJOaAAIY Puv

W9}S) + J, ‘INP ‘(peoy)IOA9S *SUd}UT -UBW9S :son IJayO SL®B + Ay “yeoyout

‘Weq9,,,JUoL

-UBD, AUQ, pssodod)

‘uy, “3q ‘Ioqiog “Was -BAOUUI :U01} SB -OUAS

VWVdSx
Sx VSI SUdJUL “INP 15 + WI9}S) (jos YIM) 7e9q 0}, 198% VSN

OLLINAS

YIBUT) (2ST

pud}xo 01,

“PBS-x

“QSIOUIUIT O},,

0} Jayem Aydde

“BOS

"LOS

Rx VVd

(quous|dunt ue YIM) OYLNS 0}, “Qes-% “90S

-QO$-x

-deam

ostUy,,

(jyns “UI, + D)

YIM) SALIS 1,

Wo}s

IIS, “IQS » YOd “YOM eos

OLLIHSNO

/SJUINAISGC [DIDIDd pud ADJOIALY WVd

JIGVHO

[00}JO

w19}S) ,3UO],,

ziupes

ey “WON

JILOWO

Sy

(9

(-Jgns ‘fpe zr rey + ‘ygns “fpe uy +

SLC

OL

.4sea 0}, -Yos vIAOY :0}9WIO

JILOWO

“Sx YOM

(f) .4sem 0},

JIGVHO

-Yoos,

4,9101998 0}, (ys nyxeeX< yey uinq

“Ol

“snj

Jad ‘Jo sovjuns

,YdI09S 01,

“JIP I, + WIS)

- IZ Y3§-

nyeeX OF

,2deIOS O},, “JOOS JLLIHSNO

NVILdADH

(9% Sd

> S ‘Ig “puay

-X9 Gy + UI91S) ,SPNOOUTEL,, qc ys If

eyyes

uPUIM 104,

“(IO Sx +

WI9}S) ,, WIN O},,

Jyes Vy (uy

-UOU U,, + W9}S)

jjo adeios 0},

,[2ed ‘purr oy}

ujes *“V “(sn

OLLINAS

Idd,

-JIp J, + Wiis) ,JUNOD ‘91LIM 0},

S% ‘Stuanasqo¢ [DIDID_ puv avjoar|y VWVd

UMOP 19M 0},

-Yoos-x

“OS

“TTS

-Yes-,

“60S

,() wing 0},,

dB IOS “yoreIOs 0}, ~JOO$-%

+ 3,

‘ued -was

(urys oy) uado dOT]S O} :sonue

WId}S) ,,S3d] 9yI uo spunoM,, SeYS Uf

VSW -ygus “U

Uy,+ Wd] SUAS,

w(S900) = WSN + UPYYIS "V) .2FTUY,, (2) T19J 0},,

US 4s

“MAES-e SxAP

x/-m ITS -M

(‘qIOA

+ Wa}s) Ay -op

-Jyns “U Uy +

(9 o3un “ING < 8x OSd

ua}s) v4oyey),,

OLLUNYS

“ETS

WX “oyTs

OBJ] UT ‘YES,

SB UdVS “PIOM

TTews,,

+W9}s -JIp I,

uATp2 ouoyoo 0}, “w4OveY} -yoo osunmg :OS

NVILd ADH

wlOTG

SNOWIO -NJIp w9Ppots

BYSO Temyes ‘SN.. oxATS,

1nd 01,

JILIHSNO

W9}S)

("919

wostUy,,

"W9G YIM

0},

UOTJBAOUUT poreys *"snj “f) »9SOT9

-yeadar

$y ‘SJUINASGC [DIDID_ puv AvJOaA[V VWWd

,J9A09 01,

“TIS

(9% Sd Sx >

ca O},, 1, Csngjip + 0}, “,JYS PP

\\

D9 (%

3nd 0}, ,ATpoyeodar

“YOS-%

%» \

xTAS

~ sty} Jo uoIs JopiM)

JILOWO

JIGVHO

ELG

‘MOT[PMS 0},

WI9]S) ,INOAOp

jeAjes ‘(103 ‘uly by +

WI9}S) ,,[oSIOUI ®@ MOT[PAS 0},

les VW

(stem Jo wIs,,

pbns ‘vy

WIS)

(Cury

UWOITS ‘MOLICU,, ~—MANNS» JO -yNNs,

sbes‘y

“CTS

OTS “Jees-~%

‘Apooid

(adeys 1001 VWWd

(CAND Hy +

yus 0}, oq

1001 VWWd Puz)

-LOT# s}Oo1 SARY

O},,S- 04,

MANS-» — ,Usybaq 0}, oq “PIS

[tbes ‘jyns ‘u I~ + WiaIS) ,,SqU MoTAq 3Ssnf usu ~opge JO Sopris,

OLLINAS

WId}S) ,onsuo},, yus

snourAuowioy dWI0D0q 9SoU}P[NOM

dnMx ~MATRR ‘dey 4oT]+

JO -,AN§-x%

“4NNS* Odd

an popunod ‘yede

(j juvdp yoouy 0},, Joly] -IB9 :SONULUIIS Syed 5, + Wd}s) Jeaq “TY 0},

WWd ADJOaA|Y puv [DIDIDg ‘SJUaNMSGC $4

Moe,

,,poexs puz) ut (‘3q-oid

O},,,,“(peay) oFLS “010YOIYM JOJ) 99S NVILd ADH ‘Jo J¥s

04, (adeys “dey wi Ss -mACTER

.(3UTy : OAd O},, 2q peyxoouy

jO) “ANS*x

JILIBSNO

JIGVHO

JILOWO

SLC

6LZ

.

JILOWO

Aqzreau :sonue

(vale

-TUIT So[quuasol

,UMOfUOS ‘9181

-ut Ay + Wd}s)

(yeoyd

+ Wd} yUS,

Oy Sd UT mx

‘Sx WSIN
@/V_C (C = [+sibilant] or PO *p ([f] /V_). Examples involving PO *x, and *xz having not yet been identified, it remains unknown whether or not shift #15 might have had a more general environment /V_C{-contin./-voice]. The Chadic shift may have operated before just voiceless fricatives (see roots #365 and 449, in contrast to #397):

PCh #5. PAA *N > @/V_C (cases of C = *ts, *h are known).

Finally, it should be noted that PAA *n, *q(™), and *p apparently all fell together with PAA *n in the Boreafrasian languages:

PBA #5. PAA *y (¥), *n > PBA *n.

+

(WIdIS

‘u ‘quye

‘Jad

-By) ,JOeM,,

ps

({ -2W-~)

‘JOM

Sdd 0},

/MU1,Aut, 9U099q

(oane -oyout Ay ‘My + WIS *6861 JOYA) JOM

0}, Oq

(‘wousp Ay + ,,Jo1eM,,) ,AIaI@M 3q 0}, AMu

+

My

“QIOAOp

(now ‘D

Mu

oe! OSd Wx

(‘jyns ‘wou Uy + WId}S) ,OUM

“69S -W-x

(-sew wax *-ss@Ww BYdoj

‘¢ (¥) sew yousg “SNBO Sy + W19}S) ,YSeM O},,

WON -SPUl,, (-Guix :f) (we “3N)

——_-OS ‘AUDIOS,

‘(ways + ‘Joid

"U “QL -Lx)

ueuw “Vy

Wx VVd

“Uy “89¢

) (, oyuyoput Apoqowios

‘(ggns ‘u Ay + wia}s ‘ oATUTE,

WI9}S) ,JOIeM,,

ayruryap

,odeys ‘UWLIOF,, u1

OLLINAS

unouold

‘dUO,

Urey

(u)nuw,-oid

-UI uosiad-pi¢ Wy

NVILd ADH

‘QUOOUOS

,

‘ueu ‘uosiod,,

“UW,

JIGVHOOILTTHSNO JILOWO

wia}s

(N) ,,2o78M,,

-we, PSUOD

‘WouU ,OS-pue Uy + (uew "dy

‘uoid e uosiad-}s] [enp ,2]doad, ew, WE ‘uoid SAIsnjouUr ‘1d uosiad-3s| Sunou

:Jjas W3}S)

WWd [DSDU Wy :SJUDUOSUOD

BYOO|() ,,UTeL,,

00

Od Sly
:sonue

-yn Ww

0},

0}, + y

JOM 9U092q

W 1§0p YoUIg)

WId}s , “WIS INP askez, -Z00u OF, [Poy

Wax WON

-Zpoowl, WON JLLOWO

WWd

vOe

(Plo

Jarprea

(‘IOM

1001) 69¢# + Yx

“Yeeu-,

uye

“pgs

WIS) + Tx

UMOUY &

oyeul

[ovw

+ Jy

(CULJ wiajs) ,,(puey)

01, pus}xo ‘JuUas ,J0199S (‘9U09

‘ZuLIq-o1d

“21

O},,

‘Vv

gz

“Esc 30M 0}, 393

‘(19U

01, ,,J89q WI9}$) + Uy, (‘ulj-uoU

w

3SOYo UIIMoY} ISU WIS) + Zy *SUdUT JO -UBUI

Ay -ul Wd}s) + (posiow SUI090q -WIT (uoneord “Yeew, qu

,UMOIP O},,

wORNE

-de *A SUTALIOp -d1 ‘(‘wuouap Ay

-YTW-»y

01,

‘youne

y[TeSse

/-OT, DAd

JOM 393 0},

amnydeo,, (Cu)

OYLNS oY} 0},

+ ‘INP 1x) Aly +

O},,

:

“Joid U ‘OZIOS 0},>) ‘ploy -qoid ‘31ye we “O71 ,,amides

qQUUe -Ty 2148

OLLINAS vy zyew

;

“78

JIGVHO

PSU0D 022:0Ul,

\*

-YIW,/-4ew,

NVILd ADA yw

Uy, ‘/STUDUOSUOD JDSDU WYd JILIHSNO JLLOWO < A Aq *A>‘U) ,UleIs ZUNOA,,

SOe

WO (-20Wy, “A

Dd)

‘INO.YHOJ

Yous

01, Dd (94 ZPx < -ORUL,,/-ONUy

aut09

“yoyo

-feu-/-fnu-»

“YIOJ 0}, OWIOD yous ‘as1oui3 Jno

O1,

0},

“[eU-~ “uMOTJ

UINo1 ,

oum09 “yorq

ePUa]~

-JBRUI-4 JO

“YeU-» 0}, 389 ,d0

VW yew 0}, Yons JNO

m ({1d

SUIAOUW9suas JO pajatdwi -d8 os

‘(adeys I-Wogd Aeypeud,, ,Ysty,, :

Joo1 :Dg]

uRoy >

[o1sWIQ

,snd puz) VVd

9ZOO “INO

“YR

-y-W, w2INS,,

JILTHSNO

“STUDUO UW,sUOD

0},

wy

,,9UI0D W19}S) “SNJJIP -OI

exw

(uol)

0}, + Iy

Ay

SMyU MAY ,pooy,, Ways) + A, “qroaap +

NVILdADH

JOSDU

yuu Vy Zjew 0}, dop -019x9 ,,.XeM,,AT w-, “UTR “KEM, “$90q Jud W9}S) + 0}, ‘9Z00 aq Oy *PUud}X9 (xem puz) VVd 1001 SNOOSTA |, WI9}S) :sonueUlas > + Yy S91 “WS JouUeUr oy} ‘adey ">s yjnu ( :sonueMOT} STOO}S SABI 1no (31q-£q-31qoy} (snue

“L8¢

,

“98¢

“CBC

OLLINAS

WWd

,,Jo1eM,, (f)

su0s UDA [Wx

id

JIGVHO

,UINIII

Feew, ~

(nur

‘0,

IOJeM UT

Sur
°A

0},

err

OLLIHSNO

:SJUDU Uy,OSUOD

39],

‘OA

u

NVILdADa

“URUI-» “Vv juew uu 0}, ‘SOT “YOR] 0}, “asnjo-UTY l *SSOy,, O19} ST 9q TOYA ‘Iap ‘WuoAald JOU *>) uowulu 1001) TLC+ # *(,,J0U,, Jadar , Wid}s) + + U (‘UlJ-uOU Ty 1uu “ed Se

“68S

"8g

JLLINAS

YWYWd JUSDU

CA

,

Aq’a>‘u ) TUL,*-U]

:0}9WIO eIAOY ayurm

JLLOWO

LOC

OLLINAS

“JUW-»

,PUTq

“T6S yy psu NOI, -3p

7

yesew

“Vy

“76S

wa}S)

, Sno

xJew “(JOT Yy

ods 0},, ‘qsaf WI9)S) +4X -XO “pusy (‘Hoy

,Pulq ‘D)

“INP
hq A
“a't

Be) ea

Py + W215)

“yes < “INp

SG 8 OTT, pyeu Vv

(JOaJJO JO SUD}

“UL Yx + WIS ,punoss oy}

OJUT YUTS *QUO}S

B OMT] YUIS O},, yogu if

‘( [eeu

“Ys sgns ‘(pe |, + Ws) ,pastdsap ‘ssoy “YOM “peq,,

OLLIINAS

M95 JO ‘OPT quatoTJap 9q 01, -wHEU-~ “819

-feu-» “L19

MOT 9W099q 0},

TQU x

( gbou uy

L:A Jeb, YOd peuns -aid ‘u, TeIWOS

-[903 Suns ssaidxo 01,,

< fx Yd “D)

A190 ‘doom 01, [nu TeIWOS -YOU

JILIHSNO

-jel, Od JICVHO

Gy, /SJUDUOSUOD JDSDU WWd

Vv qgeu

-F10-»

“SEQ

01, ns

-Wol-»

“9E9

oY} 0}, 9SNyynoul Joyjo0) UY} UT (Sune

“Vy

,9Z90Ig

gugu

JosDU

-yrG,,

(‘yeoys Ox
JorpIea “A 0},, oq ApeaI3 -yeam

w1PoM,,u19}S) Bp

./-Ieu. “aU ,[Teay *

Ad 0}, 3q

1apue3(‘Jjns

WA Fy

2OS MY eZPB Au saya oy 3ury}, Jo -Iodns jeinjeu ,Josuep was) + Ud Sx Juase “u "JJNs

OS UM “weux 0}, 2q ,200d

+

JILTHSNO

-SJUDUOSUOD Uy.

{Jou -oadxq *

0}, 9q

0},

adexos ffo (ssouysno1

NVILdADH

JDSDU

“YOO YO" euos

“Li9

“9/9

OLLINAS

d YY

JLILOWO

+f)uoljoNySsUOdAI

.2B9F

01,

e3u0H-IlU, BYDOJ)OJTU, IOS, “O1QIXOTJ MO][OUI (,,

0}, 3q “WJOS “OSOO] w2®IQIXxaTy

Ee

“619

“p79

:soljUBU -UI aS UONS SI asIp -d8 potuAq eduios SunaJo3zury jusxmoax (ured

IOI -

yd ur (m

(mojaq

SUIOD ‘O}

XIJU

‘’sneOd o

-meu-,/-Aeu-% “SO

‘T[OMS ‘OSTI 01,

,puedxa

-X9 ‘OSI 01,

SI SITY} OJeOIpUr "Wd SpIsIno

-X9 Gy + “SNF

,Sol] Jon ‘Iopurys 0},

$d}eUsOd Spud}

stu ge Vv seu

‘uOWUUNS + W9}S) {UOT} -JIP 1, ~=—s

OLLINAS

wATPNOT

Mu,/KUx Sdd

puod -dy)‘(| ,pued

jou yenAe_y -Ipoul [eUISLO (TeIONTLN

(Atu, ‘Avuy :f) (N) , ued 0}, eux e,

(6 ‘wos < Uy + Wd}s) USUI

Odd -"Jg0s U I,

2OS UM

JOUUBU

(3dy3u

fugu

*Ssud} JO

“mXBUx

JILTHSNO

:Sjubuosuod Gy 40 Uy,

NVILdADA

jospu

1,

oud , (6861

Sdd qu,

(‘sudqur

:-Dq

C#/

eewioog = SN

‘

“ygeu

JIGVHO

(q

AOIYSNe]S gaat 0}, ‘utof -u0d Was) + [Px O},, oye B ,AepI{Os W938) ‘CAU astoaId Nd ,UI + Jy SINP -WOS ipdu Od) Uy ‘U) SPPelq,,

snonb

TFV ‘OW JILTASNO

upT

sm

‘(ysIy MOIS

Uy + Wd)S)

“ot “Uulj-uou

w4sty Mors

‘punois ou} JO NO O¥PI 0}, upe’s VV

-dy) ,,({pooy yyos “prnbry)

O3% Sdd

MOTTPAS 0},

JLLINAS

yos

wT

“osTed 0},

urese

op ‘teadar 0}, “7US4 999

“ONS-% 0}, MO]JPMS

0}, ,,{ULIp youog) -tsn, WON

JLLOWO

Jo pinby

YONS1941930} 9qZz SUle] 85 0}, dy O},, + [Ze °V wia}s) (‘UIJ [x WHOIS) + ,Apuaned

