Monographic Journals of the Near East General Editor: Giorgio Buccellati Afroastatic Gnguistics Editors: Robert Hetzron, Santa Barbara Russell G. Schuh, Los Angeles Advisory Board: Bloch, Berkeley Talmy Givin, Los Angeles ‘Thomas G. Penchoen, Los Angeles Stanislav Segert, Los Angeles Volume 5 Issue 3 September 1978 Word Formation Strategies in the Hebrew Verb System: Denominative Verbs by Shmuel Bolozky ~ =~ aS Undena Publications Malibu 1978FROASIATIC LINGUISTICS AAL includes contributions in linguistics within the vast domain of Afroasiatic (Hamito-Semitic) languages. Articles of general, theoretical interest using Afroasiatic material, descriptive, historical and comparative studies are included Editors: Robert Hetzron (698 Zink Av., Santa Barbara, Ca. 93111, U. Russell G. Schuh (15337 Hart St., Van Nuys, Ca. 91406, US.A.) Send all manuscripts to Robert Hetzron or (or Chadic) 10 Russell G. Schuh, Consult stylesheet on back- cover for editing format. ) MONOGRAPHIC JOURNALS OF THE NEAR EAST MINE is a system of journals on the Near East, with each journal devoted to a specialized study area, and each issue consisting normally of a single article. Current journals in the system are Afroasiatie Linguistics, Assur, Computer Aided Rescarch in Ancient Near Kastern Studies and Svro-Mesopotamian Studies. General Subscription. - For a prepayment of $15.00 the subscriber selects random issues from within the entire system as desired, up to a total of 200 pages. The subscriber is also entitled to (1) periodical lists of abstracts from all journals in the system, and (2) reservation to any journal within the system, whereby issues of a given journal are sent on approval immediately upon publication (and may be retumed within two weeks). Library Subscription. A prepayment of $15.00 for cach journal in the system secures all issues of a single volume as soon as they are published. This subscription schedule does not allow the selection of random issues. Library subscriptions are available to both institutions and individual scholars. Individual issues are numbered sequentially within each volume. Each issue has its own pagination, A volume is closed when a total of about 200 pages is reached. A title page and a table of contents listing all issues within each volume are sent to all subscribers at the close of a volume. Periodicity in the order of appearance of issues is not predetermined. A volume, however, is generally completed within one year. Institutional and Professional discount of 20% on single subscriptions entered within six months of the beginning of any given volume (higher on larger orders). Payment must accompany orders from individuals. A handling fee of 80/ will be charged to Libraries if order is not prepaid. Order from: UNDENA PUBLICATIONS, P.O. Box 97, Malibu, California 90265, U.S.A © 1978 by Undena Publications. All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any ‘means, electronic or mechanical, including photo-copy, recording, or any information storage and retrieval system, without permission in writing from the publisher.Monographic Jowrnat’s of the Near East Ajnoasiatic Linguisties 5/3 (September 1978) WORD FORMATION STRATEGIES IN THE HEBREW VERB SYSTEM: DENOMINATIVE VERBS by Shmuel Bolozky Tel Aviv University and University of Mlinois A hierarchy is suggested for the familiar syntactic and semantic factors determining realization of licbrew verbs in the particular conjugations. Other considerations apply whenever applicable: no realizations are allowed khich would form unpronounceable clusters; realizations presery- ing the consonant clusters of original nous are preferred: if a slot is occupied, the semntically-closest conjugation is chosen. To examine these cl:tims, native speakers wore asked to themselves innovate denopina- tive verbs, Or make judgenents about hypothetical verbalizations. The results support the basic semntic/syntactic hierarchy; speakers even invent new devices allowing incorporation of non-syntactic and non-s factors without the hierarchy itself being significantly affected. snantic ‘TABLE OF CONTENTS L. INTRODUCTION. 6. se. ee aaa ee Be ere ea) ‘TRADITIONAL AND RECENT APPROACHES 11 ea 2.1. The conjugations of the verb 22 2.2. The Productivity of bényan meaning and syntactic function 222022 4 “TMM INTERACTION OF FUNCTIONAL, SEMANTIC AND PHONOLOGICAL CONSIDERATIONS. IN DENOMINATIVE VERBALIZATION 2 ee. we peer sreee sees 3 5 3.1. Precedence relations in recent verbalizations’ 1002 1D ot 28 3.1.1. The realization of most quidriliteral and quintiliteral nouns) > § 3.1.2. Verbalization of triliteral and biliteral nouns and adjectives . 6 3.1.2.1, Intransitive denominative triliterals. e+ 0+ +s 7 3.1.2.2. Transitive denoninative triliterals. . 2.2240. 8 Occupied slots. ee ee ee ee ee 0) Additional phonological factors’ 2.0 Dt see 1.4.1. Glides and second radicals in’aigek 222200 2S 10 1.4.2, No stem gominates 10... ee ccc eee ry 3.15. Sporadic analogies... u S-t.o. Preserving the structure of the original noin/atjective en 3.2. A’proposal for a verbalization hierarchy . . taney eee #113 4. PROWCTIVITY TESTS OF VERBALIZATION OF NOUNS AND ADJECTIVES 2... ee 16 4.1, Tests isolating the syntactic and semantic factors... 2) 222220 Ie 4.2. Tests involving initial and final consonant clusters 2.00 222000 18 4.5. Variations in hierarchy .... 2. aaa see 5. WNCUISIN. 6... ee. ee oe abba ecoeo8 22 ‘Shmuel Bolozky [AAL 5/3 APPENDIX I — Sowe recent Literature on realization of roots in different cata ee reenter ree re areees pinata Sees eet ee 2 APPENDIX II — Sentence frames for tests I and IT Sentence frames for tests III and WV... 2... eee 24 [ape 5 og CHOCO DUO ODS A baAHRanon es ++ % 1. INTRODUCTION* When a Linguist is introduced to the morphological system of a Semitic language, his first reaction is quite enthusiastic: here is the ultimate productivity in a morpho logical system! This enthusiasm soon dissipates when he finds out how many exceptions and unaccountable fo:ms the system actually contains, Obviously, Somitic morphological systems are neither fully regular nor totally irregular. This paper attempts to put the productivity of one aspect of Hebrew morphology in the right perspective—the formation of denominative verbs—and establish a tentative hierarchy of the variety of factors involved. ‘Traditional and recent approaches to Hebrew verb mor- phology are mentioned first, with particular reference to syntactic and semantic classi- fications, but it is shown that phonetic and phonological considerations play a role as well. Recent verbalizations are then separately exanined, and precedence relations are tentatively established. Finally, the claims for these precedence relations are examined in terms of the behaviour of fifty native speaking subjects when asked to themselves innovate denominative verbs or nake judgements about hypothetical verhalizations. 2. TRADITIONAL AND RECENT APPROACHES TO HEBREW VERB MORPHOLOGY 2.1. THE CONJUGATIONS OF THE VERB When @ noun is verbalized in Hebrew, a decision has to be made as to what conjuga- tion of the verb it is to be realized in.’ Below I briefly review the nature of traditional verb-norphology in licbrow in terms of tho existing conjugat ions. Hebrew (usually tri- consonantal) Toots can be realized in onc or morc of seven morphological verb-pattems or Conjugations (bényandm)} with the following stem forms: CaCaC, nésCCaC, CéCeC, CuCac, “I wish to thank Ruth Aronson Berman, Gad Ben-Horin, Daniel Boyarin, Peter Cole, Gabi Hermon, Ora Schwarawald and Yael Ziv’for their valuable conments on’ this paper. ‘Native nouns, adjectives etc. also follow morphological patterns (méSkatém). The difference is,'that the mmber of méSkatin is far greater than the nunber of binyanim, and while verbs can oxy be realized in one or more of the seven binyanim, borrowed nouns and adjectives are usually introduced as they are, regardless of existing native méSkalin.1978) The Helnew Verb System 3 NéteCaCel, MECCK, and hu*CCaC, ach of these conjugations has becn claimed to either carry its’own independent syntactic function, or at least denote a core (or a number of cores) of meaning common to groups of verbs realized in it. [ven today school gramars continue to assume roughly the same semantic characterization of the binyanén offered for Biblical tlebrew in the Nineteenth century, and to a large extent even by gramirians Of the Middlo Ages. Melow is a list of independent syntactic functions attributed to the bényadn in-a typical school grammar: (1) ame oF BINVAN TRADITIONAL SYWTACTTC FUNCITON EXAMPLES parat? unmarked base form migrate passive of peat niStax? ‘be sent" passive of Satax "send! piret normally transitive sider “arrange pural passive of péret sudan “be arranged’ hitparer middle voice, normally intran- hittabeS* ‘got dressed’ sitive countérpart of piree or parat he grak normally transitive hixtiv ‘dictate" hugeae passive of hégrit huxtav "be dictated" Cores of meaning common to groups of verbs realized in a particular conjugation are usually arranged in a markedness hierarchy, i.e. if moro than one core of neaning is noted, the most Likely (unmarked) meaning is Listed firs (2) siuvaw Name MeawiNe TLustanrrons négrae change of state nédtak ‘turn on (Light, int.)" ys. datak "be on (Light) piree intensified form of parat Baber ‘smash',, intensified form of Sava "break" hétparee reflexive hutaaxee wash (oneself) change of state (inchoative) Metxaxey ‘widen Cint.)" reciprocal hetwarxw "they argued" agree causative higdie ‘enlarge (tr)! change of stato (inchoative) deliv “ecome ye ow Obviously, many variations on this classification may be found, but the classes above seem to be mentioned by any school grammar. *the root pe (historically pt) has been used traditionally as a prototype, p indi cating the first radical of the root, ° the second, and ¢ the third. ‘Stress is indicated only when not final. ALL verbs are represented in the morphologically most hasic forn—Thind Person Masc. Sing., Past ‘Tense, unless otherwise specified. Thus, niStax ‘be sent! actually means ‘he/it was sent" and Séden ‘arrange’ actually means" arranged". For clurity of representation, glottals will be represented in these cases, though phonetically they arc not realized hefore unstressed vowels. AAL 5, 1134 Shmuel Bolo=ky IAAL 5/3 2.2. THE PRODUCTIVITY OF BINYAN MEANING AND SYNIACT FUNCTION Linguists describing Modern ilebrew are aware of the fact that the "independent" syntactic functions of the biwanén are still quite productive, though sone arc mbre productive than others. The active/passive relationships are very productive in the case of pé?et/pu2at and hig*ét/hug?a’, but less so in the case of pa*at/négtal, As far as transitivity is concomed, nég*at ‘and hitoa*el are typically non-transitive—i.c. do not take °e¢ NP (et is the definite object marker); pé*e€ and hég?