(T"puad

‘(‘pua}xd-ndod) d19A9810d ISP} ,(S8Op

3% /W¥d fo Stupuosuo, jvasukuvT

uPOOJ

“S99

99S hq A ‘U) < ‘A> aop: ‘WOS (SUIMOT #/ UIP ‘WQS oI ,UTeyUunOUL

1 (4)$n yousg)

Ox 899# < Ty OSTR JOJA

CVe

“L99 “PES

0},

dus

“IPRS-*

,uosiad , “PES “899

“699

‘mpny MpIS

Wl,"u

[TRI

(jgns

W9}S) GL

‘Moyd -MoTd

oppuey,, JO 24}

ng wp

WI9}S) + *(‘QIAOP

“quiqy,, -Woul

,13q My,

“Vv

Totes

bpes

WIS) + x “UI "sud}JO *JDOJJO “art poievadal Surjord (dn

01,

NVILdADA JLLIHSNO

Slupuos fO uoZ “WWdSx

01,

,UOs ,dus

«/-[PESx “IPMS

* -rad Jo Apog,, “PH x/-PES x

OLLINAS

jwasukivT

JIGVHO

(-Pox

JLLOWO

O},,

3n iy,

~

yousg ,4oyed

‘WON

WOS) ,,uosied,, “Pl

Cre

wAT

-3U0IM UOTIOUNy JO op «ITI “peq

Vv ppes

V MQeS

‘ZUOIM

(‘YEOYOUT My + WI9}S) ,,POOs ST

-UI ‘Liny 01,

yey WO WIN} ‘ropurys ‘aml

,ounid ‘doy “nd 0},

JLLINYS

.JJO 19 01, -IPNNS-» JO -IPN3-% “OL9

O},, 2q

1Jzt

NVIId ADH

TL9 -Zpes-x ‘U1,

+ “inp WSIS)

‘,AODSep ‘dn yovy 01,

(jyns

‘1aiYysneyjs,, w9}s) ,,aloesseu l

Abel] -pingye

‘Jord “A -[Bx UA)

+

seeb meyy :O”

< ZPx

( 73 Mbely < Px Od -W9is + uonoe ‘basuod jo

“IO Op ‘SUIOPSUOIM uPOoyesyey Jy 0},,

"WOS ,SSouyyt yous Yeeuloos

Sx /VWVd fO Slupuosuo, jvasukivT

:DS JIGVHO

“u ly ,,{SUOIM + (qgns 5 IP

Tysoq ue Od -Odeo O},, Aq ,,UOS ‘U) > ‘A WOS 1-Wd}s + [3] SS[MPI,, -Iad A 4s ([s] 9-sned = “(AA Aq
“A YUM “FJNs “U I, JO “SNjfIp Ix + W19}S) ,,SSNOSTp

‘jnoqe ¥Ie} 01,, £ fou

NVILdADH

3% /VWd fo Stupuosuoy oasuksvT

(mayo

0} ‘o'r “pusd

-X9 Wi, + WIS)

Moyo ‘dI1q 01,

wifes Vv

CInp Yx +

01,

“TLIO

“JES-x “€L9

-Be3-% “Vl9

oye}

wlazy Nou

e

0},

enn

wiajs) , ATposny -uoo yeads 01, YJeS

‘(SNJJIP 1, +

[-J03-x/-S}S-*

0},

-Jeey-x

99S

Ways) ,, POIs 0}, Jes Vv

(‘jd Ww, + WIS ‘dnpaz) ,,SoAo,, w-d3dy

onus()

OLLINAS

(jgns

My +Wd}S) ‘U

:

O1,,

(189

WV 1s9v WOS 4

‘,Wioo1,, -S114/-SIB WOS

,,41001,,W9X]

Aze 2 ‘[

JLLOWO

QU ~QTx/uG}00} 01, -931q : WON)“2% Ppoyxooo,UTeIs

Mebev ‘SN ‘popunod,,

JIGVHO

OV.

pepe

O},,

(A[snousael POO} OFUT 931q 0} Apeal oq

“MAPRSx JO -¥eRS*% OSd

,Sotoods ureds ,, bevy,

vie!

“-JBOYOUI

(JgNS “U1,

+ ‘Jyns ‘u My +

80g “dry

‘yeoq “youue 01,

YeS VW

(6861

Q3x Sad

wollq 0},

OLLINS ~

yoy)

Jo ,,UTeIs Wia}s) uonesedaid b,, 16,

NVILdADH 3% /WW¥d fo Sjupuosuo, joasukuvT

JILTASNO

+ jo YIMyor] WIJS) ‘suajut Bde‘Iosuny mMds1 ,,pooj 0}, d, AT My +JouUeUL

poene

S-x “MACE

0},

| 0},

JO -yees-% “LL9

Pa

:u0ly -mebeey,

(-Jgns

L13-4/-feS-x “9L9

U My + WdIS

. UTeIS PIyxood,,

‘YD -BAOUUI

39) ,UTRIS,,

“Bees. “$19

“TS

(uI1a}S

45 45

‘dnpe) ,,ury9,,

(Jo ‘To

‘D) ,pusose

Jy

‘dn junout 01,

(TeSx) 1S

,oAoge ‘UO,,

w9}S)

“Hulu,

(‘Imp WAL

OS

+ 14

OLLINAS

dn 03 ‘pudose 0},

“Te3% “O89

Mel

dn OStI 0},, -483-4/-AlS-x “BLO

“TeS-x “6L9

wil929Y9,,

“Te Sx Od

,UTeyUunNOoUL,,

(fpe

Jaleo pouns -a1d > ‘A “‘Jyns (pe 1, + W91s)

31g MOIS 01,

JI¥3% VSN ‘(‘yeoyo -UT My + WI9)S)

w4Sty poster

“paistoy 2q O},, Mbes ‘(qeod

yoral “ort “Uy

Tx + w9}s) ,UTey

-UNOUI PuddS¥ O},

NVILdADH

-rwb es MbeIT OS

JILLTASNO

0},

(ats < Avis
Aq’a

ISLx

) ut Sir [eur pur Sey JURLIVA wa}s(A

(4S€Zx)

wes @| 0}, ISISOP “WO ‘ALIe} 9q 00}

ue Ux “IOI 0}, oct 99 ,,dn wiajs) -DAOUL (y70 “Dd ‘uo UOT}JOYVINJ yey) S}OLNSal

NVILdADH

“669

Sx /WVd fo StupuosuoZ jvasuKivT

yeewoos -AT[Iq 10 ‘ured

([3]

P

Uns, (f)

JIGVHO

SIP] WI9}S) +

/-113-x/-783-%

“INS u(Z'Q) 04, dojs “TOL

783-% ~

0},

TG)

9U1099q

O},,

VV

0623

‘ULIT

01, 2q

+ My

,ounyiodunt

‘D) 49 (

“Jez,

01,

BUOYS po

‘prey

Po

‘JAT3 ,JURIS

UWId}S) + M, -op (‘wou

"Vy s$e5

0},

“(‘jyns

Wd}s) + Uy "U

JLLINAS

,OSPOI3

ony ‘(g)

“Ped

SWS VOL

“INP Py +

PIE ,pueiq,,

10Y,,

(-jgyns [pe 1, + Wa}s)

04

“80L

(€99#

SOJVOTPUT-19}

‘U g]qIssod Uy + WId}s) ‘Jyns

w93S) ,,yUOU,, *(‘U)

0},

19@2°V

“QTL-%/-Qeel-x

‘,,ULYS oy) YoI09S ‘pueiq

“qe

199

O},,

(‘Inp < ‘yeOYd “ul Ay, + Wajs) oqqel

NVILdADH

jwasuckuvT

BSE

0},

‘SUDJUT JO

GCLe Cee Pe 0}, 108 Ul

ppel

Wd)S) + Py

q£ O},,

PE

(‘INP =

pnqnd ‘dois yey ,9se9d

‘ALIR} ‘Aeys

OSd

:Dq

O},, ySUT[IOI

-QIlx

IPx< Ud (Be

YOM -G-x

Pern(f)

JIGVHO

0}, 103

a10qry pe~ 248 Od) ,1O8UP 0}, ‘X9A -I0} 0}, OQ SOBPARS 01, s1OQIV {USUI ,OAOUS ‘ 3q SQAISSOBSB[Px< (Pp wepez Oq ‘AIZUe Benet “IO3UP,,SAUD Syorne (Josue OS Ud W9}s) + Ul, sommnbed) -1sod IO -wbepiez, ‘u (jgns 3un p-3q < 0}, IBIAS ,4JBO w1a9}s) + -qoid UMOUYA[UO. ut ATqe6, ‘uly sty} ‘pominq ‘105 SutA[dut JOU 134 -19jUNOD ssoujda109 JO poyedIput (xoyII9 1ST “WHOJOd

Vv

Vv 0},

(ZQ€»

-eJoU)

Lx OAd

“Oise: 0}, ‘JIS ACIS

SIsoy} JO

(20ue ,,

wy,

(souUeU

‘99191,-MO]B

Sur Jo) “Iped

qz¢

JILIHSNO x/-Z98

SLEL

Z,,

~ZQIL-4/-ZQ2U-%

y9ZP]q

NVILdADH

SlupuosuoZ fo “WWdLx

“210

yah se‘ CH/p

+ 60L# 1001)

“OTL

OLLINAS

joasukivT

JILOWO

O9¢€

JILOWO

JIGVHO

(‘qIOA

-Sqes

~Yeel-x

,Ueoul 01,

IO -Y¥J-» OTL

IO -A®I-y “CTL

0},

~JNNT-~

,MWOTQ

CEL

,osUeLIe 0}, -mB¥L-» 10 -Byl-% Vals

-OP My + UNOS) u4sis pur POAdtIs 2q 0}, uted “ArosTU,, Vv yyel my €

01,, AT) [Ply

-9uaq ‘[njasn,, qe

‘D (Jom £ JM

,ssuny,,

OLLINAS

-AUCI-» 3q 01, ,AIODEF

(apes =

JO) MOTO}, -ye MbEI] OS ,4snoud aq 0}, -Aveqy JO -Atly

(909 ix + WO}S)

(DS Cx “MARL

O},,

JIGVHO

¢/-m “MATT {ex /ey Nye,

JILIHSNO

lx Te 01,

,,29S(f)

SIT], $1D9I09) YU,(N) "3N) eye 101Nq L861 ‘Od ( *f) (MY-»

Ux Yee

OSd

0},

i4ieo,,

Tle

YL

NVILdADA

stupuosuo fo “WWd ) Lx

OTTZx

“OTL

“OTL

[99U

“V

AIA

3seq,, jo

“Telex

*(OxTyOZ uCU)

“mACBl-»

o1,, wingJO)

9q 0}, JOY

SIL

“LIL

OLLINAS

joasukuvT

~[@x 0},

‘WON OPT iMOUF

JILOWO

I9¢€

FAL

“Ul, HI WON

(0661

Ia\jfe) ils

a

(Z PUx) AUx

yooAIYS “pL6] yer

0} :sonUPUaS

(OUIYS ‘MOTs 4 < yedY JjO das “6FINS “U $1,

SHOWLY + Wd}s)

JIGVHO

ree

TULL x

01,

(J[asouo JO} PUSIUT “°9°T “UDA Fy + WIS) ,2s00Y49

~Fe Wel, Ud OS

(1/Uy < Wy

01,

‘INP 14 + Wd}Ss) [O09

“We lx OSd

“yoy “uly 6,

fewer vy

AIDA 9qQ 01,

,(91zeI1q

NVILd ADF

u¢

dn wing 0},,

+ W19}S) ,,JOY

isnf se ar0y

(“Oy

Wl Sdd

(6861 19249) ,PUSIUT 0},

les

Jo) ung ‘(-0)

,osnqe 0}, “y[NSUT OSd -49l% JILTHSNO

OLLINAS

JY

O},,

-Wel-x

,PUSIUT O},,

JO -UT{y/-UZx

PA

“PTL

“ETL

~-Wel-% “CCL

PCL

,ATPno]

~SUI-P

ORIN ‘WON

JLLOWO

Lx (WWd {0 Stupuosuo) joasukuvT

yeads

“UT 4/-UP

yeads SOC 0},

Ulx

(1 Sdd pusddy) #lx

0},

,UOOUL,,

COE

Jno

z 1urA youeg

(,,(enqjod) noK,,

‘pL6l SUTwAL{ 0} SUIPI09NR) (SUIS)

noX,,

“WET

(H#- Al#/ Sx OUd < iTS,

ub

“Sx OHd

SUx Od)

ul

“UPlx OSd *,0M,,

-UU TZ x/-UURL x

(-DP-»

uzayed 0} Sut -p109oe8 < [euly -wia}s Jo ‘dnpor

kq ‘1d VWd +

(SUIS)

noX,,

(sur “Y)

,ueul ‘uosied,,

eu-y

OLLIINYS

UULy

2 OM,

pZL# Jo was |‘uord ‘jd uosied -IS] poxijns ‘-uvul, adeys

uut

vom,

Joes > ‘uoue "D)

NVILdADd Lx “WWd fo Stupuosuog joasukavy

JILTHSNO

Wx

SUL

BSF * Nx
)

JILOWO

/euou, WON

eulu, /eunu,,

Odaid

ele

uCBuIS) NOL,

-WUT,/-Ue, IO

“WELy/-WeLe “LOL

(‘uoid ‘Jopur) ,uossad

‘QUOSTOS ‘SUO,, “SUTUx “OTL

(xojJor “3q

ut pardunt st

“) Teulj-wsis

suIpasoid = 2d

Araya -3(®)-x

jo

1d)

voM,,

sutppe Aq p7L#

-UU J ,./-UUR, IO “UU TUy/-UURL x “CTL

VAEES

=i SsCx

yousg).”

—-“([f] -Iey

0},

JLLOWO

,MOUX

(¢ J2

el “WON

00j,,

(N)

, 909

YO)

est,

0},

5 eS(®)

JIGVHO

(N) 39/1005,

YSZ

0},

0},

03

JICTx

-gee AvlINg ‘Og

-

99S

0},

,dATao1ad 0}, J45

(jenp 4m + WI9Js) SIR,

NVILdADH

AMqU, -MXUTU4/-4XUPT

-JOLx

(,,998 01,, Sd) ,WOUZ

(7 puaddy MesV)

,0} uoQUaNe

+189,

Aed ‘uaysqy] 01,

JILIHSNO

Lx (WVd fo Slupuosuoy joasukuvT

uSSQ]_ -daajs ‘ayeme aq ‘O¥eM 0}, besel Vv

(‘yeIs < ‘pudixa qy + wi9}s) , Juapnid

qnsel WV

“OSIM oq 0},

JILINAS

“TEL

89]

SyyeM 0}, “SPl-%

“OEL

“STL

“6CL

(JoaJJo JO *‘SUdIUT x + 6TL# 100J) JO dIVMe SUIOD -aq ‘d0N0U 01, “AFel-%

01,

“Jel

,MOUy

(mXx / ([2])

/O/ < Wy, “HOS “PUdIX mXx

+ €Z# 1001)

289 [UdISTT 01, ~MmXUCl-%

‘UOT u81q (We ,,

nM

Jy ‘SUJUI JO (‘wou

Wd}s) + Ay -op

Wa}S) +

Pud}x9d OU})

:O7 Ieyy bomop

0},

w9]s) ce 3x

“ed “O° oq Os

SUI

x/-SIPL “SINC

(daqjo

,OSIO},,

,yisuay ‘

“SITU

UI9}S) +

‘suoy,, {Me

Jas INO OJ

JIGVHO

WON -SJ8,

Me WOK

(¢ Youdg $18

Ue} ,,

ne

JOYsTY

bnez‘y

0}, 2q

x

-MBLx ‘[[eI,, ‘ysty

SOALLIP -SIA

JOU WIdJS)

OILTHSNO

VYDOJ) B OS:

,uosiad,,

JILOWO

BYDO/[ OMME

Tals,.3Iq

Iedi, me

O},, CALLE

“PEL

441

‘OJUI OF pee0}, qeewoos -OF

“SI@Ux “CEL

RCs “WV A1el 0}, 3S INO ,JOF O},,

(‘yeoyout Ay +

CEL

NVILdADH

Slupuosuo, fo “WWdLx

:WON yousg

“BurquiT}9,, -IIxo OSI YSIY 0} SB refy ‘Jono b=

([3]

JLLINAS

jwasucKivT

712

“DdSUC] 0}, 2q ‘(‘wouap A,

C9E

(,, M249 0},

WO} “pua}

sey L861 174

“MXZ

-X9 4X + UWId}S)

JO -ZUL» -Z¥4y

O},, Abus x BL

JIGVHO

:

ULAS,, (, ,

Uy

0228, ‘wAPOq,, JESU, +, BYDOY

Yousg y28

1zA:@ elARIOM) ‘ose OVoyseg SA:D WOK "Bg2 “YSdTj,, Jou

-$skK:ey JLLOWO

uPLOY 94,

Tgse ,eoul,, Od)(C/ Sx wumy ¢< OILTHSNO (Mes ISUMY

NVILdA0DH

oo

Sig Lx jwasukuvtT SlupuosuoZ {0 “WW¥d

,PIOOSTA,,

(‘pual

-X9 +WI9)S) UW,

OUT ,sa0erd 9Y} +4399} 331 TT? Yt 31g 0}, wZzel Vv

(pus)