ét are typically transitive; and paral is neutral in this respect. The only significant change affecting the traditional system in (2) above has been the teplacemont of paraé by piel as the productive, unnarked forn— which was first noted by Sivan (1963). It is possible that pa%at being neutral with respect to transitivity was one of the reasons for its almost ceasing to be productive, though there mist have been at least another two reasons for it: the fact that it could not accommodate quairiliterals and quintiliterals, and most of its slots being occupied already. 1 know of only two recent denominative paat verbs, one transitive, the other intransitive: ganaz ‘criticize severely’, from gémzu ‘proper nane of a vitriolic’ Israeli critic’ and xanrap ‘nap’ from xkop ‘nap’ (X)', from Yiddish. ‘The latter started from xtop > Laxkop ‘to nap", owing to phonetic similarity of the noun to the infinitival form, which deyeloped into’ xatap by a sort of back- formation. I also heard Lidrop ‘to drop (a course)" among Israeli students in the States, which may develop through similar back-formation into daxap. Ora Schwarzwald also reports (personal communication) that in the Hebrew of Israeli kids in the States, one hears faput ‘to put" and tinok "knock', which behave 1ike "hollon" paral verbs. ‘So pa*at is not yet “dead”. Generally in the speech commmity, however, its productivity is very restricted. ALL in all, honever, the observations conceming the productivity of the binyan syntactic functions are generally true. On the other hand, the productivity of the specialized meanings of the bényanén is necessarily restrictéd, since these specialized meanings char- acterize limited groups of verbs within each bényan. Although there might be sone dis~ agreenent between Schwarzvald (1975) and Berman (1975) on How minor the redundancy rules expressing these predictions will be, it is gencrally agreed (cf. primarily Oman 1971 and Schwarzwald 1975) that predicting meanings of forms on the basis’ of specialized semantic groups in the binyaxém is a complicated matter. Hirst of all, as Schvarzwald points out, only 2.3 of Listed roots occur in all seven conjugations, and 41% of those are ononymots c, realization of a phonological root in sone conjugation may be utterly unrelated somantically to its realization in another conjugation (for instance, saxat ‘scratch’ and hisnct 'film'). Furthermore, even anong non-homonynous roots, the meaning relations often do not develop as expected (for instance, fézmin ‘order, invite'—one would hardly expect hézdanen to nean ‘occur, cone by chance’, hétpa*et often being reflexive, reciprocal , etc.). Numerous examples like these are’cited in the literature on the Subject. It’ would appear that the percentage of verbs one can classify into expected classes is small enough to make the meaning groups quite minor. What Ornan (1971) concludes from this state of affairs is that scmantic productivity is no longer a property of the verb-system, except for the active-passive relationship of hig?it/hug"al and péret/pural. To the best of my knowledge, this strong version of the non-productivity claim has never been acvepted by other Linguists describing the Hebrew verb system. ‘There seens to be ample evidence around for at least some degree of pro- ductivity: “recent formation of verbs, mistakes made by children and other learners of the language, and systematic and alnost-free innovative capability of adult speakers to express thenSelves in an original and "creative" manner by readily coining hypothetical verbs (as is evident in the tests reported in this paper), which are comprehended by other speakers with no difficulty at all. what is normally debated is not the very exis- tence of this productivity, but how it should be described, and the description may some- times be quite subtle, as in the case of Cole's (1976) description of causatives in Hebrew. MAL 5, 11d1978) The Hebrew Verb System 5 Appendix I contains a brief discussion of some recent literature on realization of roots in the different binyanim. 3, THE INTERACTION OF FUNCTIONAL, SEMANTIC AND PHONOLOGICAL CONSIDERATIONS IN DENOMINATIVE VERBALIZATION This paper concentrates on only one aspect of the productivity of the verb system: formation of new verbs from existing nouns and adjectives, whether borrowed or native. ‘There are other aspects of productivity—like the realization of roots in béswaném in which they hal not formerly been used (e.g. hiztik "inject", fiictatem ‘have one's picture taken"), fononatopocic innovations, (c.g. tirter ‘make’noise’, téktek 'tick"), and so on. Iwill restrict myself to denominative verbs, however, as they constitute the most productive source of new verbs in the language, as Sivan (1963) earlier noted, and because there is usually no question as to the fom fro which these verbs wore originally derived. Note that when realized as verbs, borrowed nouns are subject to the same rules and constraints ag native ones—e.g. bileg (‘lie’, from blog ‘Lie (N)", from English b(ug4) ~ béfga "she Lied" vz bilagti 'I Ted", compared with native mékem (‘place’ from makom ‘place (N)") ~ mékma “she placed’ ~ makdmté "I placed" Although Sivan and Bernan have already introduced most of the factors involved in verb for- mation, including those determining formation of denoninative verbs, neither of then attenpted to establish a hierarchy, or priority of importance, anong the different types of factors involved. Obviously, however, the innovator must impose SOME hierarchy on the relevant factors and establish’ precedence relations. In the rest of the paper I will try to show how these precedence relations are reflected in scent verbalizations of nouns and adjectives, and ‘then in productivity tests. 3.1. PRECLDENCE RELATIONS IN RECENT VERBALIZATIONS 3.1.1. The Realization of Most Quadriliteral and Quintiliteral Nouns Pronounceability seems to be the initial consideration. Thus, verbalization of most quad- riliteral or longer nouns would have created an unpronounceabie cluster in hég?il (e.g *hitanid from torpédo, Shikgtin from kagton ‘button’, *hépnéin from panSex "puncture, flat tire, mishap"), in hgrat (e.g. *ritapad, *néhgtar, *népnéar) or in (the future of) pa2at® (titnpad, *tzkgton, *tépndor), owing to’ the basic’ stem-initial cluster of these bényanén, to the fact that the stem-final consonant cannot be expanded into a cluster throughout the verb-stem,® and to their stems being monosyllabic. Such nouns can be realized in either peel (e.g. kigter) or hitparel (e.g. hitkagter). ‘The stems of those, unlike the ones of fiégit and the future of pa*al, are disyllabic in their unmarked’ forms, and more con- sonants can thus be accomodated.’ If the verb is transitive, it will be realized in pé?ed, since hitpael is typically intransitive. Thus, totpédo is verbalized as tixped, Spa’ut has almost ceased to be productive anyvay—as explained in 2.2 above. ©1 am aware of one exception, in child language, viz. feéivaya "he ran very fast", from Sug timpetus'. This is due both to ng functioning Like a single segment and to the tempt to preserve as mich as possible of the original form of the noun * hat is. without any additional inflectional affixes marking such Louna categories as non-third persen, non-singular (= plural) number, non-nasculine (= feminine) gender. AL 5, US.° Shmuel Bolozky iAAL 53 (3) kagton "button (N)" kigter ‘button’ panéer —“nishap" > pingex "cause a mishap" faxzava‘disappointwent' >< *ixzev— ‘disappoint’ and so on, and each of those will have a parallel passive counterpart in pu’aé and usually @ middle-Voice corresponding Métpa7et: (4) kugtax "be buttoned" hétkagter ‘button up" turpad “be torpedoed’ > 2hittanped ‘fail (intr.)" puntax "be caused a mishap’ > —hitpanden "fail because of a mishap" Puxzav "be disappointed’ > hit?axzev ‘become disappointed" ‘The sane would apply to most nouns containing four consonants or more which are realized as transitive verbs, Tf, however, the verb is basically intransitive, without a corresponding transitive and middle voice, it may be realized EYTHE® as pé*ee or’as itpael, since while the latter is typically intransitive, the former may be either transitive or intran- sitive. In such cases one must refer to the more specific meanings denoted by groups of verbs within the Génjanér, ‘Thus, if the noun is to be realized as an independent inchoative, reflexive or reciprocal,® the choice is hétpa7et: (5) rezaax citizen" > hégPacnéax "become a citizen" pirxax —‘hoodlun' > hétparxdax "become, begin to act Like, a hoodlum’ putmus ‘controversy’ > hétpatmes ‘argue about" For all other intransitive verbs, pi%et is chosen: (6) tagkéd "job, function’ tifked? "function" xantanid "nonsense" > xintreS ‘speak nonsense" télegon "telephone (N)' > —télgen "telephone! This is the normal procedure for realizing most nouns containing four consonants or more— though not for all of them, as will be seen below. Note that the generalizations claimed apply to ANY noun or adjective that is verbalized, whether native or borroved, old or now: old borrowed nouns (pulmus 'controversy", Greek) , newly borrowed nouns (tonpédo, t8£egon), native nouns (tagkéd "function', piaxax 'hoodiun') , slang nouns, whether borrowed’ (xantariS, Arabic) or native (hitpaaper ‘fool around", from parpar ‘one who fools around’, from ‘butterfly'), Also, as explained above, once a’ noun or adjective is vetbalized, it behaves like an original verb in every respect,’ regardless of its source. 3.1.2, Verbalization of Triliteral and Siliteral Nouns and Adjectives If no phonetic difficulties are involved to start with, transitivity is less useful as a “or, of course, if the form can only be realized in hétpa*el because this is the only bénjan that allows a pronounceable configuration, e.g, the five consonants in dézengo{ (nane of street) can only be verbalized as fizdaiges "go up and down Dizengoff Street °Gad Ben-llorin has pointed out to me that in sports, tégked may be used transitively, meaning ‘assign role (by coach)". AALS, LO1978) The Hebrew Verb System 7 criterion; pé‘et, though often transitive, is not alMays so (es, bites "Iie! and biter ‘complain’ below), and hig’, which is normally transitive too, has a group of intransi- tive verbs, essentially inchoatives (e.g. hvréd "become pink’ and fiédmén ‘become fat" below). Since pu*at and fu*at do not normally exist independently of corresponding pé>el and bégé forms, it is again impossible to choose a bényan merely on the basis of whether a transitive or intransitive verb is to be formed. Pa°ad being neutral with respect to transitivity does not help either. Since transitivity cannot be very useful in determining yerbalization of triliterat nouns, one has to resort to the meaning groups within the bényaném. The picture that emerges somewhat resenbles the one introduced at the beginning of the paper, with certain changes and further specifications. 3.1.2.1, Intransitive Denominative Triliterals Inchoatives-over-colours, i.e. "become colour’, and inchoatives-over-physical -hunan-qualities, i.e, ‘acquiring a particular physical (human) quality’, are realized in fifi. This was a typical léf° hétpager "die" Savac ‘heart attack’ > —fuéStavee "have a heart attack" xatix ‘good looking’ > hitxatex "become handsome" In hitpa*el are also realized all middie-voice counterparts of transitive verbs realized in péet, including reflexive, inchoative or reciprochal actions: (9) merntaz ‘center’ > hétrakez "concentrate, int ~ rikez ‘concentrate (tr.)' marci real, reality’ > hitnaneS "be realized, becone a reality’ ~ minek ‘realize (er.)" zalylin ‘penis’ > hézdaynu "they engaged in coitus’ ~ zéyen '“lay" (a worzan) Other intransitive verbs, i.e. non-inchoative (or non-ingressive), non-reflexive, non- reciprocal, etc., are realized in piel, which is the unmarked binyan for innovations. ‘Thus , (10) blog "bluff, lie (N)’ > — bites Lie" hater ‘cat, one whocon- > iter "complain" plains like a cat’ ntjes ‘a nag! > nije ‘nag! koxay ‘star (N)' tixev ‘star! AALS, 1178 Shel Bolozky [AAL 5/3 ‘As we saw above, the same also applies to quadriliteral roots or quintiliteral ones that could not be realized in lig*ét for phonetic reasons, but which would have been pronounce- able in hétpa’et: (QI) tagkid "job, function’ > xantaxi3 "nonsense! > tigked xintred "function" ‘speak nonsense" *hitghid and “iixnteid would not have been prounceable, but *hittagked and *hiteantred are. Semantically, however, they do not belong in hétpa*et. 3.1.2.2, Transitive Denoninative Tri literals Note that whereas for quadriliteral nouns, owing to problems of pronounceability, transitive realization was normally in pi?e€, the realization of triliteral nouns as transitive verbs is more of a problem, since both péet and hégd¢ may be transitive. Furthemore, transi- tive verbalizations in both biwarém may have a causative sense; and in deriving verbs from nouns and adjectives, one expects to find lots of causatives. One of the comion causative uses of hig? is ‘cause somebody to vo something", eg. (22) raked yabav hagac To form new realized in realized in "dance! ‘sit down' jump causatives of this pa’at or sometimes hegrie: ause ~ hiakid hodiv hikpio "cause "cause (sonebody} to dance" (sonebody} to sit down" (somebody) to jump" type, one would usually take sone existing verb, normally in nég?at (see Ariel (1972) in Appendix I), and have it (43) texan ‘contribute’ ~ ~—hitrim "cause (somebody) to contribute" ?azav "leave! ~ hereziv ‘make (somebody) quit (slang)" dahax gallop" ~ hédhix ‘make (a horse) gallop" xatam —'sign' ~ hextim have (somebody) sign’ Since, however, we are dealing with new verbs derived from nouns or adjectives, such causa~ tives’are irrelevant to this paper. On the other hand, we Ap interested in existing denoninative fig? verbs with the causative~ over-inchoative méaning ‘cause to become N/Adj.", Tike: (14) méeex "king" > hémbix ‘make (someone) king" gadot "big! > hégdét ‘make (Something) bir’ xatas ‘weak’ > hextés ‘make (someone) weak" Interestingly enough, although one finds many existing causatives in pie which also mean "cause to become Adj.', like (5) xazak ‘strong’ > xézek — ‘strengthen (tr)! kacat ‘short > kécen ‘shorten (tr.)" xan "hot" xémen "heat" AALS, 1181978] The Hebrew Verb System 9 recently-forned denominative causatives meaning ‘cause to become N/Adj.', or ‘make into N/Adj." (i.e. causatives-over-inchoatives) indicate preference for hif (16) xarég sharp" muxaSé ‘real, tangible néoxé 'eternal' tagel ‘tasteless, without salt! kved "very soft (slang)' hexaég ‘sharpen (tr.}" hima’ ‘make real, tangible’ hinctax 'eternalize’ hétpie ‘desalinate' hibvéé"® "squeeze , cause to be very soft (child speech) " maxaze "play prin "inside (N)' hémciz "he made (novel, etc.) into a play" hdgnim “he internalized’ vvvvvyy Also, one observes tendencies in children to shift denominative causatives of this type from péel into héf*de (the first is heard in sub-standard adult speech as well): (17) kexew ‘draw (something) nearer’ > hékeiv (cf. kaxov 'near') kirer ‘make (someone) ugly’ > “héx*ix (cf. mexoar 'ugly') If the notion ‘causative' is restricted to ‘cause to BpcoME', i.e. causatives-over-inchoatives, and ‘cause (someone) to 0 something’ (Like hitkid ‘cause to'dance’ etc.), then all non- causative transitive triliteral nouns and adjectives are realized in pied. I cannot tell, however, whether this generalization is captured as such, or speakers follow more specific’ senantit categorization extracted fron groups of verbs within pied, like ‘put, place in N', as in (18i), or ‘use, or bring about use of N', as in (18ii}: (18) (4) makom "place" > mikem ‘put in place’ moked "focus (N)' > miked "focus" zanud "page" > émed ‘arrange in pages" Suk ‘market (N)' > Kivek ‘narket" tik ‘file (N)' > tiyek file" bana ‘stage’ > biyem ‘stage (a play)" (ii) géen "bridge (N)' > géhex ‘bridge (a gap)" mexona ‘machine’ > —méken_—‘mechanize' vast ‘regulator’ > viset ‘regulate! anon ‘noney" > ménen ‘finance’ moked ‘mine (N)' > mike’ 'mine" griz ‘grease (N)' > gerez ‘grease’ zéyin 'ponis' > zéyen lay" (@ woman)" déget "flag" > diget ‘present ams (Like flag)" “ora Schwarzwald also reports having heard kéved and kivéed from kveé {personal conmumica~ tion}. AALS. 1910 Shmuel Bolozky [AAL 5/3 (a8) cont. yaran "wood" > yiter ‘cover with trees" but stamp"! > biyee ‘stamp (envelope)" is ‘man (N)" > %éyeS ‘man! rot “letter! > yet ‘spelled and so on, Most other transitive denoninative verbs which do not (or are hard to) classify into semantic groups are also realized in piel, which again energes as the unmarked binyan. Thus: (19) pézmeon "tune" > pézem "hum a tune’ navat ‘navigator’ > nivet ‘navigate’ nativ "trail, route’ > nite ‘mark trail, route’ xuga "dial (N)" > xiyeg ‘dial’ sémek ‘symbol! > simek ‘symbolize’ diaxdox "report (N)' > —dévéax report! vaday ‘certain’ > vide ‘validate’ 3.1.3. Occupied Slots Besides the above considerations, possible realizations are checked against existing items, and if the slot concerned is already occupied, the innovator normally chooses the semantical- ly-closest binyan in which the slot is still free. Thus, for instance, the adjectives palut ‘simple’ and matun ‘moderate’ can not be realized as causative-over-inchoative verbs in hig, as predicted from the procedure above, since there already exist hig?it realiza~ tions for both of these roots—-hégSét ‘take off (someone's) clothes’ and fémtin ‘wait’. The innovator resorts to the other transitive binyan, pérel: pidet ‘simplify’, miten "cause to be moderate’. Sometimes the occupied-slot consideration coincides with the semantic choice of bénayn. For instance, nécax ‘eternity’ is realized as hinctax ‘eternalize' because in pirel, nicdax is an existing item meaning "win, conduct’, and 4a{e£ ‘tasteless, without salt! becomes hitpée "desalinate’ because of the existence of tinel 'take care’ (of sonebody)'; at the sane time, both are assigned to luig>dt by virtue of being causative 3.1.4. Additional Phonological Factors 3.41. Glides as Second Radicals in hég?ie In addition to the unpronounceability of most quadriliterals (and longer) nouns in bényanim other than pee, purat and hitpa7et, thre are also language-particular restrictions on certain sequences of consonants which force choice of one bényan over another. Thus, @ "the fact that bul, %3 and °ot are realized with glides indicates that they are con- ceived as originating from 'defective' roots in which the glides are not realized in certain environments. Ruth Bernan coments [personal communication} that derivations Like but > b{yet being, 30 common in spite of their apparent opacity to the native speaker proves that "defective glides are still viable in the morphology of Modern liebrew. AALS, 1201978) The Hebrew Verb Syston noun Like xayat ‘soldier’ would be expected to be realized as a luifé causative, but since except for one word, hexya 'he revived", Hebrew glides are not realized as second radical in (agral, y would be expected to delete, as it normelly does in hif7ée: the hég?il reali- zation of the root k-w-m is heim ‘raise’, that of b-y-n: hevén ‘understand’, and so on. ‘The form hexil, however, is already occupied (‘cause to apply' or ‘hope’, or even ‘contain’, if the orthography is ignored), and furthermore, the original noun xayat would hardly be recognized. So the other transitive binyan is chosen, and we get xiyet "he mobilized, enlisted" Note that realization of biliteral nouns as transitive verbs in pied rather than hé(2é can partly be explained by the same constraint. In those cases whore y is introduced, Like tik > tiyek, *ot > *éyet, that y will be unacceptable in hég2il, for the sane reason that *héxyi€ was ruled out, and if that y were Nor introduced, the resulting hi{?dl forms would have been confused with foms with deleted /n/, ¢.. /higpil/ "he caused to fall’ > hénét. Could it be said, then, that all biliteral nouns (aiso site > Sivek and diax > diveax, where historical w is Tealized as v) are automatically realized in pé*et for that reason? It may very well be true, if the requirement for deletion of glides as second radicals in hig?ek can he firmly established. ‘the existence of hexya above and hexvir 'becone pale’, from xiver. 