-X9 Ul, + W9}S)

ATUL .poyoene

oq ‘orToYpe 0}, wZel WV

OLLINAS

GEL

wd 01, ~ZBl-x “LEL

uP[OY 94, “hel

, “Yysozy,, boul

“CEL “SABC

“ODL

“6EL

“SEL qy

NVILd

x

(‘wou

OATSTO

“mp 2 =

S40}

({uJ

+W9)S) Yx

,, SUTPULIS ,, Wi9}s) + Ay ‘(‘qioaap Iejy YISTo SuTyZNYs,, Suoye uO s$,au0 -jnq

(D7 IESY

01, 9q ‘TeIsoy OCU u[PANSOF Wa}s) + A, -op

“gee, wyeads

“Qnd,/-Qe

JILIHSNO

‘daoiq,,roddn ,ULTe

OSd 0},

SWINOUL 0} aye1qQ2[99 ® -1) Sydun ,,Je@AnSsoy

u(TMOJ

yreo01, Jo) Ysty pue

ADA

qy 0},

Aqy

,JSUIVZE 0}, ‘qni qni JJO , W9}S) + Ww, (‘pud}xo

-OTY-»/-98Y-» Vv wgey

(HOy

wa}s) a Sx

o4,, pfoy YBN =,

“QNY-%/-Qey-» x SQu

“Qeey-x yeods‘jno ,Uutepoord

uP[OY

0}, qnI

0},

0},

VVd4x

OLLINAS

oasukuvT SlupuosuoD fo “WWd Yx LOE

JIGVHO

UO

,,

(2u0

1$8(Y) 1 ,JUSUOIUTO Wd}s) + ‘U1, “Jyns“dT yey} yor st poeqqni

:WON youeg

,

JILOWO

“Trl

“Ip

0}, RJ

“P24-% 9U099q MOIS

uPunoI

“JAIN

“924-%

puaq

O},,

UrB sny

gpeq

pouinsaid JoTpIea “U -Ud,, ‘ansoypo(.,PIOJ

‘Udy “Jgns

0}, [[OMspue 393 prey ,, UId}S) + 1, ‘pe :jyns ny Janp “qory},, UoTJO (, MS

Jpeg

‘as[ng aq ‘xoAU oq09 uP2yoeqyouny Wd}S) + Gy, -X9 ‘pudy < (‘eIs

0},

Vv

*a >

+

utds ured

NVILdA

Wa)s) + Ay -ul

0},

Is

DA

Ay=

OSd

,oimeug

DSd)91,, 2q (,JeJ

‘dn

“poy, 01, ‘mod Mors

Od (94

vous,OSd)

AY, ~

JLLIHSND

Stup fouosuo) -WWdYx

(,,Peom},,

asopousapes ‘Jeoy < d “INP Ploy ,, W9)S) ‘Dd ‘soy ‘sooYy

[Te

punouins

Izeyg-v

YIM &

0},,

JLLINAS

[vasukuvT

“INP

(€OOT

uew ‘ssp4s JO} 190 [aTered ‘suOTR 99SALIAp ‘SLZ# “787 "€S9 ‘TE8 “P98 ‘S98 “€96 "P86 Pur

AZojo jo ur ‘ystsugq ~To

:sonu ‘Jo e -A}9

ouos YO PYUx2 Jed], ‘f) -W9s

(ayeus

:sonu > eures JouUe joUI JUOUI jo OAOW

woyeuS “f)

WHS + Ax

sutOs YOM YS-%

JIGVHO

JILLOWO

89E

69¢

JILOWO

JIGVHO

(Ax Od > 5»

Odd) ,3ump

-dWOS Op O}

oyeul jtse 0},

Jey Aey ‘uayeyy Ol, Jey,

(te OSd < (odor

JO 3ULNS OJUT)

IP Od)

Joqyy STM] 0}, -pny, Yd ‘OS

JILTASNO

{(sirys punos

“QIsAoOp My, +

a]qissod 10J sjeosuArey] JO UOISSNOSIP 99S

"UTS [x + W9IS) ,Jer9ues UT ITeq ‘yonp-Aooap,,

MEY (-3yns

Ud, + ‘SNZIP

I, + WI9}S) 910} -sod Sul[Med9,, 1¢J4 ‘(DUI < 9ANRIS se -Inp py +

psy

+ Wa}S) ,JIS 0},

NVILLdADH

MAvY

INOUIIM LIsse ‘uotutdo Jo 3q ‘gsoddns 0},

‘(ways ‘dnpod) sUTeISqe ING ‘Suryeads jo yutod uo 9q 01, AeydAey ‘Vv

(T pusddy)

dy, Sdd

,uMOPp Ind 0},

(“HOJ “uly

dy + “SNJJIP

LLNS

Ix a w9}s) AT

Vy -JY31) STM} O},,

JHd

Yx -WWd {0 stunuosuog joasukivy

qeipey

bey,

OP O} FI SB OYLUL ‘UTOF 0}, -Av-» “SPL

OT;

-jey-» “prl

Ab]

uC) ISTANI 01, “Tp Ant» JO -[PNY-* “EPL

OLE

JLLOWO

‘

e

;

(u) dis,

-WIY DSd

“wey,

eel be

(f) ,(poom) dAIeO 0},

(‘suojut

‘pusddy) ,,(‘010

(I dAulap oi ‘dnp

Aeme dis 0},

—_,,((919 “ULys “Trey)

(‘UDA Fy + W9)S)

,UMOP 03 0},

Jo -yaad-x “OPL

-m42a4-»

Wi9}s) ,,Uonep

(‘yeOyouI My +

JILIINAS

JUSTOTJNS -UNOJ

syeq Vv

»-ULTYs 01,

wy, Sdd = - Wr y-,/-wey-» “Ly

‘urys “rey)

AOUISI 0},

NVILdADH

wy

—--94) (4) STUY $,UOdSINS,,

‘(moy “D *ju

Jaiysneys ‘Aeys ‘ayBNSv9 O},, swy YT

S + Wo}s)

Jy IO “WOJ

Ux “WWd {0 Slupuosuo) joasukivT

JILTIHSNO

Jo -yaay,

-m42eat x

,UMOP 03 0},

WY, YOAA eulos

JIGVHO

‘INP

“OSL

“6p

“SPL

purey

Jewye ,,par,, W9}S) TA -e,, ‘que ‘Joid + 1, ‘[pe (ygns

9xeUW‘JOY

f*JINs

+

0},

Ayjetized)

(2U09)aq 0}, 3q

*(Wd)S

3nd-S,

"sned

‘OprAold dinba , ‘D) 3uyes0},, -O1d

‘dnparways se (‘SUSIUT

,10y

OSd ,ULIeM

-wey,

*sud)JO

WS+ Zy -UI

YOO UYx

JIGVHO

-UeYy 0}, Ind

;

,9oerd

JILOWO

WON Ife

(19UUvUL

0}, ,,PTINg ‘(f) YOM Pury, 0}, puris,dn“f)

JILTASNO

“Wet, :SOMUBUIOS

nos < J9}1Iq < ‘Ayes ‘dD (nwy

0},

uy

NVILdADH

Stupuosuo fo -WWdYx

JTRS, W]s) + Iy 0}, 03 ‘peq ‘fpe ‘yyns + ‘W1, wods

1gwy

*) wy wo

rorprea(‘{pe

mumey

Joy *

0},

Vv

*A >

“fpe “yyns (IM

+ 1, “INPIO 1,

qewuey 0}, 9q “uel portods ,, W9}s)

“UeY-4/-UTY-»

WHA,d0N

-weq-, JO

0},, 12S INO AR] ‘ino 3nd ojut ,osn

0},

(Inos

vy

[lods u(Qut) 0}, 9q “INOS “was) “3q -ul ,,proe WIAJS) + :uoljBvAO 0} U [Px °CA'U

[1ods < 0}

0},

“Wey-~

OLLINAS

joasukuvT LES

pny

“TOL

“UOY-»

-o1d 490fo1d ‘dn yous 01,,

ALIv90})

(INP x + OSL# JOO) ,,(WIOIJ IO

0},

~“yuvy-y/-JUTy-%

( yuey ‘ yuoy

-Inq 3q ‘1ajjo

yuy

‘yuasaid 0},,

‘(‘QIOAOP My

NVILdADY

“TCL

quyuy

-O1) , SUIT[OMS,

IMuy

‘u ‘(ygns

Cygns “U1,

(4#/9 jo dnpor fq ‘a

+ W9}S) ,UIOY,,

1, + WOUdP My + W19}S) ,, peo joumoi,

> u) ,snqeyd,

uuy

ojuoy

oreyeq -OS

APMP O¥P}

‘QAOUIAI O},,

-YUTY

JILIASNO

,AQAU0D

JICVHO

joasuksvT Stupuosuo, fo “WWd Yx

0},

OLLOWO ~

JLLINAS

‘D) .Peusp Gy UWId}S) +-X9(‘pust

‘971189, -sassod ,,SUOTS W9)S) + Muy

quy

GLE

ESL

0},

,OVUT

puaddy) ([

oq

/-utty-,,/ = Sqd -unny-, uy,

0}, O11q ,OJUT

-Ueey-,

OLLINAS

,,PdoIS.,,WId}S)

“Udy CJgMs Jy *(WasJO

,aeyed,,Wi9}s) + *SUSJUT JO

yury, -

(asuas

+ ie

SuImMay9o -o1d

“IOI ae

BATTS,UIO}S) ‘Ppud}xXd + sodonp (BAlTes

9,

WIS) + Oy -X9 ‘(‘puay OATd “juouey,

OSd “Hund, 0}, 9q -19A09 ‘sno ,Apoois 0}, MIU ,, WI}S) +SuIpaoed) + [Px “AUOd— ‘U [Px< OSd ‘(x A, “wouap Dad OUT. /:SONUBUIIS Aiguny ‘I0Jul oueey, [eouoydeyour O},, ,,AdY UIO}S)

Aqy

+

“Wd, ‘borg (yyns

‘dnpse

quy

+

JILIHSNO

Stupuosuol fo “WWdYx

NVILdADA

jvasukuvT

JIGVHO

OLLOWO ~

PLe

-utd:

SPJ -1)

eyooj ede

0}, 3q

VaL© JILOWO

Uy >a + Wajs syns ‘dnp “fpe

sorprea (-fpe

w[QFyINoW,,

Ploy 01,JO 2ye1 “PSL

-dnny-, Jo -dny-% “CCL

-det-~

-Op Sy, “snevd SB (‘wou

(inp

-dey-»

,ualsey 0}, “OSL

nyeex(3y

Od‘ VY%


+ A Ul Wd}s) “Jjns

+ wi9}s)

w9}S)

OM}

my + U9}S) JSPJ FCM 0},

OLLIINAS

ADH

,d0RIQUID "U1,‘uU>

Vv

joasukavT

jody NVILd

OF Y- 0},

Sd

OSd0},

+ YT ady

{XO 4yeolq,, -dnnu,

-WOS1°A

SPJ -WOSTq3

surdegy,,,0UepOd dy. < x

SlupuosuoZ fo /WWdYx

Jo -dn.

JILTIHSNO -dey, ,OSP]D

nyee

JIGVHO

“MeY-% “OOL

“6SL

0},

0},

w4oyelOS

w[IeY

(6861

,ode1os A) yoy

*ydyeIOS

O},, 9q uPoyxeu UId}S) + Ay -ul

peq

uSSUT[I9}+ pueddy) (]

Sad SYx 01, [995 “Peg SAIS JUSA0]

0}, ‘dARYS qni JJO OY} .AT@Y

(WO

,

TMY

uX®,,"Uu) >


) Od Yx we

*JB9q “SIU IO -oyxLns u4Sary} Aq A W9}s) + Ay -ul ‘yeoys < “Inp

O},,

IOI, 2x [pe

,,2[qELIO}W9}S)

PAYS

“424%

-OS oreyeq S9dIIT],, OBPARSY S18 F,S1,

(66€#

(ygns

998

OSd 0},

JILLTASNO

SJupuosu fo “VWWd o) Yx

NVILdADA

joasukuvT

“yeoyo SOT 2q CPL podeios ‘oreq 0}, MOTd , W191) Aaya { Joy + Si, (SNJJIP 2 > Vdd‘ex

0},

Sad IY,

-MOy-4/-,eq-% ssey-y

uPeq

~S2Y-%

“8CL

0}, [SeJ

“FQU-% 0}, odeios.JJO

“LOL

OLLINAS

key

AT £sq

JICVHO

+

,UIOU}I,,w3}S)

UI, (Jgns

BYDOI 18 01:8,

‘

JLILOWO

3 oe SH

(‘inp

0}, 3nd JJO {(peoy) 3nd no

‘USIS,,

‘yeoys
wiajs) ,,(194)

Wy,‘u

(‘yyns

-OS ud (eur)nAny, wiajs) ,Aaunof,,

+

“OJUT IND O},,

Cygnus

-ul Ay W9}S) + yys NVILdADY

TAY

O},, [ABD “TOL

Px OSd) Uses

JILIHSNO

JIGVHO

Slupuosuo, {0 “WWdYu

OLLOWO ~

joasukivT

Oil

JLLIWNAS MYx/AYx Sdd

“peed. OSd

ie

€@ 29 JLLOWO yousg :WON

31q,,

JIGVHO

Yx -WW¥d fo SJunuosuo) jvasukivT

-WI9}S-QI9A

INOS JILIHSNO

“Pou

9q “214 O}1,, x

OSd

Z, = > “O"l

NVILdADH

“IOUUUI

dzy

jo Od OSd) Px ‘suowur dy, +> (Ayreuy 01, (119 Joyjeo

w9}s) 0}, ,Uopres.,, SIp | (0 Fr)

im Iy pouinsaid -BAQ]9 was) ,Punois + “yyns Jejzey JILLINAS ‘IN ‘YSNOLI,, ,[TJ Jy poy 0}, [pe>°A uzey Vv + Uy UWId)S) +‘U (yyns ‘{pe rJorpeo “(,T1,,

UI JUNOUIP 9SBAIOUI

“ZtU-% “EOL

JO O},,9ZIS

QUt) uC

-ZOY-»

0}, SIP “POL

yx

“LOL “Yx

“BY

[Px

“suoWap Was

(AU

‘suowap) ‘(aseq ur) "Vy Ul Jeou *suOW Ul IOp DeLIA UI S Je Jey)/STY,ou0 },, pue Ieau-Op (‘suour

JYI/STY}

pneq‘y 164

AD NVILd

+

JNO

(jd

JPU},,

y BY sty}, ,,2U0 Sd) x BU* OT TV ye “( OSd NY,JO Any,

Yea, u21Q99J ‘fpe) > ‘A Aq A < (AA

SAQPUIOIe

SULIO; UT (4,

20]@M-POOysy,,

SnotA ‘A +aid) (Jyns

OSd *-$22U% ,,UdIPTIYO,, W9}S) ‘-Laay, “-S2QU IO ‘UI, ‘JJNs -122Y, +

Moo

1, “u

1ey

“SOY,

[tds , 1014) :0861OTE YT

OSd *-l04x ‘-3OYy “-LOUx

JILTHSNO

stupu fo osuo -WWdY%

399]]09 ‘oye 01, ‘dup .dolp Oyeul & -I9Sa1 IO WId}s) + Ay -ul IIOA | WIO}S) + ‘yeoy < o ‘(np 04,

uCDur) 0},

wom

-32Y-, JO -Joay-,

-LOY-»% 0}, [yds

0},oq

VVdUx

“99/

“SOL

JLLINAS

joasukuvqT

("919

‘Aey epAC‘-ey

‘suoulwax uvy ‘ oe

,,STU},, aseq) Ul Jesu-Op

“BY,

JLLOWO

8LE

Oi

(A A/ M BYOOW < G@x

BUOTE O8 0}, ~PeRy

+ UI9}S) Yous

"WIOS < ‘SNBd Sy,

® ILM OO

-OUIOS 189q 01,

Od *S05JIP I»

(vey

Y

(inp
*irtt [*t >t assimilation]

> i3tt¢). But on the whole the exam-

ples of medial weakening involve only those penultimate *r where the ultimate consonant is a suffix of high productivity in Middle Egyptian, normally the suffixes i~ y, w, ort. These cases are thus suspect as instances involving affixation subsequent to the operation of the shift *r > 3 /_#. In other cases the affix may still have been felt by speakers to be only optionally present and therefore still indirectly indicating a potentially word-final affixless environment; in such an environment the *r > 3 shift would have applied, with generalization to the case where the affix did appear. Even the word for "milk" might be considered a special case of this process — its particular history giving it the appearance of involving a double suffixation of t. The reflexes of PAA *1 were only somewhat more varied than those of *r. In PS, PC, PCh, and PO it remained *1. Newman fails to account for it in his reconstruction of PCh, but Jungraithmayr and Shimizu (1981) provide a clear demonstration of its presence in proto-Chadic. In the Gonga subgroup of Omotic, *l became proto-Gonga *n word-initially when the following consonant was labial, and proto-Gonga *d in other word-initial environments (see Mocha examples of *1 below, and Table 1, Chapter 2). But in the rest of Omotic, and non-initially in Gonga, its reflexes are generally //, except for assimilation to /n/ in the restricted environment of a proximate nasal in several North Omotic tongues (again see Table 1). In ancient Egyptian, a distinct /l/ was not marked in the script. In the early written records, PAA *1 produced Eg. r in word-final position and occasionally in what would have been final position, but for addition of one of the

highly productive Egyptian suffixes mentioned above. All cases of wordfinal r in ME and LE data — except those of gr, mr, nr, rr, and wr and of &r

and xr noted above — can thus be taken normally to represent PAA *1, since *r would regularly have yielded Eg. 3. Word-initially PAA *1 also yielded Eg. i (*y) in instances where the following vowel was *i. In all other envi-

Laterals and *r in PAA

|

391

ronments PAA *] appears normally as Eg. n. In at least one word-initial context, preceding a PAA *b (#_Vb), PAA *r would have fallen together with *1 in Egyptian, because it produces the samé outcome, n, as does PAA *1 in

such an environment (roots #927-930).

The Egyptological literature on the consonant *1 often holds that Egyptian had a spoken [1] but no grapheme to represent it, and so relied on n, r, and i to fill the gap. This seems extraordinarily puzzling, given that the scribes did not adapt their script from that of some other language of considerably differing phonology, but evolved it on their own; that they seem fully to have represented in their script all the other early consonant distinctions; and that they seem eventually, if belatedly, to have taken account of most other changes in the consonant system of Egyptian. Moreover, the regularity of the orthographical correspondences revealed in the comparative data presented here — - coupled with the fact that [r], [y] (@), and [n] are such natural, common, and

unremarkable phonetic outcomes of original *] in the described environments — would make far better sense if the graphic correspondences were phonetically rather than scribally motivated. And in fact, a reexamination of the Coptic outcomes of PAA *1, *n, and *r, viewed in the light of the correspondences established here, compels exactly that conclusion — that ancient Egyptian orthography reflects the actual phonemic history not of just a scribal dialect (Vycichl 1983), but of the language as a whole, both spoken and written. In words that were not borrowed and that are certainly traceable back to Middle Egyptian, Coptic (and Demotic) I is as likely to derive from PAA *n, *y, or *p (which fell together as Eg. n) as from PAA *]. Moreover, Coptic / appears as a reflex of these several PAA nasals and of PAA *1 only if one of a restricted set of consonants co-occurs with it in the same word; otherwise both PAA *1 and the PAA non-labial

nasals yield Coptic n. These data thus impel one to the conclusion that most PAA ¥*n and *1 merged as actual n ([n]) in Old Egyptian, with new / arising subsequently in Demotic times by split from such Eg. n. Other / entered the language in word borrowings, principally from Semitic and Greek, completing the rephonemicization of /l/ in Demotic and Coptic. Two periods of sound change lie behind this phonological history. The first period started off with six sound shift rules dating no later than the beginning of Old Egyptian. First, Eg. #30. PAA *r > pre-Eg. *1 /#_V(V)C[+obstruent,+labial] (#927933, and 941). A shift of more sweeping affect then removed most word-final *r from the language:

#31. PAA *r>Eg. 3 /C_(t, w, y)# (C#9,m,n,7r,w, 8, or). It was followed by

392

Proto-Afroasiatic Consonant Reconstructions

Eg. #32. PAA *1> Eg. r/_(t, w, yi,

remedying the defective distribution of Eg. r created by sound shift #31, but conversely beginning the elimination of the phoneme /1/ from the language. More PAA *1 were removed by two other shifts, #33. PAA *1>Eg.i ([y]) /#_iC (instances of /#_ib and /#_ifC have been observed so far:. #808, 809, 834, and 835), and

#34, PAA *1> Eg. r /#nV_C (#633). Finally, a sixth rule,

#35. PAA *1>Eg. n, eliminated from early Egyptian all the remaining *1. These six rules together account naturally and succinctly for Old and Middle Egyptian orthography — for the shift of most PAA *r to 3 word- or stem-finally; for the realization of

PAA *1 as 7 in the same environments; and for the representation of all other PAA *1, and some word-initial *r, as n.

A second set of sound shifts, dating to the period between Middle Egyptian and Coptic, then completed the process initiated by shifts #30-35. First, new / were created from ME n in a variety of restricted environments, most of them involving a proximate voiced labial or another nasal in the same word:

#36. MEn

>I /#_Vb_Vb (single instance, #810, but no counterexamples known);

#37. MEn >I /#{S, h} V_{m, 1}, /#_(V){s, §, h}V{m,b}#, and /#_V{m,b}{s, §, h}, where s, 5, and A are the attested Coptic consonants (hence in #401, 629, 659, 661 (first

entry), 810 (third entry), and 827, but not in #661 (second entry) and 817); #38. MEn >I

/#_Vk (#619, 822, and also C. 1k6 "sycamore

fig" < ME nk Cwt); #39. MEn

>1 /f#C,V_C) (either C; or Cp = m; 316, 587); and

#40. ME nn >1(V) 1 V (#637; single example, but no counter-examples have been identified).

Outside of these instances, however, all ME n — whether deriving ultimately from PAA *1, *n, *n, *y, or *r — normally produced n in Coptic (for Coptic n < PAA *1, see roots #442, 484, 803, 804, 814, 817, 819, 831, and

Laterals and *r in PAA

393

969; for Coptic n < PAA *n, *n, or *n, see 48, 198, 323, 366, 391, 448, 568, 589, 590, 609, 611, 616, 620, 621, 626, 627, 631, 685, 689, 690, iGnes 970, and 983; and for Coptic n from PAA *r, see #927-930, 933,

an

é This history of PAA *1, *n, and *r in Egyptian has the added virtue of providing a plausible and simple explanation of one seemingly irregular source for word-initial / in certain Coptic words. These particular / replaced the grapheme for 7, which had been used in the Demotic or Late Egyptian forms of the same words. With one exception, none of these words can be traced back earlier in Egyptian than some point in the Late Egyptian period, and most are generally argued to be Semitic borrowings in which the orginal consonant was PS *1 (Cerny 1976; Vycichl 1983). Given this latter information and the phonological history sketched out above, the explanation seems relatively straightforward. Initially, in Late Egyptian, r would have been adapted to mark the rare borrowed [I], lacking as a native sound in the Egyptian of that period; and then, in later centuries, such [l] would have become

part of the new post-Late-Egyptian /1/ created by rules #36-40, and so would no longer have been represented by 7, but by a new symbol encompassing all of the newly phonemic /I/. The one case of Coptic word-initial / which truly derives from PAA *r (root #926) is one of three examples (another notable case being #679) attesting a separate pre-Coptic sound shift rule in a very restricted but sharply defined environment: Eg. #41. MEr

> Coptic / /#(C)_o(C)# and /#o_#.

Of the remaining three PAA laterals, *t was surely a voiceless lateral fricative; the correspondences point clearly and consistently in this direction. PAA *dl and *tl', on the other hand, were most probably articulated as affricates. A minus-continuant articulation is indicated by their reflexes in all four branches of Cushitic: by the Beja retroflex stops /d/ for PAA *dl and /t/ for PAA *tl'; by the Agaw voiced stop *d for *dl and the ejective stop *t' for *tl'; by the PEC stop *d' (either retroflex or implosive) for *dl, and PEC *} ({'j]) for *tl' (Ehret 1987); and by their maintenance in PSC as the lateral af-

fricates reconstructed respectively as *tl and dl. The same feature is attested in Omotic, in which *dl and *tl' fell together as the PO implosive stop *d' ({f]) whereas PAA *¢ and *1 collapsed into PO *1. And it is paralleled in Egyptian, which changed *dl and *tl' to the respective palatal saffricate or stops d and £, but PAA *¢ to the equivalent fricative palatal 5. Only in Chadic, where the two consonants widely yield lateral fricatives, and in Semitic,

where PAA *tl' shifted to pre-PS *s' (*s) — filling the slot in the consonant system vacated by the Boreafrican collapsing of PAA *s' with *s (NAA shift #1) — are continuant reflexes notably present; and even then the PS *s may actually have been realized as an ejective affricate [ts'] (as a variety of scholars have argued). Newman's published reconstructions (1977) posit only *t for PCh, but

Proto-Afroasiatic Consonant Reconstructions

394

Jungraithmayr and Shimizu propose four laterals, which fit very well with the

four laterals of PC. Their PC *l corresponds to PAA *1, their *4 (or *4,) to PAA *tl', their *42 to PAA *¢, and their *& to PAA “dl. As with their reconstruction of the alveolar sibilants, one of their reconstructed laterals, *¢ (*41),

acts as a catch-all category into which a lateral fricative or affricate is placed whenever the data are insufficient to establish which specific lateral is involved. Thus PAA *¢ is normally reflected in Jungraithmayr and Shimizu's

*4, but occasionally in their *4. The evidence cited from Ngizim, finally, is

quite clear: PAA *dl yields Ngizim /dl/, both *4 and *tl' become Ngizim /tl/,

and PAA *l commonly gives Ngizim /n/, but sometimes /1/ (e.g., root #804;

the factors conditioning these particular reflexes remain to be worked out). Table 8:

Afroasiatic (Afrasian) Lateral Correspondences

PAAR om) Poe

iiig

anak

PCh

PO

narat

*]

*] (Ng. /n/)

*]

“lewd aly thd

*d

*k (Ng. /dl/

*q!

*f

5

*4

*4 (*4) (Ng. /tl/)

*]

t

“1 ({t'])

*t(*h) (Ng. Al)

*d'

*]

“I's

*]

*f

(Note that the notation *tl in Cushitic symbolizes the ejective lateral affricate [tl'].)

A fairly extensive co-occurrence constraint in Egyptian limits the variety of possible consonant sequences in Egyptian involving the PAA laterals. In particular, it disallows lateral obstruents (or, as they became in Egyptian, palatal obstruents) before following velar obstruents in the same word, with the result that PAA *¢4 and *tl' became Eg. s (rather than 5 and ¢ respectively) and *dl became Eg. d (rather than expected d) in such environments. The constraint apparently affected some but not all sequences of the opposite order, with a velar as C; and a lateral obstruent as C2. For further discussion of

this shift or set of shifts, see the materials on Eg. rule #3 in the section above on Afroasiatic velars and labiovelars. Another Egyptian sound shift of consequence for the laterals,

Eg. #28. PAA *tl'> pre-Eg. *dl > Eg. d /#_VS, accounts for the absence of the sequence ¢ © in the language (roots #922 and 923). The distribution of PAA laterals may also have been subject to a very an-

Laterals and *r in PAA

395

cient and widespread constraint against the co-occurrence of any two nonidentical lateral obstruents as C, and C2 of a root. Whether this restriction existed in Omotic is not as yet known; perhaps like the constraint barring the

co-occurrence of non-identical labials (discussed earlier), it did not operate in that branch. But it seems on present evidence to have been a general feature of the Semitic, Egyptian, Chadic, and probably Cushitic divisions of the fam-

ily.

,UMOp

-C2T-» O¥LNS

Uy,

0},

1,

“INP


‘Ol,

u ‘01, ,Je -FT°W) “If ¢-7 (919 (-u

~Tx

NVILdADH

~TT-x/-B Lx

“S08

“708

‘Te, 0}

“£08 Ble

(anu ‘9d

VVd Ix

JILINAS

spsaqv punTdy Ul

ye

“LO8

0},

Is

-QTT-x JO ~QTTT-x

1024) (6861

0}, dey ,dn

“GIT

“808

0}, de]

1q1

SITU}, W9}s) +

Ajqissod) ‘u >

AISITU},,,Ueud

20),

OLHSNO

/WWdlx

,S4011NG,

‘U)> ‘A hq (GAA -ul se ‘dnpor

“Ip 1, + “SUd}

,,dedy wiaqs) 0},

(#7 $°O 6s O4Ad)

10 -mbUT-% “718

.qeads 0}, ee J-~ -mb

ABUT» “ST8

“Iaq]e 01,

qefery

AQuanb ut

zbel-V

({) JNI 01,

chequ

,UINJII O},,

“Heel, DAd

(‘bary se ‘dnp -a1) ,yooods

-d1J SPIOM yeadar 0},

ferfey

(was ‘dnpor A][ng :,,JoyUeQ

‘(‘pud}X9 Gy + WId}S) , JOWIPIO ‘jno AIO 0},

b6u

JLLINAS

Ol, 2u2Q0U

2), (Dou se wia}s ‘dnp Yoed 01, = -O)_,, AP

BPTxOSd “mB

NVILd ADF

(LAA < A Aq’‘a > ‘U) SMU,

JILIHSNO

Ix “WVd U1 dx DUD S]DABID']

“LT8

“YOl-x 0}, ‘dst08 ,dn

1zyu

OILINAS

+ “WOJ ‘puay -X9 ZP, IO Jou -UBU JO “sud}

zak 0},

-UI Zy + WId)S) iy

mdyu

+°A -OP My

ind gno Mou. “sjooys(d30

“WVdIx

JLLIHSNO

OSd “SEYOTx 0}, nd 3NO AOU ,YIMOIsWId}S)

in ee! “~11ed a

sjp4a10T puDdy Ul

NVILdADH

“(wos “uly dy + Wd}s) Sul -UJOUI OU} Ul Ajred OSU 0}, dyu

snotAoid) ,3ut -uow ATIe9,

‘((QIOA

Ay “yeoyout ‘SD (3sy3u

JIGVHO

JILOWO

lO

JILOWO

JIGVHO

(-Ya]y JOIOSUT

sey L861 314g)

JNO dUI09 01,

“OT “OARS

(daay ‘1asuep WOI, MPIPYIIM 0},, wynu ‘) “puslxe Wy + Was) ,,({J19S -9U0) MBID »

“WIM {Jyjo Aled

wiu

‘KEMP OAC} O01,

(94M YOO]

o1,, Yst[sug

({ pued

-dy) ,, premol QUIOD 0}, Ulx Sad

(‘ulj-uou Uy + WId}S)

,puayoidu09

Vv Zeqel

JNO QU09 01,

-Yo2T-x

-Y2I-x

,uonUaTe Aed ‘a0n0u 01,

+‘(pusgix9 Oy

‘pueisiopun 0}, uyey

wi9}s) , puayoid -W09 ‘30T]0U ‘presal O},,

OLLINAS

"JO :SOTJUBUUDS SUTJ [x + WSIS) ,2[QuIaSoL 01, “Y2Tx OSd

NVILd ADH

Iu Ix (WWd Ul dy PUD SIDLIIOT

OLLIHSNO

-YOO] 94} UO aq “IOJ USISTT O},,

JO} INO

-Y2Tx

COV

“618

“818

0},

puod (1

01, ,,dey (6861

Sad 4Tx

oy),,01 Indon3uo}

“AP

S¥OT‘D

Hoy(¢

WI9}S) + \S%

Ways) + 3,

~JeFy

(UOUIA}S

sajou ATuo

32249)L861

91qIS

(ued

JIGVHO

,9nsu0}, (¢)

eu0s YOMLx

(PAL 01, AH“f) MOTTRMS WI9}s + Z,, -UI

oreyed “SexYe]

FBT, - JO

yep, &Set -Sod -

rary”) 01, “O11q SUNS,dn 0}, dey

-dy)

0}, dey dn

Odd (94 < O},,

vAep,,

JO *SUud} -UPUT ‘(19U

“CCB

OUd

‘DS

Od)

“WeI-* 0}, poy GUA oy} ,on3u0}

(urys

(ZU,

,UeTT,, -o1d)

poumnsPIO ‘3q

~98Tx

Ix Sad

‘amy -ZPP] JO) JO 143T],, -[e[x w(Aep “OZ

ffo 1q jo

Su

JLLIHSNO

ZPx

“TC8

+ J Ways)

Syed ‘a'r wing

,,dZ1

‘puriq -Ioyned

Vv 52et 0}, suInq

NVILdADd

-VV. Ix

OLLINYS

sjpsaqvT puDddy, Ul d JILOWO

COV

(‘suo]UT “Inp Jy + Ws)

‘paye} 20110)

Ay,

‘yeoyout
,Puly

‘SAIOIOd, "31 “-1]-op Jou Inq "snevo-op :son

-UBU9S *yeOYd

,PUTJ SOATIO

-ut Ay, + W9}S)

-Iod ‘19)uN09

Abny ‘Abry

-UBUI JO *Sud} -ut dy, + Wd}S) ,Appomnb pusy -d1du09 0},

Jber-y

“CCB e

~FRCL

(WWdlx

(u, BBU0D < YSep

/U/ Tx 0} inoge YIM OSdO},,ITF Wy

“Jyns QJ Yx “InP Was) + 01, .MOys

“WISSe “AUR “NOge

JO -,4¥Tx

JILTHSNO

NVILdADA OLLINYS

210JOq QAOU ‘jnoqe JO ,Apjomnb ~OIGVHO

:esu0yH Ov'S

03 01,

-WU

JILOWO

Ix Od)

vOV

JIGVHO

Jd

>

Vx

-SULTy

(‘pudxo Wy + WId}S) ,,(ssou

Px + WdIs)

uo daox “RIS

‘(Aq ATa0[du09 03 “d'r “sud} -UI ‘INP 1, + wi9}s) ,ssed 0},

VSI

9q “IJOs 0},

NVILdADH

iqel Vv

oteIs

pousyeom

SSUCT-x 98

“SZ d2]-% -

0}, ‘au090q (amp

OLLINYS

.duary “yea

(dx

Odd) ,,suoye

OAOU 0},

~.99Tx OAd JILIHSNO

Pax

QUW099q 0}, UL “yeoM oq B

Suoye

0},

wesel MOIS JUDTIS

JILOWO

Ix “WW Ul dy puD S]D4sIIDT

AOU

-LB3M WOI{)

Vv 0},

COV

x

su ‘Od

‘USP, JOyIeI

-Iy/-Cx

+

GI3A (1001

-o1d

-SIp

:,,9nsuo},,

-S8T-~

0}, POY ,01U0 “878

Sd (9%

dio VSISx >

W13}S) + Xy

VSIA XSTx

0}, 103 PloyJO

WIM‘A

SIG =, SIs JO Ay (U>

aloqiy]

JILTASNO

!VVdIx

Od Sx

OeS >

GA/

opeyeq“Sey

-O$

071, [Ind , -yeq)

“OYIEIOU ‘UHI + ““qiaaop

‘[nJYyINOUL,, 0}

wey} snf UST], SB UI AOUOYPIC] °C/6T WIS + Ux 01, dey dn ‘u (ygns Spucar tsdep-Z

~

10249) ‘(6861 UST, ,,9N3UO},, UST[,.)