'pale' (v originating from historical w) indicate, however, that this require nent is not absolute. Note again that these “defective” glides are not as opaque as generally assumed to be (see Herman's coment, footnote 11). 3.1.4.2. No Stem-Geninates One would have expected manaké 'real' > himné’, but Hebrew, though permitting geninates across morphene-boundary, docs not allow then Stem-internaily. Instead, we get mine’ "realized, tr.', then. 3.1.5. Sporadic Analogies It sometimes happens that a certain noun or adjective is verbalized in some bényan, and then another two or three nouns of the sane semantic field follow, even though the meaning relationship is not precisely the sane (or has specialized in some unexpected manner). Thus maxaze 'play' > fimuz ‘he made (novel, etc.) into a play’, which was realized in hig? by virtue of being a true causative, probably! triggered sétet ‘movie’ > hisrét "he shot, or projected, a movie’ and €éxan ‘tune! > hixin ‘he composed a tune'. It seems to ne that *Sitet (or’*seret) and *Léxen would have been chosen if not for the analogy with the realization of maxaze as himcéz, since they are not causatives semantically. 3.1.6. Preserving the Structure of the Original Noun/Adjective Varlier it was noted that quadriliteral (or longer) nouns are not always necessarily realized in piret (or purat or hitpa°et, for that matter). Generally they ARE realized in pirel (or purat, or hétpa’et), since realization in hifi is often quite unpronounce- able, owing to the stem-initial consonant cluster characteristic of hég7it, nigral, and the Future of pa*at. Note, however, that among recent verbalizations, one finds at least two quadriliterals that are realized in hég?ce (20) Spric > hibpric ‘he squirted’ Sung impetus (colloy)' > —fuilving ‘he ran very fast “19 the best of my knowledge, iimeéz indeed preceded both héstit and hélxin. However, since att of the three are innovations, I have not been able to verify this yet. AALS, 1212 Shmuel Polozky TAAL 5/3 ‘This happens in spite of the second being intransitive and the first non-causative transitive (with an underlying object). The reason is quite simple: piret, i.e. *Sépree and *Siveng, would have broken the conso- nant Clusters of the original nouns and rendered them unrecognizable. Yet innovation pre- fers the origin of the denoninative verb to be as transparont as possible, and since con- sonants are essential for recognition of meaning (the original vowels are’ gone away— replaced by the bényan vowels), breaking consonant clusters is avoided if possible. Thus, original vouzus can frecly be deleted—for instance, tétgen ‘he telephoned’ gets rid of extTa vonels in the source noun té£egon without seriously affecting its recognition, but the Spr cluster of consonants in Sprie remains unaffected. Obviously, the sten- of hégrit makes it even easier to recognize Spric in iéXpric, but preserving the original vowel is got that essential. Note that in a sense, the reluctance to break clusters in Sptic and Sung can be interpreted as assignment to'biliteral roots: Spt-e and u-ng. Similarly, we have some triliteral nouns recently verbalized in hig*d where piel realiza- tion would have been expected: (2) flék "blow" > higeik — "give a blow! Smit "blow > hébnix give a blow! {toc "fart (N)* > higdic "fart" Svie 'bragging' > uébuée brag" ‘The first two are non-causative transitive, and the other two intransitive. Obviously realization in hif*il is intended to keep the original consonant sequences of the source- nouns as transparent as possible. At least for borrowed words, then, we have a tendency on the part of innovators to preserve the relationship between nouns and verbs derived from them, so as to have a kind of para- dignatic uniformity across syntactic categories. This tendency may even result in morpho- logical processes affecting a bényan so as to allow preservation of original consonant Clusters. Thus, since (it 'flirt' cannot be realized as *igtirt (because according to the procedure for quadriliteral intransitives outlined above it does not belong there and because—except for hiSving, where ng is conceived as a single segment—verb-stens cannot end with a cluster of two consonants), it can only be realized in pi°el. liowever, since realization as *{ltet would obliterate the original clustering of consonants in the nou, the innovator resorted to reduplication of the last consonant and sticking the stem- final vowel in between the two identical consonants—{lirtet ‘flirt (V)'. Thus he avoided a hégril realization and at the sane time managed to preserve the original noun intact. Reduplication is an accepted device in liebrey, used largely to express diminution—both in the verb system, for, instance cétkek ‘giggle’ vs. caxak "laugh, and elsewhere (vaxod 'pink" vs. vaadtad 'pinkish'). Indeed, {éxtet may be interpreted as Some sort of diminutive— but I suspect that this interpretation did not trigger its formation: *4iéret would have been diminutive as well, "flirt' in itself being a diminutive of ‘making love'. In other words, it was the need to preserve gLétz that caused reduplication, and the diminutive interpretation is redundant. ‘That the diminutive interpretation is not the triggering factor in such cases can be proved by similar reduplication cases involving triliteral nouns (no other recent verbal- izations involving nouns like (litt are available) : (22) signa digit’ «= > adguen ‘assign digits’ AAL 5, 1221978) The Hebrew Verb System 13 doxnét ‘plan’ comnen’® ‘plan‘ boks "box" > -hétbokses "box {with someone) (slang)" urs ‘course’ > —hithauses "take a course (slang)" une "trick? bkincee ‘play a trick (literary, rather rare)’ Except for the Literary bineve, perhaps, no diminutive sense is implied. It might be argued that normal pi*et realization would not have been possible because of occupied slots— 4éper and tiken respectively (the second radical of pi?eé is normally p, 6, & rather than §,v, x); Still, *bckes, *koxes and *kinec would have been “free” for the Tealization of boks, kuts and kune respectively. Roduplication is quite productive in the language (cf. Bolozky (1972)).'* Among. denoatina- tive verbs, however, the number of reduplication cases is quite Limited. It is impossible, hovever, to evaluaté the significance of this fact, since verbalized nouns with initial and final consonant clusters aro not at all conmon’in Hebrew, and all (or most) of those that exist do preserve the original consonant clusters. Is this significant or accidental? It seems that the best way to test it would be by productivity tests. 3.2. A Proposal for a Verbalization Hierarchy To sum up the picture that emerges from recent verbalization of nous and adjectives, the following general diagram might be suggested: (23) unpronounceable in hif?if, néf?ae and Future of pa?at tranSTeive — inchotive mI rs reflexive Feciproval piret hitpeeee piret bigter ‘button' hit*aznéax "become citizen" tigked "function" tinped "torpedo! hétpatmes "argue about" tigen "phone! “tomnet has a different meaning—‘he programmed’. +t in toxnét is a suffix. “specially for colour terms and a few other adjectives, meaning '... ish", as in: taxot ‘blue’, > txateat, ‘bluish’ fe 'green' > yerabrak ‘greenish’ famen ‘fat’ > Seainan "fattish" for diminution in aninal words, as in: kétey ‘dog’ > klaveav ‘doggy’ xatut ‘cat’ > xatattue "kitten" for some verbs, especially onomatopoei: citeet ‘ring’ > zimen ‘buzz" and so on. AAL 5, 125[AL 5/3, Shmuel Bolozky saaypodyord, vain oveSrary, yaaye Berq, rp0g7y .>8prsg, woz76 orge ams, aay osetd, waxes 737 Pert sonparsuesy a0 sorsnta ae surtesep, 2dx7y {Tear agen, guy seatiesnes \Z saya Stexy epee aaeey B axey, saerrgry 403709 ur S8e8u9, wafypery HEIs, nay ,aIBIIUDDUOD, zaANATY .TT, $2779 sPTP, ¥26ocxry ord Peveizry erd 50 sized-xe3un09 ‘2570A-9TPPTU TOU pUE *Tesordr301 “oarxoTgex seaparsuesquy 2130 soatssauSut pur SOATIEOUDNT LaV30 \2eE auoseq, weg squrd ewoseq, prvary rey Soarqeoysur Aarrenb SOATIeOYDUT-InOToD upwohiuyg Te Wy oTqea2unouoad (re) AAL 5, 1241978] The Hebrew Verb System 15 (25) Unordered factors, applying whenever applicable: (4) Avoid unpronounceable sequence: (i) Avoid “occupied” slots , OF sequences disallowed in Hebrew (ii) If possible, preserve original consonant sequences of source N/Adj. (26) Sporadic factor—analogy with sone limited semantic group (27) Automatic derivation: (4) transitive piret > pur kégter ‘button! > — tugtax "be buttoned" item "place > mukam "be placed’ (i) transitive nigeit > hugeat hime’ ‘make real" > fumaS "be made reat" hitpie ‘desalinate! > futpat "be desalinated" (28) Frequent derivation: transitive péel > middle voice in hitpa*et rukez ‘concentrate (tr.)' > — hétaakez ‘concentrate (int.)" mine "realize (tr.)' > _—shétmame8 "realize (int.)' Note that since the unordered factors do not have to be included in the diagram, the section for "unpronounceable in hég?é2" can be dispensed with. The branching in it is identical to the "pronounceable in all boryanin" diagran—with hég?dl and the cluster- preservation branch excluded. (25i), however, will automatically exclude unpronounce able realization in the former and (35iii) will take care of the latter. (25) and (24) can be collapsed into (29), then (since derivation of middle-Voice hétpa*et from transi- tive pé°et is also almost automatic, reference to middle voice is unnecessary as well): (29) tragsitive intransitive causatives other tran- _colour- incho- other incho- other intran: sitives atives, atives and sitives ingressives physical -human reflexive quality inchoa: tives reciprocal hagri hagrit AAL 5, 12816 Shmuel Bolozky [AL 5/3 4, PRODUCTIVITY TESTS OF VERBALIZATION OF NOUNS AND ADJECTIVES. To find out whether trends emerging from recent denominative verbs are still productive, four productivity tests were run with fifty native Hlebrew-speaking participants (most of whom were students at teacher training colleges) for cach type of tost. 4.1, TESTS ISOLATING (IIE SYNTACTIC AND SENANTIC FACTORS The first two tests were constructed in stich a way as to (hopefully) exclude the phonetic and clash-with-existing-items factors, in an attempt to single out the somantic and syntactic considerations. The nouns Selected were triconsonantal, and contained no con- Sonant clusters that might cause reduplication or use of extended’patterns. Turthernore, they were almost all non-native, to avoid the otherwise-alnost -unavoidable clash with already existing, or at least possible, forms—though it is inconceivable, of course, that one would be able to coMPLETELY avoid any sort of interference by association, or sporadic analogy with some existing verb or some small group of verbs. ‘The first test was supposed to invoke “active” formation of denominative verbs; to represent the innova~ tor's point of view, subjects were given only the noun and target meaning of the verb it should be realized as. The second was intended to represcnt the more "passive" acceptance and rejection by the Speech comunity of proposed innovations: various verbalized forms of the same noun were suggested (in addition to the noun and target meaning), and subjects were asked to choose the one that will best characterize the given meaning. "Although the best testing method would have probably been exclusively oral throughout, the need to simultaneously propose alternative realizations for cach item in the second test made it quite impractical. To avoid interference of spelling considerations as much as possible, and still maintain a uniform method for the two tests, nouns and given meanings were intro- duced orally, but the answers proposed in test II (subjects were asked to re-write their choice) and subjects" own formations in test I were all in writing. A few subjects were given a sample of test I as intended, and a few days later a sample of the same test con- ducted orally throughout, and the results showed no significant difference. In test I, the nouns were read aloud, paraphrased target meanings were suggested (trans- lated in table I), and subjects were asked to fill in their suggested realization for corresponding verbs in very short sentence frames prepared beforehand (cf. Appendix 11), with complete representation of diacritic vowel marks.!? Given meanings related to the Sane noun were not consecutively ordered (they are in the table, for convenience). No strict tine limit was set, but reading of nouns was not repeated at request. Tho same method was used in test I1, except that below cach of the prepared sentence frames a few altemative verbal realizations were proposed (in Hebrew script). These forms were also read aloud, but subjects could sce them all at the sane time. ‘The number of subjects was always fifty, though the total of answers was often less than fifty, owing to subjects’ inability to make a decision in sone cases. The numbers in the table’are doubled to give percentages. To save space, the results of test II are in- corporated in brackets in the table for test I. “5]¢ took sone time to explain that one should not worry about certain distinctions anong the diacritical vowel marks which no longer have a distinctive phonetic function in Modem llebrew. Subjects were told that they could use, for instance, either kanae or patax (both of which stand for a) as long as they used ove. AAL 5, 1261978) The Hebrew Verb System v Results of Tests I and IT ibe, including nieCeat No. Noun, Meaning/Target-Meaning —_ >4#CCeC. hirocic hitecacec or CaCat: 1 Yerig. "sherift'/‘he made hin Eiseg 20 (16) hikes (Banag-2) a sheriff" “idneg 6 72 (82) 2 Berig ‘sheriff'/"he became a hidnig habeancs sheriff! 4 90 (100) 3 Sénég ‘sheriff'/'he made him- hidtaneg® self, proclaimed hin- 100 self, sherif¢" 4 Sénég ‘sheriff'/"he served ines hidbnig hidtancs (Bara g-2) as sheriff" 74 (68) 6 (0) 16 (20) 5S vasat 'vassal'/"he made some- visek hivsek one a vassal" 40 (48) 60 (52) 6 — vasat ‘vassal'/"he became a hévsik hitvasee —nivsat vassal! 2Q) 94 (98) 2 7 vasa ‘vassal'/*he lived as viset hitvaset a vassal! 70 (82) 24 (18) 8 — saéné 'an armchair revolution- séten histin ary' (from saton 34 G30) 66 (70) ‘parlour')'/"he caused someone to become an armchair revolutionary" 9 satbné as shove /*he became an histin héstaten —nistan armchair revolutionary" 2a) 94 (96) 4 10 sa8éné as above /"he spoke like siten histin héstagen an armchair revolution» 56 (52) 2 (6) 36 (42) ary" 11 panet 'panel'/*he covered pind hignit with panels" 68 (70) 30 (24) 12 Rees "hook'/'he fitted with — keres hiknis a hook" 66 (70) 30 (30) “ost subjects explained afterwards that they yould have preferred a causative verb plus a reflexive pronoun, i.e. hiSnif *et *acmo "he made sheriff (of) himself’, Tee ein seatroctionm exitet tr siigieel vockt, Shay bind bo Seance a0 ABtaYaL, AAL 5, 12718 Shmuel Bolozky [AAL 5/3 I was not able to find a colour or physical-hunan-quality triliteral non-native word to See whether it would tend to be realized in hig*é2; otherwise the results roughly confirm the findings summarized above for recent verbalizations. ‘Thus, causatives (i.e. examples 1, 5, 8) are realized in fié*é more often than in pi*ek—although this is obviously not a’universal trend, which might indicate that with pé°et having become a base form, it is more productive even for this type of causative than what actually emerges from récent verbalizations. Other transitives are realized in pie, as is illustrated by 11 and 12. Intransitive inchoatives, other ingressives and reflexives are realized in hétpa?et with almost no exceptions (cf. 2, 3, 6, 9), and other intransitives in piret (cf. 4, 7, 10), though here it might be that hitpaet’is more productive for these other intransitives than what energes from recent verbalizations. ‘The results of Test IT were not different from those of the first test in any significant way—except for a STRONGER tendency in most cases tovards forns already preferred in the previous test, probably since exposure to it strengthens the subject's confidence in the appropriateness of the choice. One case where the tendency weakened was the causative realization of vasaf. 1 think that the stronger preference for vise here (which incide tally is my own choice too) is due to a certain preference for pi*et realization of liter ary nouns 4.2, TESTS INVOLVING INITIAL AND FINAL CONSONANT CLUSTERS Testing whether pronounceability would in fact cause most quadriliterals to go into pire was trivial. Obviously, if a verb simply could not be pronounced in hégil, it would go into piel or huétpaet.’ In the third test, then, the nouns used contained three or more consonants with clusters in initial and/or’ final’position, which phonetically could be realized in either pi7ee, hitpa*el or héfril, The idea was to check three things: (a) whether the choice of bényan as established in the two provious tests may be affected in any way So as to allow preservation of the original clusters; (b) whether within a béwyan, variants preserving original clusters would have preference; and (c) whether the restric- tion noted above on sten-final clusters in the verb (i.e. that a stem-final consonant cannot be expanded into a cluster, e.g. tunpédo > *hitrépd) would be maintained even when original clusters are involved. ‘The fourth test was, again, the sane as the third, except that alternative forms were suggested as verbalizations of given nouns and given meanings, and speakers were asked to choose rather than "create". The assumption was, once more, that the "active" test would reflect innovative capacity while the "passive" one would recreate the normal situation of speakers hearing innovations and rejecting or accepting them by intuition, The Results of test IV are incorporated in the table for test III below, in brackets. AAL 5, 1281978] % of No. Noun, Meaning/ CéCeC’ and Target Meaning extensions 1 snob ‘snob'/*he 40 (36) made him snob- bish' 2 snob ‘snob'/*he became snobbish" 3 talk ‘talcum — 72 (82) powder" /"he sprayed with talcum powder (transitive) ' 4 marks 'Marx'/"he 90 (98) “Marxicized" somebody" S marks 'Narx'/ "he became a Marxist’ 6 patent ‘patent'/ 100 (100) ‘he registered as patent (transitive) ' 7 Enate 'schnaltz'/ 36 (40) “he made some- thing schnaltzy" 8 mate 'schnaltz'/ "it became schmal tzy" 9 zbeng "blow'/ 32 (28) the gave a blow (intran- sitive) 10 *aagaee 96 (100) ‘asphalt'/‘he covered with asphalt (tran- sitive)" 11 apoat "sport'/ 68 (56) "he dealt with sport The Hebrew Verb System Results of Tests UII and IV Distribution of CéCeC forms hirer, séneb 10 (12) hisnib snobeb 12 (22) 60 (64) Sinbeb 6 (2) Pisneb 12 tikek 26 (20) hietie tithe 40 (62) 28 (18) mites 28 (18) himeihs méxkses 62 (80) 8 (2) pitnet 44 (32) pitent 56 (68) Bintoe 6 (8) hidmite Bimete 2 64 (60) Smékcee 28 (32) zibegg 12 hizbing zbiggeg 20 (28) 66 (72) Pishlet 8 (2) hisgere Pésfolt 88 (98) 4 sipret 2 (2) spirtet 10 (4) sportet 56 (50) AAL 