Sdd SIx 01, OI] ude]

qsu

poseys uoqeALap ‘raqiog STI ‘JoidUSTTx® pue oues :UOljBAOUUI

SI ,on3uo},, Sd ‘u OY} ayeredas *3q y/-SeTC ‘quae “YO)pue ‘Wajs 10unWO 0},

wl]

(ser ‘D

“StI-%

NVILdADA

[(-stte

sloqry powed = ‘DAD#/ oWOsS -x9) osensuel « ueo, = -aud powunsaid)

TT

syosaqwy pundy Ul

Lf (€S]-x

-e, “QLaIe "U ‘Jad + (ui9}s

,903U0},, “N)

Se]e,

JIGVHO

0},

wT]

-WOS sud ~, SUIT

JILOWO

907

(,,9ZIS

:UOIBAOU -UL,, ,,9SII9 < MOIS,

+ 0} De *(9

UL

(U1) “urag) “8g YQ “ul

“1E8 XBR» 0}, OSBIIOUT

“MEET-» 01, yordudn

JM

Qu

(INP

Ut9}S

“QATQNOI

+ Jy “ION
‘somUeUaS ‘IPx Od > LP] = x) ,Yy3sm,

—-ebbox IeJy :Od

-ba TPs

,0} UDIST] O},,

(d nyeexX ‘U) ,,MOPIM,, 9U:N,1 BYDOIL

JIAGVHO

GCA SSNS

1¢p

x OSd 0},,[Px 2q P21 Od) OSd (De
‘dn 08

ae]

-ul Ay + UWId}S) ,dn o3 ‘dn IT] -OSTeI 0},

(prnbry oj 10

(‘yeoyd -UI My + W9)S)

MZUS ‘VY

w43Ty 2q 01,

CSNJJIP Ty

+ WI9]S) ,, SUITS

JTes

“UI ‘pug 0},

*(,,UMOp

PoMmod,, Jol[Ies “ < ‘AW [Py + WI9}S) ,, you oy) pusg 0} QUO 9SNed O},,

(, SSQUOT],, “QUI x

,UIS,,

jou *JyNs ‘u

:,,predogy,, pug

0},

Sd Jo ayeusoo

uy, + Ws VVd ugel, “UTZ ‘(OJ

([,,Wuey,, “Ol UL > S°A°

YOM ‘f) ,4Or, 6, + WId}S) ,OZeI ‘AOQSOP 01,

$,Q190eS

-Bges -Seul ‘sIaquinu

Aqu

-Jod Suny ureul -d1 ‘\stsiad 01,

oq,, + ,ADONSIp syoulsq] “A + Jolie W918)

NVILdADH

JLLINYS

‘pason ysuod ‘ “Qni-

; (ou, egnp' Uy, +W9S VWYq

-O

JILIHSNO

Aq -n-,, “11)

0}, ‘AOQSAP

01, 9q Sop “L76

BYyoO| *FyNs “u

JIGVHO

-ged ‘SN

ead UL |IPy O1,,

-U-qnp, B3U0H

JILOWO

w2lp

IJV

Jy SWWd Ul dy PUD S]D419IDT

-Qel-x

"@)

pqu

[Px (AU + ,,SISAIS WI9}S) “(gus

puz) ,2sdi09,,

Cvy

ged

"676

0},,

SUTSIA)

“qOOl-x Sdd Gx uaiysn Aq

0},

wD

(6861

yoIy_)

gqu W9}s) + ‘UI,

Cygnus

-qool,, 10 ‘A> Aq

Aq‘a

WOS /-9:0P, “G:9P

JILOWO

° 6

(00,

vvY

0},

gypsy ‘Idi
‘A 0} QUITS *(SITeqS

-UI-M-Iy Wy

NVILd ADH

Jy SWWd Ul dy puv SjDAIIOT

JILTASNO ,, S-OSiT MOIS

“Pb6

00F,,

“ZTTIy

“ZN I-x0}, SOT “CV6

-UNI-x 01,, ons dn

2U1099q

“Ob6

,, PaMoul WI9}S)

+ Gy ‘(‘pudg}xd

VV qces 0},, 3nd yoy Ssvis ulese Jaye Suroq

‘puoddy) (1

‘dn uIOM ‘no Iso] ,UONIUN;

0}, 393 pasn

Sqd ZI

OLLINYS

(uon

JILTIHSNO

SWWdJy

OSd

0}, 393

“Ud ‘duuny,, ‘dung punour ,

ISO]

‘readdesip

y/-ZT “ZERIT uLId}O0j,, QOS Jo ,,1003 -dV) pued ‘7’ $19901109 9144 -O86T (6ZE

TI

NVILdADA

SIDIIIDT PuDddy Ul

pi3q sty} ,J00F,, “1 saouap IIQIS O6# | Wy .

JIGVHO

OSV

YISUZT JOU 194 ‘poutey| Od dxa ZZy< BYOOJ

-juasqe ,.popuru A)

OLLOWO ~

eyooyw -[fop

0}, 3q

(ff

°I8J‘D) -sod -MOA “que Joy10q QdINOS-uodsa1109

saneuiayye 1NQ

(14°D)

({ ‘puod

wep ‘uns, — -dy) ,, Si O1, 34% Sdd a4

.Pop

-uadsns aq 01,, - 5ees opeyed :OS

34x Sad

uPastel oq 0}, “SUC “876

“SeCI-x “L6G

=,SUOTR SAO 0},

(‘duos "QIOAOP Sy

STI

,Peoy,,

JLLINAS

aS w9s)

({‘puad -dy) ,d9A0Ur 0},,

NVILdADH

wMOTTOJ ‘Auedwiodov 0},

-SeUdy

JILIHSNO

/WWdJy

|

JIGVHO

SjD49IDT PUDdy Ul

JLLOWO

Icp

452

Proto-Afroasiatic Consonant Reconstructions

THE PAA

GLIDES

Finally, the PAA consonant inventory contained the two glides or "semivowels" *w and *y. Both have been retained quite generally across the family and so are easily reconstructed. In one environment, medially in triconso-

nantal roots, *y and *w coalesced with a preceding short vowel in PS, Egyp-

tian, PCh and presumably Berber, but remained distinct in PC and PO. This

development was posited in Chapter 3 as a shared innovation, PNE sound shift #1, supporting a North Erythraean branch of the family excluding Cushitic. Medial /y/ and /w/ are of course again today common features in Semitic languages, simply because many new medial *y and *w were created later on in the Boreafrasian subgroups, via the addition of extensions and nominal suffixes to biconsonantal roots of the shapes *CVy and *CVw (e.g., see root #836, among others, for a shared Egyptian and Semitic innovation of this kind). Other *y continue to be produced by modern Semitic morphological operations. Jungraithmayr and Shimizu (1981) reconstruct both *y and a less common glottalized glide *'y, although it is not clear from their data that the sound *'y is actually to be traced back to PCh. A number of modern Chadic languages, among them Ngizim, also maintain both /y/ and /‘y/ in their consonant inventories. This distinction is a reconstructible feature of proto-NiloSaharan (Ehret, work in progress). As only *y seems traceable to PAA, *'y probably thus originally entered Chadic phonology via Nilo-Saharan loanwords in early Chadic. In Ngizim both /y/ and some /'y/ can derive from PAA *y; a possible conditioning environments of the split in Ngizim may be the following: PCh *y > Ngizim ‘y /#_VC,

>Ngizim y /#_V-#.

But more data are needed to substantiate this suggestion. One additional glide, *h¥, may possibly have existed in PAA. But thus far only four potential examples can be suggested, all with the hypothesized consonant occurring in stem-initial position (#1012-1015 in Appendix 3). If valid, *h¥ would account for PO *h2, the Omotic outcomes of which are /h/

or /w/ in different of the Ometo languages, /h/ in the Bench and Maji subgroups, and /w/ in the Gonga and Yem subgroups (see Table 1 in Chapter 2). The possibility that such *hW might trace back to PAA is at present hardly more than a conjecture, although one certainly worthy of future scholarly consideration.

O},,

"A)

Sey

OY} *UIAISPA,, muqm

asu,, ut ‘dao

-y1]3 ‘(uns Jo)

-UI SB “UIJ-UOU Uy, + Ws) 19} SUIYS

ysysuq “3°3 , ouTys “ung

O},, JO} SQIOA WO dALap AT -UOWIW0D ,, MOT -[a4,, IO} SpIom :SONUBUIOS)

wTOK “MOTTO, gooM

TTeeuloog ‘gq

( 89M

MOY,,

NVILdADH

ot he

AM

P» + Was) uPOPTBdS 2q

‘(inp

wYSLq 2q 0},

mA + WAS)

‘OSI

q4% YOU

My /PM MESY

ud Jey,

eM,

“ 604M,,

(f) ,JSBOI 01, (N)

0},

-My

"3 Sduros

9qM

uqm

(Jgns “U 1, + WI9}S) ,,UNS,,

iGleyM,,

mojjak

OLLISND ~

(‘905

x + Was)

JILLINAS

wCDuT) wing 0},, “Q29oM-% “OS6

uGteyM,,

My VVd

“616

pue

JIGVHO

/-BM%

¢‘¢ Jego

JILOWO

uGSP]F O},, "Vv S$qem

404 (yI0q

Mx «SAPD VWVd

“IM,

:wON

:(, Jo) (uns

pqm O},, ‘ung ‘yeoy

CoV

yousg

ASM

ADA NVILd

+ W9}s) “{pe Mx ‘[njiamod [x ts [enue JO ost 1, (‘qIOA ‘SUONS,, ‘Ayppeom(jgns -nygur -op

“VY gnznmn

, Uy, Ajouny,, ‘prey poppe “poystqeiss Suruesu ‘Ajaane “IRM-» “9QM-% ‘aouaY 9q ,,PITOS -uoly 0}, “3 (,3u0ns ,ssed 0}, “TC6 “CCG

“UI2¢) -BAOUUI -In3Ij "Dd

(ujqsOM

USM 0}, sonoeid 0}, [DARD -JOTJ INOYIIM-IQ\UI SAT oq -uryun ‘uondni -uod polop ,, WI9}S) + Ux [hx +“OOF ,onuy WI9}S) + “UTJ-UOU -Ws SonuP JO ‘90J Ieopoun “3x9

JOM, 0}, 9q ‘Pexly “ULI “PHOS uPOYst[qeyse WIS) + ,Yx “UT “sud}JO (JD9JJO

OLLINAS

VVd

Ale=

0},,

«SAPD Mx

Od (9%

,SSOID Sd)

JILIHSNO

JIGVHO JLLOWO

qy (‘puaixs

+

Wy

days


(‘UIj-uOU

U,

uonrppe jo

,ddd]S

hq

“A 0},

:UOIJVAOUUI uOneALIAp jo

“Jyns °N) JOTpIva = + Uy,

nyeex gq >

onyeex

Jd

nyeex ngogo ,,WIBO;,,WaIs)

,

'-U-S, WId}S (‘Ulj-uOU

win}
-UIN]

‘qioaop

MOq]9,, ,, W19}S)

+ Ay.

:SOMURUIIS

(‘Inp

‘WON Wak BhIOM

JILOWO

{Jyns -W9s

Suljo

(‘pus}xd“u

Wd}s) + Y~

/-YIOM ,. JIM-Y 0}, 03 ,Punol

JIGVHO

“LL6

+ qy

balz + ‘Ay1,

sewajs

JAIIM ,JOUeYO,, -O1)

‘anbrjqo ‘dnp

,Teuoserp wia}s)

‘4ueyse

“VY QUsem 0}, 3q ‘AIMB

JILIHSNO

-JIM-» 0}, WIN} UT) ,(

NVILdADY

9/6

OLLINAS

VWVd -SOPUD My vOV

UT “uC #/ (0

C9” yousd -WON

‘ 3SS@M'd

SM» Sdd

-SARM-»

(Or# “1661 34H $]091109 £,SUT]

,(8ZIS UL) BSVaID -UI ‘MOI8 0}, Qs M

0},

“SoM

-[OMS ‘][IMS 0},

Io -sA@my,

d0UIY

quam

,JSs UO

J + W218)

JLLUNYS

,Puedxa

ISMy

wa}s) ,,yIpecsq “Opim “peolq,,

(, W1007q,,

WI9}S) , JOMOTJ,,

o10y,, -UM ‘Dd ‘inp se ‘dnp

-O1 W9)S) YIM uum

ulSTX9 *9q 0}, NVILd ADH

OST® 1001) UIs€36# BUI ‘UB(fpe A[qe ‘sneo se 10 YIM “WoOusp UdEq -qoid UI + s,

“P86

‘uon

‘ureids -dajUr 0} oNp dn ][aMs 0},

‘-sh@My Ad

‘guI099q 0}, -uARM,y

“SNM SN

(f) TINy,,

OLLIASNO ~

u

JIGVHO

MOIS

(pIOMUvOT

Od” eIqQtssod)

,2U1099q 01, -u:0 * -uupM oud -wOs

‘(gus “pe A, >) 31d,

+ ways ‘eA- +

-UARM

euuT

wok -WON JLLOWO

My -SAPI1D Vd

“E86 -UARM-»

ISUIS Gy MOIS, “UWTUR “op *(,,ST

-Aey Jol sny} WV ABM, d-aid

< d0udyM -BAOUUI ,MOId,, Ux Ud9S + 86 ,,‘dAT],, UOT} J00]I) “WIS Ul ‘aq, (,1Stxo “8g 01, O18 SUulj-uou “3q

OST

0}, ,puedxo qIOAap “WIUe :SONUBUIOS 99S

(inp

0}, TI9MPp “WV yuem

UL “A > “QIOA

“

LOY

“986

SIYO

“ZIM,

-(RM-% IO

‘Tes,Tes

ppm

a 0}, sy

wrem

‘(sng

WIS) + Ly FIP

+

ung

‘MOJZ “CEE

JO) u(21 My

““barj w9}s se

w191S) JOY a

rem 0}, ({) ISBOI ,,(uleI3) wi9}s)

‘ueM

AIDA,

Ww, “fpe (yyns + A,

$jOOI

n(e)n(e)m

‘yeoyout
,YSIM “O'T) yyIM “dnpor + uy ‘u JsTpIe9yWeayJO) ATqtssod Joy (IOATT SB 9S Jo (suonsajye sind50 Url (DS

-V ‘ppnma ‘PPM ppem

:0861 (CTE

‘(np

jo ue3ZIo,,

n, 2M BRI OS

se “OSANUTUIp

“C86

“Inp

“SUdjuT < (‘inp

W1d}s) + J,

NVILdADA

-SOPID My

/-UZIM, Odd

OLLINAS

VVd

>

[Da

ey, +,

:WON

-UZIM,

uevO]

0}OWIQ]

JILOWO

897

-I09)

JIA ,,dYOus,,

wia}s) < ‘SNIP wa ,,Aled -J(OM :WON Ix O},, +

58M -JOMUY:WOS © ‘(¢-m dn “My20109 BYDOJ]) puris youog WON 01, 0},,

697 JILOWO

(amp

03 _,,08 (INp1, :adeus ‘Aq ‘AWIOD IM,ssed puz+ Was

yoreul

WT 1¢M 0}, £03

(Jus

“IO Jy + Ws)

(inp

uPulq -qino ‘onpqns 0}, Jo 4 HT

< ‘yeoyout Ay + Wid}s) ,JSuTese

16M

“886

“SUM-x

‘OARI]Ie1S ‘INO

“U8UM-x

pue Ioy}es 0}, ASEM

JO PlOY ayel ‘dsvis 0},

w4Woy 2WI09 IO O3

“(Hoy “Uy

id, + W1d}S) ,,]]e JO POY 9¥¥} 0}, q5eM “VY

0} + -W9S OSNJOI "UL ‘advosa ‘advosa Was) ind ud}YysLy *SNIJIP ,0}[x 0} + DARI [fem UaIsey :soyue ,ABMe WIS) 01 ‘yyIM 0}, AQ JZ@M0}, 1, “V (QUsTg

“186 0}, 393 0} dn

Alatf,,

IgM Eom

Cygnus

(SUs}UL < “UZ + W9]S) ,,d9]F 01, lcm

uo

"U1, + Wd}s) ABM ‘peOl,,

-[BCMy

ae)

NVILdADH

+ ‘dnpad) “SUdJUT

WO} ind ‘Aydde 0}, “SeM,x Ud -OS

JNO O8 ‘OSU 01,

JILIASNO

My -SOPND VWWd

JILLINAS

se U1, wia}s

(‘inp

“My0}, ‘Jno‘f)




ABS O},,

ke

JILTHSNO

5JeA VW

‘pusose 0},

,Puey,,

(pedy) PAx

OILINYS

wl [OAs

OF

(,puey,, Joj “u jo UONeALIOp {UOIjBAOUUL

Aes

0},

Mors,dn wia}s) + 3, Syed -wias :sonue Morsdn

01, y9ZTOS

O}OWUIC) ‘WON

JIAGVHO

,dSeI3

-pak-»

3A

uy -WON

-Pp

OLLOWO ~

Ky ‘Sapy VWWd

0},

Ay WVd

“166 -Oh-»

“P66 -Jnk-»

°766 k-» “qe

"WIdS “37)

,J00101d 01,

“€66

x

LEV

SUIOD *,,JNO

01,

:,,OSH

“N- YIM 1001)

k-» -mBn

(xyur“n

0}, 3nd Jno

0},

966

-seyeh,

*L861 (ZOS#

,,dSTeI -109)

Sutotof St [B90A 0}, yse -sanb) asuodsa1 0) ,(uon Was) + Addey (syuaao Sy (‘sned

OSd

MPIP hy “yn “) WI9}S Jovem + Uy (‘ulj-uou0},

0},

“OT “QIOA -O1

O},, OSL

(xIJUr

ayy!