5, 129 8 of hét-cacec and extensions 92 (98) 100 (100) 96 (100) 32 (a4) i) Distribution of hit+CaCeC forms histaneb 44 (36) héstnobeb 6 (14) hitsnobeb 22 (38) histnabeb 4 (4) hétanabeb 16 (6) hitnarkes 24 (18) hitmarkses 76 (82) héStamtec 22 (20) hidtmateee 8 (8) héthmatcee 66 (72) histapret 4 héitspantet 4 (4) héistpontet 2 (4) fuctapontet 22 G30)20 Shmuel Bolozky [AAL 5/3 ‘The findings of this test do seem to indicate that the desire to preserve the original consonant Clusters of the source noun is indeed a significant factor, but that speakers tend to try to find ways of maintaining it without disturbing the general procedure for binyan choice outlined above. ‘The most obvious way of doing this is, indeed, by introducing reduplication in pizee. thus, since 3 is not a causative-over-inchoative, realization as titkek leaves it in pi*ee, where it belongs, without breaking the £k cluster, or rather the whole V2 pattern. Similarly, jvé8mateee in 8 and sportet in 11 presérve the structure of Snafe and spoxt, respectively, and are still realized in hitpa%et and piel respectively, the former being inchoative and the latter a non-inchoative non-reflexive intransitive. "Hitlék, hikméee and hiiapiat would have also preserved the original clusters, but as long as thé speaker can find a way out, the basic procedure will normally prevail. ‘Also, in all cases of pé?ed and hitpa?el realizations except for 2, reduplicated forms ourNumpeR others, since they can preserve the original without forcing a shift from one béwan to another. Another way, though not as common, of preserving original clusters in piet is by using the extended 24470 form, c.g. *idneb, *ébmele. It goes without saying that if the baryan chosen in the normal procedure preserves the structure of the noun without requiring reduplication, it Will have preference over forms in other bényarim which also preserve that form—lisnib over snobeb and *Zsneb, in 1, hilmite over IméLeee in 7 (both causatives-over-inchoatives) and *ésfete over fiésgeét (non-causative transitive)—and obviously over forms in the sane bétyan that do not preserve the structure of the noun—pitent over pitnet, *isgeet over ris feet. Another finding: the constraint that disallows expansion of a sten-final consonant into a cluster nay be violated in order to maintain the normal procedure of béwar choice; thus, héiméte is preferred to Sméfcee. Furthermore, we already noted that within the sane bényan, structure-preserving clusters have preference even in stem-final position (see pétent’ and *ésgelt above). Reduplication in pé*et is not the only new device, than; clusters may be preserved by simply allowing stom-final clusters, which is another significant change from recent in- novations to our productivity test (except for héSvégg—cf. footnote 6). Another way in which clusters are preserved in the test for which we have no evidence in recent innovations is the blocking of hitpaef metathesis in certain cases. Hitpa?ee metathesis, which metathesizes the t of the litpa2et prefix with a following coronal ob- strucnt (e.g. hit+saper "have a haircut’ > iistaper) may be blocked in the process of verbalization in order to preserve a ston-initial cluster: Smate is realized as hithmatcee oftener than hiSinatece or héStantec, sport is realized as hitsportet oftener than any metathesized variant, and so on, Each of these devices helps the innovator preserve the structure of the original noun without necessarily violating the established procedure for verbalizat ion. Concerning hizbing, the normal procedure is violated here; as an intransitive verb, it should have been realized in hétpa*e?, or pé?ee. It appears, however, that we are dealing here with a case of "sporadic analogy", which is quite obvious to any speaker of Hebrew, with the two recent derivations mentioned in (21) above sik "blow > higlih ‘give a blow’ Simin "blow" > hédmin "give a blow! which themselves may have been formed in analogy to native hinbic ‘give blows’. AAL 5, 1501978) The Hebrew Verb System 2 It seems to me that although it was preservation of original clusters that motivated forms like hékpric, higtik or hifvic being chosen as Tealizations of Spric, 4léh and bude respectively, the third productivity test shows that switching to a different binyan to achieve this is not as likely any more if it involves deviation from the general procedure. In other words, it is possible that when Sptic etc. were realized as verbs, the general procedure was violated, but now that speakers are aware of other devices for preserving Clusters, like reduplication in pc?et, they will not deviate from the gencral procedure if they can avoid it. A final point: Ora Schwarzwald pointed out to me (personal communication) , that except for spoat, initial clusters show significant preference for hig*ét. I still feel, however, that although this seoms to indicate priority of cluster-preservat ion over syntactic and semantic choice, the fact that all cases concerned function as causatives is not accidental, and’ that the behaviour of sport is the proof. ‘The results of test IV were again not significantly different from those of the "active" test, except that as in test IT, most forms preferred in test IIT were supported even further, and that reduplication’constituted a higher percentage of the cases, since sone speakers were not aware of its potential unless they were actually reminded of it. 4.3. VARLATIONS IN. HIERARCHY Ro-examination of the results of the four tests, primarily the less decided cases, where there was no clear predominance of one bényan, Suggests that it may be possible to attribute the DISTRIBUTION of forms to sone systematic strategies. For instance, when 60-70 percent of the speakers opted for gil realizaiton for causatives, the same 25 participants were consistent in making this choice throughout, and when piel was chosen for the sane proposed meanings, in roughly 25-40 percent of the cases, 15 speakers were consistent in all cases but two (counting all four tests). Similarly, when 5-80 percent opted for non-inchoative non-reflexive non-habitual non-reciprocal intransitive realization in pi*ee, 32 speakers were consistent about it in all cases but four, and most of the rest (15 except for four cases) consistently preferred hétpaet. A Similar conclusion could be dran from realization of non-causative transitives in pet (665-1000—30 speakers consistent) or f§*d£ (roughly 20-30% except two cases—9 speakers consistent in all but two cases). Since a considerable degree of consistency anong speakers is involved even when the re- sults show a clear split, it is very possible that there exist systematic differences in word-fornation strategies among speakers of Hebrew. It was suggested to me by Yael Ziv (personal communication) that perhaps speakers differ systematically in some rearrange- ment or simplification of the hicrarchy in (29) above. Interestingly cnough, all the 15 speakers who were consistent in choosing péet for causatives were among the’30 opting for pi2et for non-causative transitives, and the 15 who preferred fitpa%et for non- inchoative (etc.] intransitives were ali included anong those realizing inchoative (etc.) intransitives in hétpa’et. In other words, it seems that for some speakers, the TRANSITIVE branch of the hierarchy was simplified to a single piel branch, and that for another group the INCRANSITIVE branch got simplified by one branch. The two fifteen-speaker groups overlap only partly (ten out of fifteen) which means that each of these simplifi- cations could either be dependent, on, or independent of, each other. Still, one could probably speak of something like "dialectal variation” in verb-fornation strategies in Modern llebrew. AAL 5, 1312 Shmuel Bolozky TAAL 5/3 5. CONCLUSION It was shown here that the process of verbalization of nouns and adjectives is a complex one, involving syntactic, semantic and phonological considerations and requiring various global considerations. A verbalization-hicrarchy of these factors was proposed on the basis of recent innovations, and was supported by means of productivity tests. It was clained that although new strategies were introduced, they have not destroyed the essen- tially-semantic and syntactic base of the traditional theory of verb-formation, but rather built on it within its framework. Tt was also suggested that for some speakers, the sys~ ‘tem of verb-formation may be sonewhat simplified. APPLNDIX.L SOME RECENT LITERATURE ON THE REALIZATION OF ROOTS IN THE DIFFERENT BINVANIN A modern Linguistic approach to bényan selection was attempted in Ariel (1972). His contribution Lies primarily in providing a more precise definition of some of the rela~ tionships among realizations of roots in different bényaném, and in emphasizing the notion of change-of-state (ingressive, in his terminology) in capturing these Telation- ships. Thus, he comments on the fact’ that nég*at often expresses change-of-state from paral, for instance ne*emad ‘he stood up’ vs..2anad ‘he was standing’, and that the iugrdk comterparts of the sane nif?at forms Will be corresponding causatives, for_in- stance he*emid "he made (somebody) stand’. He further comments on the fact that hitpaee forms often indicate change-of-state from adjectives, for instance hitkacer ‘it shortened’ from kacax ‘short’, and that corresponding pi7ef forms, like kécer "he shortened’ in our case (and obviously kucar, the pu°at form), are almost’ automatic. Herman's (1975 a and b) attempt is considerably more comprehensive. Having examined all possible realizations of listed verbal roots, she establishes that’ the approxinations of School grannars are generally supported, except for a few, though rather important, changes. She reaffirms the "independent syntactic functions of the bényanim as specified above, except that pa°at is Nor the only base form—pé*et also is, since there are many piPet' verbs that do not have pa*al counterparts and certainly would not be