Sutotofar,, ‘Ayrunyiodunt, Ud)S) UII-O1 “a°1 puodsal 0} ‘dnp 10} -UI uonesnsoe JO *Sud} + Ay -ap ‘(Aranb

“yeh,

“0

syed

:son 9SLI< -UI

ifeyyzr {pe

zryed ,Aueur,,ways) *JjNs -URWOS

+

eyoojw -6n&,

JLLOWO

‘°d'T) 3s¥d (ano

0}, MOI)

WON eA]

O1,, ,osTel 1001) YIM -N-,

JO SuUTUvOU -BAOUUI -MOLeU Spury ‘puodsai ul 1no -yeh-»yeads "W9¢ JO pnoy :U01} BUI0} 0}, ,osuodsa1 “3q) (asuodsa1 “Nn K-»

°L66 “WeA-»

“VW

OLLINAS

wie

-mbek-»

Bx ~ OLLIMSNO noOd)DSd -Bnks -yeks [Ind.JJ° anos < DSd0},[Indmbx Sd 0}, .TUUpe

(ygns Ly “U -ouos,auenb

+ UI9}S) ,SSou

“S66

‘Sap ky

NVILdAD4

VWd

CLV

oye)

~ QIGVHD YOO ‘U0 mbedk,wTeEM autos 0}, 2

yo $}99I

OIGVHO ~

- 4e(A) »

(7 pusddy) ,MOUY 0},

»voPly Oj,

(Ax UM < AAx Dd)

uo peoy ind 0}, -snynynd

Uy + WdIs)

S, + [dure

(‘sneo

-mAP@hx UM AOS

,Jayjour

(19q

-WUNU UT aseaId

JILOWO

(-ygns “[pe Tx +

‘(‘pud}xd Oy +

woge 21qe poy Mouy ‘JO -93 OEM 9q 0},

(u> "A -jyns ‘U Uy + Wd]s) ous ‘UTeLIIOsuryyAure SA] SUTeLIO9 -UTRLIOO IO} MOUY 0}, ubed

Jouadns ‘jU3]

w9)S) vCqued ul)

OLLINAS

zeber-yv

‘eM 9q 01,

WId}S) ,DyVMe

IYI

-[99X9 “[NI[LIS

‘AYWIOMISNy,,

Cygnus ‘UW, + W931) ,,PIetys,, at

NVILd ADH

Ky :Sapiry) VWd

JILLIHSNO

BMSETY OS

do} jo

CLV

,[220U09 0},

~myee A- 5,

‘666

"866

yekx-

A-y mR JO-

,oUeU

“OOOI -Tehy

sgea

0}, aq ‘snodof ‘aorofar,J[Rxa +

vy

(Jyns

“WTUR

gered urys,, Jo oy} yorqJO PIE ,,speuTue uio}s)

+ Gy

“Vv

OLLINAS

-jnA-»

0}, ,JNOYS “TOOT ‘dnpal se

“jjns

OdJJO

,Zznq

/ (#

Wy
[x Ws +"U (jgns

oreyeq -OS ejerel

NVILdADH NO

(‘sudjUur

yyim

Uy ‘U

uUUSsT

VWWd ‘Sap ky JIGVHO

JILOWO

99S ‘S]OOI

(SOOT# OsTe

,288 Plo uTeye 1M¢1

‘pase oq 0},

_$,Joy@AA

15 “(UDF [ete OSd

14 ¢1

XUUT qx +-x9 DO

“pus} +

WO}JO

,dSJOUNUIT

UWId}S)YIM -N-,

Tyeewoos ‘,JoIeM,,

JILTHSNO

vlag wet

=—-— :Oq ({woA] ogund O},,

SIOALL,,,BOS —“d))

wors

= *("qI9ANP My +

u2S,,

-AQS UL 9OURUSs«++- “U SNOTAAI) ,UBWIOM PpyJO,, -Iaye U/Ay
pre-Eg. *ts (de-glottalization). idle pre-Eg. *ts > Old Eg., ME s, pre-Eg. *dz > Old Eg. z (de-affrication). #8. Old Eg.z >MEs. . PAA *f > Eg. p /#_VC[+continuant/-voice] and /#_Vr.

#10. PAA *p' > pre-Eg. *p. #11. Eg.p >Eg.b /#dlV_. #12. PAA x" > pre-Eg, “y™ /n, T.

#13. PAA *x(¥) > *y(W) > Eg. 6, *k > Eg. gw, y.

29

Appendix 4: Summary of Sound Shift Rules

aioe)

#14. PAA *CV¢ >Eg. 63 (C= PAA *g¥VS > Eg. § ©. PAA *g() > Eg. d /#_Vd, #15.

velar stop other than PAA *g™),

PAA *k() > Eg. ¢ /#_Vt.

#16. PAA *xVh > Eg. kh (see root #402). #17. PAA *C/+labial,+velar] > Eg. C [+velar]. #18. PAA *%i > Eg. i.

. PAA *¢ >Eg.i /#_V[+cont./+dental].

#20. PAA *?>Eg.i /#_VS, S = sonorant. #21.

some PAA *¢u > Eg. w (/#_{f,s}.

#22. #23. #24. #25. #26.

PAA *2>Eg. © /_VC,

C=n,r, org (V =a, aa ?).

] PAA *¢ > Eg. 3 /[+voice/+dentalV_.

PAA *H > Eg. 3 /C[+voice/+velar]V_ (H = *h or *h). PAA *hf > Eg. h3, PAA *¢h > Eg.ch (?).

_ PAA *hy > Eg. 43.

. PAA *tl'> pre-a3e *dl /#_VS.

. PAA *C[+lateral,tobstruent] > Eg. C [+palatal,+obstruent]. . PAA *r > pre-Eg. *1 /#_V(V)C[+obstruent,+labial]. PAA *r> Eg. 3 /C_(t, w, y)# (C #9, m, n, r, w, 5, or x). . PAA *1 > Eg. r /_(t, w, y)#.

. PAA *1 > Eg. i ([y]) /#_iC (instances of /#_ib and /#_i{C have been observed so far).

. PAA *1 > Eg. r /#nV_C. : nee lS Sieg ME n > Coptic (C.) |/#_Vb_Vb. . . . . .

{s, $, h}. MEn>C.1/#_(V){s, 8, h} V(m,b)# and /#_V{m,b} MEn>C.1 /#_Vk. MEn>C.1/#C,V_C2 (either C; or Cz =m). MEnn>C.l(V)LV. PAA *r (> ME r) > C. 1 /#(C)_o(C)# and /#o _#.

3. Proposed Egypto-Berber (EgB) sound shift:

#1. PAA *p' > EgB *p (before *p, *f > proto-Berber *f) 4. Proto-Boreafrasian (PBA) sound shifts:

#1. PAA *s', *s > PBA *s. #2. PAA *h > PBA *h /#_Vs.

Appendix 4: Summary of Sound Shift Rules

Dell

#3. PAA *t > ? (unknown) /#_VC (C = dental/alveolar obstruent) #4, PAA *z > *d, other *C/+sibilant] > ? (unknown) in the environment of a proximate sibilant in the same word. #5. PAA *n (W), *p > PBA *n. #6. Loss of phonemic tone. #7. PAA *V > PBA @/_#in nominals. #8. PAA *z> PBA *d /_-# (noun-stem-final). . PNE *N- > PBA @ /#_CV (deleting N created by PNE shift #5 in section 6 below)

5. Proto-Chadic (PCh) sound shifts:

ake #2. #3. #4, #5.

PNE *a [aa], *4 > PCh *a [+pharyngeal] > [-pharyngeal] (PAA *¢ > PCh *?, *h > PCh *h. PAA *t', *s' > PCh's',-PAA *c’ > PCh *%'. Eryth.*dzi>,PCh *z."ts.> PChi*s.

PAA *N > @/V_C (cases of C = *ts, *h are known).

6. Proto-North Erythraean (PNE) sound shifts: #1. PAA *Vy, *Vw > PNE * VV /#C__C. #2. PAA *ee > pre-PNE *1, PAA *oo > pre-PNE *u. #3. pre-PNE *o, *e, *a (< PAA *o,:*e,,%a) > PNE *4, pre-PNE *u, *i (created by #2) > PNE *a.

#4. PAA *-VC- verb roots > PNE *-CV-. Ge Eryth. *in- (concom. pref.) > *N- /#_C. 7. Proto-Cushitic (PC) sound shifts:

#1. PAA *b > PC *m /#_Vn. #2. PAA *g > PC *k /#dV_ and /#wV_. #3. PAA *y > PC *g /#_Vx-. 8. Proto-Erythraean (Eryth.) sound shift:

#1. PAA *j, *dz > Eryth. *dz, PAA *c, *ts > Eryth. *ts

Appendix 4: Summary of Sound Shift Rules

po

9. Proto-Omotic (PO) and North Omotic (NOm) sound shifts: if ii.

PAA *dz > pre-PO *2; PAA *j > pre-PO *¢; PAA *c > PO *%3 ([s]2).

lil. pre-PO *é > PO *5, *ts > PO.*s /#_. #1. PAA *a(a) > PO *e(:) /_C- (C =PAA *£ or *h). #2. [+pharyngeal] > [-pharyngeal] (PAA *f > PO *?2, *h > PO *h).

#3. PAA *CWa > *Co /#_C> (C2 =*1, *dl, and *c, in the examples

#4. #5. #6. #7.

PAA PAA PAA PAA

so far noted). *Cw > PO *C /#_V, V #i(:). *§ > PO *%> /i, y_. *VNC > PO *V:6 (C=[- voice]). *u > NOm *o, PAA *i > NO *e. [7a. */e/ -> *[i] /EN_C (N = nasal);

*[i] #1_C[-+abial/+voice] (PO *1 < PAA *1 and *4);

*[i] /#b, p_t; *(i] #S_K G = PO *s, *s', or *8; K = PO *k, *k™, or *?); and *[e] elsewhere;

[7b. */o/ -> *[u] /#C_C[+labial];

*[ul, */e/ -> [i] /H_C; Fuh eke hak xk

CC = POM

Pat fac

*Tul, */e/ -> [i] /#(2)_C- (but not /#(2)_CC-);

*[ul, */e/ -> [i] /#C[+labial]_C2 (C2 = ts, c’); and *[o] elsewhere. ]

#8. PAA *a> PO *o /#C;_C[+labial], C1 = PAA *z, *j. . PAA *CW > PO *C (deleting all remaining labiovelars). #10. PAA *e(:) > NOm *i(:) /#S_C (S = sibilant; C = *d, *n, or *r); #11. PO *V: > NOm *V /#C[+velar, -voice]_C, where PAA root had * tone, #12. NOm *u > Ometo *o, *i > Ometo *e /#(2)_C[-continuant, -voice] #13. PO *u > NOm *u: /#C_P (C = voiced oral stop, P = voiceless labial).

#14, PO *V: > NOm *V /#C_C- + -(V)C- suffix. #15. PAA *N > NOm (or PO?) @ /V_C (C = [+sibilant] or PO *p ((f]

/V_).

References

Allan, E. J. 1976. "Kullo." In M. L. Bender (ed.), The Non-Semitic Languages of Ethiopia, pp. 324-350. East Lansing: African Studies Center, Michigan State University. Appleyard, D. L. 1984. "The internal classification of the Agaw languages: a comparative and historical phonology." In James Bynon (ed.), Current Progress in Afro-Asiatic Linguistics, pp. 33-67. Amsterdam, Philadelphia: John Benjamins Publishing Co. n.d. "Proto-Agaw Roots." Unpublished typescript.

Arvanites, Linda. 1990. The Glottalic Phonemes of proto-Eastern Cushitic. dissertation, University of California at Los Angeles.

Ph.D.

Bender, M. L. 1971. "The languages of Ethiopia: a new lexicostatistical classification and some problems of diffusion." Anthropological Linguistics 13 (5): 165-288. 1975. Omotic: A New Afroasiatic Family. Carbondale: University Museum, Southern Illinois University.

1988. "Proto-Omotic phonology and lexicon." In M. Bechhaus-Gerst and F. Serzisko (ed.), Cushitic-Omotic. Papers from the International Symposium on Cushitic and Omotic Languages, Cologne, January 6-9, 1986 , pp. 1221-159. Hamburg: Buske. 1989. "Gender in Omotic." Journal of Afroasiatic Languages 2 (2): 203-221. 1990. "A survey of Omotic grammemes." In P. Baldi (ed.), Linguistic Change and Reconstruction Methodology. Berlin, New York: de Gruyter. Bomhard, A. R.

1988. "The reconstruction of the proto-Semitic consonant system." In Y.L. Arbeitman (ed.), A Semitic/Afrasian Gathering in Re-

AIS,

534

References membrance of Albert Ehrman. Amsterdam, Philadelphia: John Benjamins Publishing Co.

Breeze, M.

1990. "A sketch of the phonology and grammar of Gimira (Benchnon)." In R. Hayward (ed.), Omotic Language Studies (which see below), pp. 1-67.

Callender, J. B.

*

1975. Middle Egyptian. Malibu: Undena Publications. Cerny, J.

‘

1976. Coptic Etymological Dictionary. Cambridge: Cambridge University Press. Cohen, D.

1968. "Langues chamito-sémitiques." In A. Martinet (ed.), Le Langage. Bruges: Encyclopédie de la Pleiade. Cohen, M. 1947. Essai comparatif sur le vocabulaire et la phoné tique du chamito-sé mitique. Paris. Diakonov, I. M. (Diakonoff)

1965. Semito-Hamitic Languages: an Essay in Classification. Moscow: "Nauka" Publishing House. 1970. "Problems of root structure in Proto-Semitic." Archiv Orientdlni 38: 453-480. 1975. "On root structure in Proto-Semitic." In J. and T. Bynon (ed.), (ed.), Hamito-Semitica, pp. 133-151. The Hague, Paris: Mouton. 1981. "Earliest Semites in Asia." Altorientalische Forschungen 8: 23-74. 1988. Afrasian Languages. Moscow: "Nauka" Publishers, Central Department of Oriental Literature. Ehret, C.

1979a. "Omotic and the subgrouping of the Afroasiatic language family." In R. L. Hess (ed.), Proceedings of the Fifth International Conference on Ethiopian Studies, Session B, April 1318, 1979, Chicago, USA, pp. 51-62. Chicago: Office of Publication Services, University of Illinois at Chicago Circle. 1979b. "On the antiquity of agriculture in Ethiopia." Journal of African History 20: 161-177.

References

DOD

1980. The Historical Reconstruction of Southern Cushitic Phonology and Vocabulary. Kélner Beitrage zur Afrikanistik, B.5. Berlin: Reimer. . 1987. "Proto-Cushitic Reconstruction." Sprache und Geschichte in

Afrika 8: 7-180. 1989. "The origin of third consonants in Semitic roots: an internal reconstruction (applied to Arabic)." Journal of Afroasiatic Languages 2 (2): 109-202. 1991. "Revising the Consonants of proto-Eastern Cushitic." Studies MS.

in African Linguistics 22 (3): 1-65. The Historical Reconstruction of Nilo-Saharan..

Ehret, C., and M. N. Ali. * 1985. "The subclassification of Soomaali." In T. Labahn (ed.), Proceedings of the Second International Congress of Somali

Studies, pp. 201-269.

Hamburg: Buske.

Faber, A.

1985. Semitic Sibilants: (Manuscript)

A Study in Comparative Lexicography.

Faulkner, R. O.

1964. A Concise Dictionary of Middle Egyptian. Institute.

Oxford: Griffith

Fleming, HC:

1969. "The classification of West Cushitic within Hamito-Semitic." In J. Butler (ed.), Eastern African History, pp. 3-27. New York: Praeger. 1974. "Omotic as an Afroasiatic family." Studies in African Linguistics, Supplement 5, pp. 81-94. 1976a. "Cushitic and Omotic." In M. L. Bender, et al. (ed.), Language in Ethiopia, pp. 34-53. London: Oxford University Press. 1976b.

"Omotic Overview."

In M. L. Bender (ed.), The Non-Semitic

Languages of Ethiopia, pp. 299-323. East Lansing: African Studies Center, Michigan State University. 1977. Oral presentation, Conference on Cushitic Origins, School of Oriental and African Studies, London, June 1977. Greenberg, J. H.

1963. The Languages of Africa. Bloomington: Indiana University Research Center in Anthropology, Folklore, and Linguistics. (Also reprinted in several later years)

536

References

Hayward, R. J. 1982. "Notes on the Koyra language." Afrika und Ubersee 65: 211268. 1984a. The Arbore Language: A First Investigation. Hamburg: Buske. . 1984b. "A reconstruction of some root extensions of the Eastern Cushitic verb." In J. Bynon (ed.), Current Progress (for which see

Appleyard 1984), pp. 69-109. 1987. "Terminal vowels in Ometo nominals." In H. Jungraithmayr and W. W. Miller (ed.), Proceedings 4th International

Hamito-Semitic Congress, pp. 215-231. Amsterdam, Philadelphia: John Benjamins Publishing Co. 1990. Omotic Language Studies. London: School of Oriental and African Studies, University of London. Hetzron, R.

1978. "The nominal system of Awngi (Southern Agaw)." Bulletin of the School of Oriental and African Studies 41: 121-141. 1980. "The limits of Cushitic." Sprache und Geshichte in Afrika 2: 7-126.

Johnstone, T. M. 1977. Harstsi Lexicon. London: Oxford University Press. 1981. Jibbali Lexicon. Oxford, New York: Oxford University Press. Jungraithmayr, H., and K. Shimizu. 1981.

Chadic Lexical Roots, Vol. I. Berlin: Reimer.

Krait, C,H. 1974. "Reconstruction of Chadic pronouns I: possessives, object and independent sets—an interim report." In E. Voeltz (ed.), Third Annual Conference on African Linguistics, pp. 69-94. Bloomington: Indiana University Publications, African Series, vol. 7. Lesko, L. H.

1982-1990. A Dictionary of Late Egyptian, vols. 1-5. Berkeley: B.C. Scribe Publications. Leslau, W.

1959. A Dictionary of Moéa. California Press.

Berkeley, Los Angeles: University of

References

hall

Martinet, A.

1975. "Remarques sur le consonantisme sémitique." In A. Martinet, Evolutions des langues et reconstruction. Universitaires de France.

Vendome:

Presses

Moscati, S. (ed.)

1964. An Introduction to the Comparative Grammar of the Semitic Languages. Wiesbaden: Harrassowitz. Newman, P.

1977. "Chadic classification and reconstructions." Afroasiatic LinSuistics dS) t 1-42: Newman, P., and R. Maa.

1966. "Comparative Chadic: phonology and lexicon." Journal of African Languages 5 (3): 218-251. Pia, J. Joseph. 1984. "Multiply-tiered vocalic inventories: an Afroasiatic trait." In J. Bynon (ed.); Current Progress (see Appleyard 1984), pp. 463475.

Prasse, K.-G. 1975. "The reconstruction of proto-Berber short vowels." In J. and T. Bynon (ed.), Hamito-Semitica, pp. 215-228. The Hague, Paris: Mouton. Sasse, H.-J. 1979. "The consonant phonemes of proto-East-Cushitic (PEC): a first approximation." Afroasiatic Linguistics 7, 1: 1-67. Schuh, R. G.

1981. A Dictionary of Ngizim. Berkeley, Los Angeles: University of California Press. Shryock, A. M. 1990. "Issues in Chadic classification: the Masa group." M.A. thesis, University of California at Los Angeles. Steingass, F. J.

1978 (1884). Arabic-English Dictionary. New Delhi: Cosmo Publications. (Reprint of 1884 edition.)

Vycichl, W. 1983. Dictionaire étymologique de la langue copte. Leuven: Peeters.

538

References 1990. La vocalisation de la langue egyptienne, Vol. 1: La phonetique . Cairo: Institut Francais d'Archeologie Orientale.

Wedekind, K.

1985. "Why Bench' (Ethiopia) has five level tones today." In Pieper Stickel (ed.), Studia Linguistica Diachronica et Synchronica, pp.881-901. Berlin, New York, Amsterdam: Mouton de Gruyter. ‘ 1990. Gimo-Jan or Ben-Yem-Om: Bené-Yemsa phonemes, tones, and words." In R. Hayward (ed.), Omotic Language Studies (see above), pp. 68-184.

Zaborski, A.

.

1980. "Can Omotic be reclassified as West Cushitic?" In G. Goldberg (ed.), Ethiopian Studies: Proceedings of the Sixth International Conference, Tel-Aviv, 14-17 April 1980 , pp. 525-530.

Rotterdam: Balkema, 1986.

Index 1

Proto-Afroasiatic (Proto-Afrasian) Consonants

*b — 1-40, 119-122, 144, 161, 182-185, 240, 241, 262-264, 294, 308, 351, 352, 375, 383, 408-410, 463, 488, 494-496, 506, 507, 546, 610, 663, 664, 709-711, 738, 739, 770, 807-810, 835, 838-840, 864, 892-894, 927-930, 943, 950, 992 *c — 123, 312, 455, 488-493, 573-575, 665, 740, 951 *c'— 49, 67, 68, 100, 101, 313, 546-567, 577, 629, 666, 741, 952, 953 *qd — 47, 69, 102-104, 118-158, 265-269, 295, 296, 340, 412, 508, 525, 529, 578, 630, 667, 668, 742, 771-773, 931, 954, 993 *d1 — 384, 413, 414, 440, 631, 669, 670, 712, 743, 836-861, 932, 955957 *dz — 269, 455-462, 579, 671; Appendix 3: 1016-1018 *f — 65-99, 124, 162, 186, 187, 198, 211, 270-273, 314, 324, 341, 353, 360, 385, 415, 445, 464, 509, 547, 548, 564, 611, 672, 673, 694, 713, 744, 811, 812, 841, 842, 865-867, 891, 895, 933, 994 *g — 70, 125, 262-292, 490, 497, 498, 530, 565, 612, 674, 675, 706, 813, 843, 844, 868, 934, 935, 958, 959; Appendix 3: 1017 *gW — 126, 189, 242, 293-308, 613, 814, 833, 845, 995 *gW or *g — 714 #4, 71-73,.127, 163,164,190, 191, 212, 243, 350-373, 457,471, W159 745,.815; 816; 369,-896 *y or *g — 188 *ywW— 128, 374-382, 580, 690, 897, 936 *yW or *y — 244, 531 *h — 5, 74, 105, 129, 165, 245, 246, 274, 297, 356, 386, 416, 532, 549, 550, 581, 614, 615, 659, 660, 716, 767-802, 817, 846, 870, 871, 898, 937, 996 *h — 6-9, 42, 43, 75, 76, 106, 130, 131, 166, 192, 247-249, 275, 298, 299, 315, 332, 341, 357, 358, 417, 441, 449, 472, 473, 510, 511, 551, 552, 582, 583, 587, 616, 644, 738-766, 818, 819, 847, 872, 873, 899, 900, 960-962; Appendix 3: 1023 *hW 2? — Appendix 3: 1012-1015 *j — 10, 300, 387, 418, 463-469, 584, 617, 676, 774, 963, 964; Appendix 3: 1019-1024 *) or *dz — 470-481, 820 1 77) 1075132, 1674168, 213, 214, 309-338,465, 491, 512,553, Hl 585, 751, 821, 938, 965, 997; Appendix 4: 1012, 1019

509

540

Index 1. Proto-Afroasiatic Consonants

12, 133, 169, 170, 193, 339-349, 513, 516, 618, 717, 848, 874, 875, 901, 998 *kW or *k — 677, 746 *k' — 23, 78, 79, 108, 114, 134, 194, 195, 215, 408-435, 458, 465, 499, 554, 619, 620, 632, 645, 657, 678, 730: $225 523; 849, 902, 903, 966, 967, 999; Appendix 3: 1021 *kW' — 216, 436-454, 850, 876, 904 #kW! or *k' — 135, 514, 515, 533, 534,536, 586, 939 *] =" 13-16,-44,.45,.80, 81, 109; 136,-he7; 171521725 1962172208. 250, 276, 277, 301, 316-319, 342, 343, 359, 360, 376, 388, 400, 407, 419, 442-445, 466, 467, 474, 492, 516, 517, 554-556, 587, 588, 633, 679-681, 718-720, 775-778, 803-835, 905-907, 968, 969, 1000 *{ — 17, 46, 110, 111, 138, 197, 251, 320, 344, 361, 389, 390, 446, 634, 635, 643, 721, 862-891, 1001; Appendix 3: 1024 *m — 139-143, 173, 219-221, 278-280, 321, 322, 362, 420, 447, 449, 459, 475, 485, 522, 535, 556-558, 568-607, 621, 636, 682-684, 722, 723, 747-749, 779-782, 824, 851, 877-879, 908, 1002 “ye | Oe 9 32447) 120143, 144A 19S. 199972235 oa, 363, 365, 366, 377, 397, 439, 448, 500, 501, 505, 518, 536, 559, 589, 608-628, 637, 686-688, 724-728, 750-753, 909, 970, 983 *y — 82, 200, 201, 281, 422, 502, 503, 519, 629-642, 689-691, 880, 916 *n or *n — 83, 363, 364, 391, 401, 421, 685 *yW — 48, 49, 224, 643 *n — 644-658 *kwW —

*n or *q — 662 *n or *y or *n — 590, 659-661 *p — 41-64, 145, 225-227, 252, 326, 327, 367, 378, 379, 392, 520, 521, 560, 692, 754, 755, 783, 881, 910, 940, 941; Appendix 3: 1020

*p or *f— 3 *p' —

100-117, 146-148, 253, 260, 282, 522, 561, 662, 756, 784, 785, 825, 852-854, 882, 883, 971 *r — 10, 20-23, 25, 50-52, 61, 84-86, 112, 113, 149-151, 174-179, 202, 228-231, 254-256, 283-285, 302, 303, 328-330, 345, 346, 368, 369, 393, 402-422, 423-426, 450, 460, 476, 486, 487, 493, 504, 505, D252 AO DAO D0: 562 564, 591-595, 641, 646, 647, 651, 693-697, 729, 730, 757, 786-790, 855, 856, 884, 889, 924-948, 972976, 1003, 1005 *5 — ae oy 209- 239, 286, 325, 394, 427, 428, 596, 698, 726, 791, 826 ?

Hs *s' *§ *t

— 24, 114, 152, 257, 395, 429, 529-545, 699.758. 792.703: 197; S27 or *c or *ts — 53 — 25, 54, 287, 288, 347, 396, 506-528, 700, 731, 735, 828, 978 a 92, 159-180, 331, 332, 366, 399, 406, 430, 538, 9392. 701% 7276

Index 1: Proto-Afroasiatic Consonants

541

*t'—

26, 88-90, 232, 240-261, 370, 431, 451, 453, 597, 638, 639, 648, 702, 794, 795, 829 ate ae 56, 333, 348, 432, 452, 598, 622, 703, 759, 796, 892-923, 979,

*tl' or *dl — 456 *ts — 28, 304, 305, 321, 322, 334, 365, 482-487, 599, 704, 705, 733; Appendix 3: 1013, 1014 *ts or *c — 41, 203, 339, 411, 494-505, 576, 981 *w — 93, 153, 258, 335, 477, 478, 517, 525, 540, 541, 600, 640-642, 649-651, 654, 734, 760, 761, 830, 857, 885, 886, 912-914, 926, 942, 949-990, 1004, 1005; Appendix 3: 1016, 1018 *x — 29, 30, 57, 58, 154-156, 204, 233-235, 380, 381, 383-398, 831, 982 *xW — 32, 371, 399-407, 461, 561, 601, 602, 652, 661, 728, 832, 1006 *xW or *x — 91 “*y — 32, 92, 259, 260, 382, 397, 407, 433, 453, 469, 503, 526, 565, 653, 654, 706; 735, 761; 797, $33, 857, 912,,915-917, 942, 943,983, 984, 991-1010; Appendix 3: 1015

*y or @ — 1011 *7 — 33, 59, 93, 94, 115, 181-208, 236, 289, 397, 454, 623, 658, 707, 710, 736, 737, 762-764, 798, 799, 918-920, 944, 945, 985 *? — 34-36, 60-62, 95-97, 116, 199, 237, 238, 261, 275, 306, 336, 434, 469, 462, 479, 542, 566, 603-605, 624, 625, 655, 709-737, 765, 800, 801, 813, 858-860, 887-889, 921, 946, 986, 987, 1007-1009 *¢ — 37-40, 63, 64, 98, 99, 117, 157, 158, 180, 206-208, 239, 284, 290292, 307, 308, 337, 338, 349, 372, 373, 398, 435, 480, 481, 527, 528, 543-545, 567, 606, 607, 626-628, 642, 656, 663-708, 766, 802, 834, 835, 839, 861, 890, 891, 922, 923, 947, 948, 988, 990, 1010; Appendix 3: 1022 *¢ or *? — 205

whee

ae

eet :

aha

tay a .

i

es

hte

pre aaa

a

Fee

nl ;

ur

a

But Yet Skee ae

=

[we We

up.

wee.

To ne

,

ay

pes

IRR Macao.ee 2

,

Index 2

Reconstructed Root Meanings beat, to

A above, to be above, up abuse, revile, to

accomplish, to add together, to alight, afire, to be all anger, to act in

appear, to apply, put into effect, to approach, to , arrange, to arrive, to ask for, to

at, by at, to

attack, to

attack, plunder, to avoid, to aware of, to be

beat, strike with weapon,

O71 221 222 80 22 Pal3) ww) (Ae 686 683 245 714 802 704 924 803 582 677 403 454 659 999

to

beaten, to be beget, to begin, to behind, to be behind, to put bend (intr.), to

back back, lower

bad, to be

bad, to go bark (of tree) bear a child, to beat, to

770 514 18

119 147 74 Be 302 319 bend (part of body), to oy bend (tr.), to 602 603 685 956 bend back (tr.), to 939 bend down, to 435 bend joint of body, to 371 bend round, to 741 844 bent, curved, to be 227, bent over, to be(come) 265 big, to be 591 bind, to 115 bite, to 440 676 1d, 857 875 jieye bite into, to 753 615 bite into hungrily, to 417 bite off, to bite or swallow, to take a 450 556 black 640 blaze, to 710

B baby

552 902

599 119 229 146 550 Tao 848 445 407 138 173

543

544

Index 2: Reconstructed Root Meanings

bleed, to blood blow, to

blow (with mouth), to

body bone boy break (tr.), to break apart, to

20 140 JENS 923 aS 733 428 637

burn brightly, to burn low, to

burn up, to bush, uncleared land

(é:

break by hitting, to

calabash utensil

call (n.) break off, to

call, to

call out, to

break up (tr.), to

break to pieces, to bright brother build, create, to bull, ox burn, to

burn (intr.), to

call to, to carry, to cast off, cast out, to catch hold of, to cease, to change direction, to chase, to chatter, to

chest, organ of chest, upper chew, to chew up, to

child, offspring claw, paw, to

[Append.3] burn (of fire), to burn (tr.), to

22 696 154 486 546 246 345

clear, to make clench the teeth, to clever, to be cloud cold come, to

come away, to come back, return, to

375 836 158 183 442 610 84 775 845 O72 622 Ta 72 313 326 815 898 487 985 287 ey fe) 479 891 456 907 562 143 694 408 570 1007 416 586