derived from pa°al or from any other conjugation, and also because piel is very productive in innovations. With regard to semantic characteristics of groups of verbs realized in certain binyanim, her findings again confirm some of the specialized meaning-groups above, except that she would further classify some of then; for instance, change-of-state in nigral is subdivided into change-of-state regular verbs—no?ecax 'stopped’ and change-of- state copular verbs~nehegax ‘became’, Also, she would add new groups and change ordering in tems of frequency; for instance, in pi*eb, she would adi a group of causatives, Like xémem ‘he warned", and would give it preference over the "intensified pa*at” group: A piece of work concemed essentially with productivity, that is, describing innovations rather than classifying all existing lexical itens, is Sivan's (1963) doctoral disserta- tion. It is full of interesting information, but has one serious shortcoming: except for a few occasions, it does not distinguish between colloquial and literary innovations. This would not have been so problematic, had the literary illustrations been reasonably accepted in everyday usage; the difficulty is, that is some cases, primarily in the less productive binyanim, thenajority of forms are the inventions of a particular writer which nobody else ever uses, and which in sone cases are not even understood by anyone else. I do agree with some of his observations, particularly with pé*ee becoming the most pro- ductive binyan, but not with the mixing of levels in sone of his data. Sivan is the only one to consider non-semantic factors, primarily phonetic, phonological and orthographic, in verb-formation. He claims, for instance, that the traditionally-basic AALS, 1321978) The Hebrow Verb System 23, pa*at is no longer productive, not because most of its slots are “occupied”, but because its unmarked past fom is represented orthographically by consonants only, and contains no affixes (c.g. kadar ‘tic’ as t-3-n, Sagat ‘close’ as 4-g-n). ‘The unmarked past form of piel docs not contain suffixes either, but in Modern Hebrew orthography is represented with the sign y, which stands for either y or < (only a and ¢ are NEVER represented in the modern orthography), which appeals to imovators. Sivan believes that innovators are heavily influenced by spelling, and that they prefer forns with distinguishing marks (such as affixes or vowels represented by glides in the orthography), partly because they them- selves notice then mre easily, but more importantly—because they wish readers to pronounce ‘them as intended, They would use paat only in poctry, since in poetry the vowels are fully represented by diacritics. But notice that Sivan's argunentat ion here revolves around the written language, and even if he had the spoken language in mind, he still con- siders its representation in writing to crucially determine the choice of the binyan. Concerning the anazing productivity of piel, and normally related pu*at and hétpaet forms, a5 against other bryan patterns that’do contain affixed and orthographical ly- tepresented vowels, Sivan's explanation is phonological. Quadriliteral roots—which in general were quite’ scarce in pre-Nodern liebrey—could only be realized in pé7et, purat, or hitpa%et when expanded into CéCCeC, CuOCaC and héteCaCCeC respectively. Furthermore. there existed a group of verbs in these bénjanén, in which the first radical was redupli- cated so as to begin the final syllable of the stem, c.g. gitget ‘he rolled (tr.)' (from the root g-t-C), guégat 'it was rolled", hétgatget ‘it rolled (int.)', Quadriliteral realizations were also forned with the increased productivity of the derivationgl prefix ABV+, which was ALSO restricted to these three binyaném only, the result being 3éCcec (Girbed the enslaved"), SuCCaC (Burbad "he was enslaved"), hiSiacceC (after metathesis; hékza*bed ‘he bocane a’slave').° Whether the reason was réduplication or increasod pre- fixation, or because the root was quadriliteral to start with, the way was opened for further incorporation of quadriliteral roots in pé*et, purat and hitpaee. In Modern flebrew this trend considerably extended, particularly owing to the formation of a few more quad~ Tiliteral "sub-bényanén" in the three binyanén concerned: é+CCeC (*vxen "diagnosed") , tieCCel (tidtek "he fuelled") etc. (I mention pé7et only, but in most transitive piree cases, realization of the other two binyanén is also implied). So whether an actual fourth radical was involved, or a derivational affix was re- interpreted as part of the stem rather than as a bényan affix, the three binyanim concerned were partly reconstructed to allow four sten-consonants, and since these bényandm are the only ones allowing four consonants in the stem, they are naturally preferred when a noun containing more than three conso- nants is involved. It is interesting, however, that Sivan fails to explain wty four radicals could be possible in pé*ee in the first place. There might be senantic reasons for each of the particular cases, but there are also systenatic PHONOLOGICAL explanations. First, Pre-Modern quad- riliterals were realized in piel, pu2al and étparel in analogy to the (now lost) gemin- ation of the second radical of these conjugations, c.g. tirgem 'translate’ based on dibber "talk", etc. Furthermore, as pointed out in 3.1.1 above, the disyllabic structure of piel,” puraé and iitpa*et’stens can accomodate more consonants than can the monosyllabic Stems of iéf*if and the future of parat. Concerning the question of why pie is more productive than iit, Sivan suggests, correctly, that denominative verbs tend to preserve the original form of the noun aS far as possible. ilis argunent is that the prefix hi of hig*d£ would tend to obscure that form more than the pied vowels do. As shown in this paper, the principle is valid, but I think the explanation here is’ weak—prinarily because,’as Sivan himself admits, iug2ée sometimes preserves the original form of the noun in the best possible way, as’ in words ike hidpre squirt’, ifdvce, ‘brag’ MéLcE ‘hit", where the source nouns are Bprie 'squirt’, Evée "bragging’, and cde ‘a blow’ respectively. AAL 5, 158uw Shmuel Bolozky [AL 5/3. APPENDIX. IT SENTENCE FRAMES FOR TESYS 1 aND 11 (Proposed forms for test II in brackets) Lew *ote bekéax ‘he made him sheriff by force (forced him to becone sheriff)" (Sireg—iSaig Karas niSnag hid taneg) 6. harikar ___bekirax hansibot 'the farmer became a vassal due to circumstances" (vasae —hétvasee —visel —niusad hive) 11, lu ___ *et hakéa "he covered the wall with panels* (panat hétpaned ——é{mée pine nga) 5. baat ha?axuza ___ ett haéhax "the lord made the farmer a vassal’ (hitvaset vast —ivade «= vések ~—nivsatt) 9. bemesex hazman hu ‘with the time he becane an armchair revolutionary’ (silen —histin satan —istaen —néstanl 3. hu ____ beewra Lo xubét "he illegally proclained himself sheriff* (Sarag —niSnag Sines —idnég ——éBtanesg) 7. fw bararuza ek habaron *arba*in Sana ‘he served as vassal at the baron's estate for forty years" {hitvaset nivsal vised vasat hivsdt) 8. hansibot ____ ?0to ‘circumstances made him an amnchair revolutionary’ (satan histaten nistan hislin si€en) 2 hw Lifney Snatiyin ‘he becane sheriff two years ago" (niinag —hedtaxeg —Barag —hidnig «Benes ‘12. hu et haxaka "he fitted the fishing rod with a hook’ (heres hanas —fuébenis—ithanes ~—niknas) 10. hu ot ha*énev ‘he spoke revolutionary-style all evening" (satan nistan séken histaten hiskin} 4. fu ___ betéksas ad moto "he served as sheriff in Texas till his death" (uéinig — anag—ibrag iS raeg — Sunes AAL 5, 1381978] The Hebrew Verb System 25 SENTENCE FRAMES FOR TESTS TIT AND Iv (Proposed forms for test IV—in brackets. Very un- Likely suggestions, like fitpa’et forms for transitive realizations, or pi*et forms for inchoative verbs, were not proposed). 1. hama*anad haxada’ Seto tote Legimee ‘his new status made him utterly snobbish’ (saiad —iandb sine —snobeb sabe) 3. fw ___ *et haragtayém Seto "he sprayed his feet with talcun powler" [tatak tiek tthe hte) 8. hadia ___ "the song became schnaltzy' (Ganate —iStantee —ndbmate —étKnakeee —_—huékmé£c} 9. ha do bekéax ‘he hit him (intransitive) with force’ (zébeng —zabaag sizing —zbingea) 2. fu _Legimre "he became utterly snobbish" (nisnab —réstaneh —feéstnobeb —sanab——étanabeb —fuitanobeb —héstnabeb} 4. hw *et xfyim ‘he "Marxicized” Chayin' (hémiks —mixeks —mitkes —minkses —maraks) 10, hu __ *et hakvéS "he covered the road with asphalt ("“asphalted” the road)" (Pisgoet spine —isgeet) 6. hw *et hahamea?a "he registered the invention as patent ("patented” the invention)" (higtint —pitnet —_pitent) Se hu van mézman "he became a Marksist long ago’ (hitmarkes heétmarkses ——tmaneks ——némnaks) 7. hu ____ et haséger "he made the book schnaltzy" (Samate Bimeec ——Smilece —Bimete-—heéBméto} 1. hw kot hayom ‘he dealt with sport all day' [sépret histapret spintet histportet sportet hiteportet hétepartet histpantet] AAL 8, 13526 Shmuel. Bolozky The Hebrew Verb System JAAL 5/3 1978] REFERENCES Ariel, Shlomo. 1972, "The function of the conjugations in colloquial Israeli Hebrew." Bulletin of the Schoot of Oriental and Ajnccan Studies 35:514-30. Berman (Aronson), Ruth. 1975a, “Lexical representation of verbs, roots and conjugations {in Hiebrew)."" In J. Grnan and B, Fischler (eds.1, The Rosen Book, pp. 25-36. Jerusalen, Berman (Aronson), Ruth. 1975. "The morphological realization of syntactic processes in the Modern Hebrew verb system [in liebrew]."" Hebrew Computational Linguistics, 9, pp. 25-39, Bar-Ilan University. Israel. Bolozky, S. 1972. "Categorial Limitations on rules in the phonology of Modern tlebrew."" Doctoral dissertation, University of Illinois. Gole, P. 1976. "A causative construction in Modern Hebrew: theoretical implications." in P. Cole (ed.