Index 2: Reconstructed Root Meanings come forth, to come near, to

come out, to

come to, arrive at, to come toward, to

come upon, to conceal, to connect, join, to

continue, keep on, to

convey, to

cover, to

cover over, to cover up, to

covered, to become cow; cattle crust

cry (n.) cry, to

cry loudly, to cry out, to

curve, to

curve, to form a cut (n.) cut, to

584 225 518 533 819 095 649 651 TES 280 998 369 926 962 191 476 566 692 800 445 512 886 [35 336 884 886 426 888 354 836 708 pS) 373 381 521 633 660 447 741 374 293 228 295 352 413 431

cut apart, to

545 635 856 960 1007 78 965

cut hole in, to

81

cut into, to

30

cut off, to

[Append.3] cut repeatedly, to cut up, to

117 177 196 425 538 762 787 100 411 670 890 1023 ols 330 432

D

damaged, to be

dark-colored, to be

629 840 368 558 529

darken, to

556

dark, to become

639 darken, become dim, to dawn, to

1013 sg

daylight, sunlight

541

deceive, to

deiok

defeat, to defecate, to deficient, to be despise, to destroy, to

86 39 618 959 927 979

destroyed, to be

O27

die, to

600 1010

Index 2: Reconstructed Root Meanings

546 die out, to

dig, to

dig out, to dig up, to dirt, rubbish, to form

dirty, to be disapprove, to divide, to drain out, to do, to do, make, to draw out, to drench (of rain), to drenched, to become drink, to drink up, to

drip, to drive, to drive, chase, to drive away, to dry, to be

477 551 764 329 61 978 256 484 427 70 239 98 69 282 204 542 332 542 561 1008 772 45 286 Sy)

E eagerness ear earth earth, ground

earth, piece of eat, to

eat up, to egs emerge, to empty, to enlarge, increase, to evening, to become examine, to

780 728 178 520 383 159 436 606 125 So2 585 449 456 598 150 596 ate!

excited, to be excrete, to

exhale, to expand (intr.), to

extend (lengthwise), to

extinguished, to be extract, take out, to exude, to

eye

869 205 433 488 611 334 654 860 866 984 508 163 400 339 720

3

958 769 fall asleep, to 977 fall down, to 171 fall into ruin, to 234 fasten, to 634 fat, to become 703 742 fear, to 455 646 feed, to 524 feel (with fingers), to 288 feel bad, to 758 feel good, to 790 feint, to 745 female 399 fence, to 981 find, get, to [Append.3] 1024 face, lower fail, to

finished, ended, made useless, to be finished, used up, to be fire

fit on top of, to fix the eyes on, to flames, firelight flank

851 77 717 353 912 641 106

Index 2: Reconstructed Root Meanings flash, to flee, to

23 207 898 394 786 798 9 36 123 312 358 540 578 587 595 913 955 1014 490

float, drift, glide, to

flow (n.) flow, to

[Append.3] flow, to let

flow heavily, to flow out, to

74

flow over, to

574

flow slowly, to

547 85

fluid, to give off 164 fluid, to produce or apply 915 fly, to a fix the eyes on, to 911 fold (tr.), to 370 fool, to foot forest, bush four frail, to be

to

-

glow, to

glow, burn (of fire), to go, to go around, to go away, to go down, to

sie 192 842 932

flower, bear fruit, to

get up to leave, start out,

603 812 901 944 283 93 83 647

go from, to go in front, to

go out, to

go rapidly on foot, to go round, to go secretively, to go toward, to go up, to

go/come, to go/come forth, to good, to be grab, to grain (cereal) grasp, to

G

gasp, to get, gain, to

248 289

grasp and take away, to grass stalks

988 993 376 388

Index 2: Reconstructed Root Meanings

548

grassy area, low vegetation, scrub

great (in size, number) grindstone

grip, to grip tightly, to ground ground, area of ground, area of open

grow, to

887 262 32 149 669 778 631 o> 383 768 161 690 742 837 864 983

he he/she/it head, crown of head, side of

heap heap, mound, to form a heart heat heat, to

heat, to give off heat, light, to give off

heavy, to be(come) held, to be here/there

hindquarters hit, to

grow (person, animal), to

grow in size, to grow large, to grow up, to

grumble, to

gullet

974 136 963 1003 281 1003 420 354

hit repeatedly, to hold, to

H hack, to hair

hair, body

hair, pubic handle, to

happy, fortunate, healthy, to be hard, solid, to be harden, to harm, to

haste, hurry (n.) hasten, hurry, to

760 120 889 247 501 607 a7 951 415 855 879 545 756 969

hold in the mouth, to hold onto, to

hold with tongue, to hole, breach hole (in), to make a

honey, to have hot, to be(come) hot, to make house hull, husk hurt, ache, to husk, to hut

209 209 Zi 278 133 172 809 374 722 846 709 429 698 206 118 482 144

363 385 580 910 226 448 472

Index 2: Reconstructed Root Meanings

549

land lap, to

I I (independent pron.) immerse, to

impel (push, pull), to increase, to increase (intr.), to

[Append.3] increase (tr.), to inhale, to insert, to intend, to intertwine, to

724 507 575 300 174 oF 101 102 136 763 831 864 1012 934 73 401 134 157 123 964

lap up, to last, endure, to

lay, to

leak, seep, to leak out, to leave, to

leave abruptly, to leg lick, to

lick up, to lie (down), to lie flat, to lie still, to lift, to light, slight light (n.)

light, to

jaw join (intr.), to

join (tr.), to join, assemble, to

554 680 900 369 488 623 392

-

light up, to limb (of body) listen, to listen to, to

[Append.3] lit, to become little, low, small little, small

live, grow, to liver loins

K

kidney kill, to kill (animal), to know, to

343 292 480 W290

L lack, to

261

long look, to look at, to

look at intently, to look for selectively, to

446

Index 2: Reconstructed Root Meanings

550

lose, lack, be without, to lost, to become loud, exuberant, to be low, to become lower, to

lower oneself, to

589 945 899 617 859 266 537 782 Ol 937

mane

meat, flesh meet, to merry, to be(come)

1000 1011 fas) oils)

1017 716 259 650 804 [Append.3] 1016 moisture 916 moon 871 mouth 65 mouthful 15 mouthful, to take a 674 move (intr.), to 14 125 pol 824 O35 move (tr.), to 191 835 954 move about quickly, to 190 move along, to 202 825 947 move along, proceed, to bie move apart, to 27 move apart (tr.), to 82 move in a confused [Append.3]

moan, to moist, damp, to be

move water about, to

much, many, to be muscle, sinew

691 681 651 832 601 868 251

N

564 220 nap, fuzz 365 narrow, slim 515 neck, base of 718 noise, to make a loud 340 not be, to By) not be able (to do), to 581 not be there 2 not do, to 294 not function right, to 629 not have, to S12 not move, to 189 498 not want, to 237 notice, to 730 818 1009 notice, to bring to one's 823 nail, claw name

M me, my (bound pron.)

manner, to move to and fro, to move toward, to

O oath obscure, to observe, to

of (genitive)

offspring, to produce old, to become one, someone (pron.)

719 125 201 250 Sp) 609 1004 284 181 470 568 726

Index 2: Reconstructed Root Meanings ooze (bodily fluid), to ooze, be viscous, to

ooze, be thick, to open, to open (tr.), to oppose, to other overflow, to

297 587 908 904 658 970 270 776 474 794

P -pass, to pass by, to

pass over, to

path pay, to pay attention (to), to

[Append.3] peak

penetrate, to perish, to person pick, pluck, to pick up, to pierce, to

place plant growth pluck off, pluck out, to

pluck out, to point (n.)

point, to form a

point, to shape to a

952 593 469 380 853 Zig 818 843 1015 47 423 76 163 621 668 124 40 95 830 457 874 936 485 499

223 423 466 785 203 644

pole pound, to pound (to loosen), to

pour, to pour heavily (intr.), to pour out, to present, to press out (fluid), to

press together, to produce fluid, to produce water, to produce young, to projection, protrusion protect, to pull off, to pull out, to push, to put above, to

put around, to put over, to

put out, to

put the hand to, to put together, to

put up, to

python (?)

Hio:l 460 132 342 441 aay 894 jal 222 702 68 139 O15 331 1004 366 992 217 467 544 850 137 783 928 269 f52 995 838 211 783 642

Q quick movement, to make

quiet, to be

893 410

R raise, to

137 194 STZ 667 682

Index 2: Reconstructed Root Meanings

Biers

raised, to be

reach, to reach out, to

remove, clear away, to remove (from surface), to

remove, take off, to rend, to

repeat, to repetitive movements, to make

respond, to rest, to

reveal, to

ribcage rise, to

rise above, to rise up, to

ritual, to perform root

918 997 695 816 948 166 beh) 48 406 OZ 113 208 666

rub, to

rub against, to rub off, to ruin, to fall into ruined, to be rumble, to run away, to run out (fluid), to

443 996 Soil 16) 33 882 148 161 bi 243 Zaz Ife Bt 367 434 497 567 654 656 690 811 817 909 914 942 997 841 563 678 241 230

salt sand satisfactory, to be say, to

216 efel/ 398 Sol 645 784 903 740 553 234 405 349 898 218 Dies 619 643 880 989

461 193 jis 991 scatter (intr.), to 41 scorn, to [Append.3] 1018 scrape, to 67 233 209 398 509 scrape off, to 628 IST 872 scrape off, scrape out, to 58 scrape repeatedly, to 90 scrape surface of, to Log scraped, to be 2 scratch, to 509 yg! 624 1p.) scratch itch, to 465

Index 2: Reconstructed Root Meanings scratch off, to scratch (with tool), to secrete, to see, to seed seep, to

seep (bodily fluid), to seep out, to send, to send, direct, to -send out, to separate (intr.), to

separate (tr.), to

separate husk, rind, to set apart, to set down, to

set on top of, to set out, lay out, to set out for, to sew, to shake (intr.), to

shape to a point, to she shell shift (position), to shine, to

shoulder, inner shout, to show, to shut, close, to

sick, to be

389 a7 215 672 720 10 pny 613 687 198 418 643 930 930 414 50 56 89 463 490 966 82 O17 854 750 152 941 176 644 209 109 523 11 200 663 834 870 318 565 1001 276 823 156 240 548 290

aiaye)

1006 420 193 sing, to 706 sink, to 306 sip, to 478 867 sip, to take a [Append.3] 1021 38 Sit, to 285 1A0! 807 17 skin Ze 356 422 695 sky 362 slice, to 614 slight, to be 305 small, to be 96 smell, to 224 500 smell (intr.), to 528 625 snuff (up), to 975 soak (intr.), to 282 soak (tr.), to 626 soft, to be 321 sole (of foot) 390 sound, to make a rough 533: sour, to 526 sparkle, to a2 speak, to 274 799 814 142 speak forcefully, to 653 speak loudly, to T21 Wick speak out, to 765 spill (intr.), to 496 spill (tr.), to 384 spill, pour, to 894 64 spill out (intr.), to 481

side

sigh, to

Index 2: Reconstructed Root Meanings

554

spill out in drops, to spit, saliva Spit, to

Spit out, to splatter, to

split (tr.), to split apart, lay open, to split in two (intr.), to spoil (intr), to

spread (intr.), to spread apart (intr.), to spread apart (tr.), to [Append.3] spurt out, to make squeeze, to stamp, to stand, to stand up, to Start, to

start off, to

stay, to

step, to step along, to stick, to stick, prick, to stick into, to

stick one thing into another, to

stick out (intr.), to

stick up (intr.), to

31 114 162 60 164 829 108 378 912 46 502 494 748 638 71 1020 60 849 293 980 863 STi, 24 HIS 214 267 298 412 530 7A! vay 475 28) 335 662 167 134 892

662 203 896 122 203

still, to be still, to become

still, to stay stomach (?) stop (intr.), to

stop (tr.), to

stop doing, to stop up, to strand, strip strength stretch out (intr.), to stretch out, extend, to strike, to

strike (with tool), to strike down, to

strike repeatedly, to strike with a blade, to

strip, to

700 7TSZ 896 946 238 693 942 491 492 Li? By 412 684 796 701 705 943 99 255 260 697 839 269 595 130 940 506 534 922 805 188 6 444 304

strip (from ground), to Strip away (hair, skin), to 747

strong, to be

submerge, to suck, to suck in, to suffer loss, to surprise, to

swallow, to

360 471 1002 620 625 840 661 307 452 688 953

Index 2: Reconstructed Root Meanings swallow, to take a

swallow liquid or soft food, to swell, to

664 665 29 94 148 Diy| 654 656 837 847 994

,

thin, to be this this one

this/that (demonstrative base) threaten, to three throat throw down

tie together, to tie up, to

tighten (by twisting), to

tip (n.)

T take, to

take apart, to

take away, to take from view, to

[Append.4] take hold of, to

take in fingers, to take into the mouth, to

take out, to tall, high, long

tap (hit lightly), to tear (tr.), to tear off, to

tear off in strips, to tell, to they thick, fat, to become

thigh thin (of thing, person)

184 320 885 50 87

top

|Beg

touch, feel, to

1022 409 754 778 988 594 42 630 781 44 734 O10 165 208 Ioffe) 303 444 128 210 703 588 439

trash, residue travel, to

trade, sell, buy, to

tread, to tree trickle, to true

trunk (of tree, body) turn (intr.), to

turn (tr.), to

turn back (intr.), to turn back (tr.), to

233) 439 309 767 767 Bye) O01 308 458 941 548 590 929 233 423 466 682 688 785 263 688 145 Be) 549 186 761 168 350 464 833 806 264 193 213 302 303 338 348 404 976 576 685 895 815 956

Index 2: Reconstructed Root Meanings

55:0

turn round (intr.), to twist (intr.), to twist (tr.), to twist (rope), to two

62 316 328 346 364 743 981 AZ 503 505 877

watch, to

water

water, body of we weak, to be(come), to

weak, to feel wear out, to weave, plait, to

Ui

wet, to

wet, to be

understand, to 249 use mouth (other than to eat), to 663 useless, non-functioning, to become 103 utter, to 386

wet, to be very wet, to become

673 wet, to become

Vv visible, to become voice

thoroughly wet down, to

876 TEX

wet place what? wait, to walk, to

walk about, to

walk swaying, to warm up, to

201 612 438 569 1002 125 180 536 766 826 861 477 212 iS 244 437 569 604 493 291 459 579 583

wide, to be

236 560 lays) 186 213 452 (ie)! 931 126 380 ho 749

wing wise, to be with womb worn out, to be worn out, to become

wrap, to wrap around, to

wrap up, to wrong, bad, ill, to be wrong, to do

piste

At NR

Wey

See stanhee

eo 8. oe

a

AHS

777 451 mE 774 906 925 aot 949 105 324

608 399 54 861 341 812 525 314 671 495

Index 2: Reconstructed Root Meanings wrongly, to do or function

yellow you (fem. sing. bound) , you (masc. sing. bound) you (sing. independent) you (pl. bound) young

671

Y

yell, to

iia 990

Sop THEOLOGY LIBRARY

CLAREMONT, CA

Dion, 657 310 311 IPA 323 707

Vol.

75. James A. |tassel Fie ee

ISBN0-520-09467-0

Dei

ne 96.Brent D.Galloway. A Grammar ofUpriverHalkomelem. ISBN 0-520- ita

O9619-8 ee) ; Vol.104. HaluaAcki aieDeward EiWalker,Jr.Nez PerceOralNarratives.

ISBN 0-520-09693-2.

z

Be Vol.Or. Jon a Dayley. Tzutujil eames ISBN0-520-09962-1 ey

Vol. 111. JamesA.Matisoff. The Dictionary ofLahu. ISBN. 0-520-09711-4

Vol. 112. David John Weber.A Crane ofHuallagatHudnuco) cables

ISBN 0-520-09732-7_ aa Vol. 113. Kathryn Klingebiel.. Noun + VerbCompounding inWesternRomance. ISBN 0-520-09729-7.

_ Vol. 115. Jon P. Dayley. Tumpisa (Panamin ShoshoneGrammar.ISBN0-520-

in

|

09752-1 sae en: : Vol. ‘116. Jon P.Dayley,Tampisa (PanaminnShoshone Dictionary.ISBN 0-520: . 09754-8 Fa

Vol. 118. Xiao-nan Susan Shen. TheProsodyof MandarinChinese. ISBN 0.520 hy 09750-5

Vol. 119. Maurice L.Zigmond, Curtis G.Booth, ae Pamela Munro. (ed. by Pamela Munro). Kawaiisu: AGrammar and Dictionary with Texts. ISBN 0-520-09747-5

_

em Vol. 120. Thomas Edward Payne. Thehon Stories: Participant cone: in Yagua

Narrative. ISBN 0-520-09774-2 aS “ihe Vol.121. Derek Nurse and Thomas J.Hinnebusc, SwahiliandSabaki: A Hingis

History. ISBN 0-520-09775-0.

Vol. 122. Haruo Aoki. Nez Perce Dicnanars ISBN0-520-09763-7

|

he

Vol.123. Mary Ellen Ryder. Ordered tei The Interpretationof Noga Nees

Compounds. ISBN 0-520-09777-7_

ia Vol. 124. jean Gailoe rg.ativity inCoastTsimshian(Smaleyav ISBN0-520mavens Hs -09788-2.

. Vol. #25. Villiana Calac HydeliedBricElliot.YumaykYuma Long Ago. ISBN ap

Sei Se

me

ity