}, Studées in Modern Hebrew Syntax and Semantics, pp. 231-47) Amsterdan: North’liolland Publishing Co. Oman, Uszi. 1971. “Hebrew grammar.” Encyctopoedia Judaica, Vol. 8, pp. 77-175. Jerusalen: Keter. Schwarzwald, Ora. 1975. “Comments on root-patterns relations in the Hebrew Lexicon." Lin fiebrew}."" Hebrew Computatéonal Lénguistics 9, pp. 47-59. Bar-Llan University. Israel Sivan, Reuven. 1963, "Patterns and trends of Linguistic innovations in Modem Hebrew-— general introduction and part [=the Verb [in tlebrew|."" Doctoral dissertation, licbrew University. AALS, 130AFROASIATIC DIALECTS Faitot: Thomas G.Penchoen (University of California, Los Angeles): Berber Advisory Boar Giorgio Bucceliat (University of California, Los Angeles: Akkadian RssellG. Seu (University of California, Los Angtles): Chadie Stanisla Sepert (University of California, Los Angeles): Northwest Semitic Aftoasiatie Dislects (AAD) seeks to provide concise descriptions of individual languages which belong to the Afroasatic language funnily. Its primarily directed toward an audience consisting, on the one hand, of students of one or several [Afroasiatic languages, nd, on the other, of students of linguistics. In these volumes, both these groups should find su inctteatises such a8 to provide familiarity withthe base structure of the language in question in a comparative perspec tive, Hach description will he comprehensive in scope and suficiently detailed in exemplification. But atthe same time the aim will be to cut through to the essential and to avoid specialized argumentation. The goal then is neither to publish 2 compus of exhaustive referonce grammars nor to provide a platform for the analytical defense of theoretical questions. In this snse the sores is property data-oriented, Though the authors will necessarily be of a vaiety of theoretical persuasions and cach will have his own sct of proferences for presentation, not the least important goal will be to achieve as high 2 degree as posible of uniformity in structure, and in the conventional signs and terminology used. This being accomplished, the readetshoukl have no difficulty in finding points of resemblance and divergence amongst the languages which concern hh with regard to some point of inquiry. The term ‘alects' inthe serie’ tile refers not only to moder spoken vernacu: lars but to historically definable stages of any language ofthe various branches, Publication of studies of at many’ such dia lects as possible would provide, we fee, both an encouragement to comparative work and a sound documentary base on whieh alone thiswork may feutflly progress. AAD 1 - Berber: TAMAZIGHT OF THE AYT NDHIR by Thomas G.Penchoen. 1973, i¥-124 pp. $8.50, The Ayt Nubhie dialect which is described belongs to one of the major Berber languages, Tamazight, spoken inthe Midle {Atlas Mountains of central Morocco. The description is based in the main on research undertaken with native speakers of the Ayt Nahic territory surrounding El Hajeb, — While diected to the non-specialist, a numberof points in the desesiption proper will be of interest to the specialist as well: the presentation of noun and verb morphology points up 8 number of regularities which more often than not have been obscured in previous descriptions. Also, phonological rules ate given hich account fie the major share of morphophonemic complexities. The reader will find inthe appendices and ‘optional’ sections conjugation tables of typical verbs-including detailed observations on the placement of shwa in verbs a chart showing the main morphological patterns involved in verb derivation, a description of the phonological rules appiying in complex sequences of morpheies of the erb group, the "basic" vocabulary contained in several wellknown lexicosttistic word lists, and a chart ofthe Tifnay alphabet used by the Tuareg. AAD 2- Ancient Egyptian: MIDDLE EGYPTIAN by Jobin Callender. 1975, 150 pp..810. This grammar deals with the iterary language used in Egypt from ca. 2000 to 1200 B.C. and considered in even later times tbe the clasical written form of Egyptian. The book is directed toward the gonerl linguist as well asthe Egyptologist: examples ave glossod and weitten in transcription and there san index of grammatical ters and Egyptian morphemes. A comprehensive sot of paradigms ofboth verbal and non-verbal predicate types is included as an appendix. together with an appendix on negation and one on the historical origin of certain cnstructions.- The grammar eontains three main pats phonology, morphology, and syntax, of which the lst receives most emphasis. The section on phonology sketches the laws of sound change to the extent they ean be discovered, The section on morphology stresses the paradignatie character ‘of verb tonses and thei derivations, A distinction is made between truly paradigmatic tenses and tenses borrowed from Old Egyptian for quotations or special effect. Following Polotsky, the “ermphatic forms” are treated as nominalizations under the ubric “manner nominalizations.” Unlike previous grammars of Egyptian, this grammar discusses syntax accor ing to transformational categories. The process of “clefting” interrelates emphatic forms. the “participal statement” and constructions with pre + ratves, The process character of negation is emphasized, and the implications of wo considering itare developed in» special appendix. A sample text ir als included, accompanied by a vocabulary and a translation, AAD 3-Semitie: DAMASCUS ARABIC by Arne Ambros. 1977, vi-123 pp., $13. Based on both previous works and the author's own observations, the grammar describes the Sedentary Kasten Arabi dialect spoken in Damascus. While strietly synchronic and written without presuppesing knowledge of classical Arabic, it follows traditional arrangement and terminology as closely ax posible without failing however to do justice tothe indivi: dual traits of the dialect. Appendices deal with 1) the regular reflexes of Clascal Arabic phonemes in Damascus Arabic, and rules governing the reduction of vowels, and 2) a discussion of morphological substitutions which cannot be inter. preted as describing the historical development from Classical Arabic to Damascus Arabic. A prices ate postpaid. Payenent rust accompany oners fom individu. A handling ff 04 wil be charged to vats if nde ot Prepaid. Discount of 207 on al eres recetvd within ne ea of pubation date, Ons om LUNDENA PUBLICATIONS, P.O. Box 97, Malib,Caitoris 90268, 18.SYRO=MESOPOFAMIAN STUDIES A journal devotes civilizations of acho Haan Sei rom late preistony to the fies millennium BA providing an outlet tr the publication of prinry sures aad fon For the anchusulegicl, historical ain Knguisie anys oF pertinent phenome forthe sty of the itor M, Kelly Bucs Asssciate Lait: Olisier Rowaute ot itor: Willian Shelby Miviviry Bont: A Hota, IJ. Gath, G6 Asi vi TH, Hacusen, ML. Vane Hoon Volume One Articles by G. Buccllt, M. Kolly-Buccellat, D, Sshmandt-essert, L. Mount Williams and WR, Sty. Volune Two Issue L: Michael C, Astoue, The Rabbvans A Tribal Society om the Euphrates From Yale Lim 1 Suds Caesar, 12 pps $1.00. stchies fot Io the Benrinte confers and ctl ope of the mse whan Sake the upper Mille Pophates. traces the hoy of the tibet thecal of the Lae Brose Sint cxamite etl th stele te coneporancas et powers. Thon, ae 2 Ln historic 2. it ns the descendants of the Rabie ects wae the sane te, Ins wry Mma pil, dng the tal for Syn et cemry BC: me a Issue 2 Paolo Matthias, Pretiminary Remarks on the Reval Palace of Ele, 20 pp. 8 plates Communiqué of the Halian Archacological Mission to Seria Concerning the First Mecting of the Tuternationad Committce for the Shuly of the Tests of Ebla, 2 yp. S385, This atte prwnts pin pte of the 1973-1975 excavation of ts Tl ManiAh Royal Paice dating. to Tans ore A ex 400-2280 BC). It ws 1978 Hat some 15,000 eu Tube a raze wcre found Ae pac, wane of thewn wat In 2 nw NovteWestrn Swi [Srp The rari metre of the alae fs br essa oe dew and the Bali placed ‘nithin is chroot tod torial amework,Crunicsvidc ace prtani t thy croton is pulsed tore fr the fin tie. The pope (ion re taaton by De Trances Pek) an wa at te 2M Henconte Avsraique Fatt at Birmingham fn July 1976, Issue 3: Piotr Michalowski, The Neo-Sunerin Sifver Rite Texts, 16 pp. 1.35, stay fe tention 1 parte type oF docetsfremn the Ue HD pata whi Hd nv Thee texte ape analy a0 nly he sk Pat 8 {poe achive an in wats to sini st tens Tom other comermponay Boerne hs TW props ht thw ste re part acti orem An heen of ies ite Desc which prose vats pon encountered fat teats as to props he aie Issue 4: Kassem Toveir, The Strign Archueological Expedition 10 Tell AVABL Zrejehe Clay Paeurines of the Dhird Millennium B.C., 15 pp 20 plates, $6.25. ‘cal of some 239 guns fons the Late tied lnsiam sesypation of Tell Ab i Spin in the fond aco the abs de, pry saed ad party rom the sara.” We a conalopeat Seaton can he eb + Iypcl cacao dating dlerent erage of obs By are tn fy epee catcprk Thete af 77 hr urine, 138 sna Heures ad 34 eat fragcate se whcel Comparison wh pail ups mater fom ter sex ae fs Issue 5: Ascad Mabmoud, Terga Preliminarr Reports, No. 5: Die budusirie der ilamischen Keramik caus der eweiten Seauon, 16 pp.. 4 plates, $2.3. tory Asha Mk stuated the south of the tell fant vate Me perimeter of the aie Si. Noms crams wes plan umglieds wot fue eo net he ila. thee HisMINE STYLE SHEET UNDENA PUBLICATIONS The primary goal of Undena Publications is to publish at lowest cost possible while keepit the highest possible standards. In order to do this we need your cooperation when sub- mitting manuscripts. Please read the information below. MANUSCRIPT: Keep a duplicate copy of your submitted manuscript since this will not be included when proofs are sent to you. PROOFS: Each author will receive proofs for corrections. CORRECTIONS AND CHANGES: Corrections and changes must be kept to an absolute minimum. Major changes, ie., changes that affect more than a couple of lines,or a great quantity of minor changes, will be at the author's expense. You will be presented with a bill based on the amount of work involved. All changes as submitted by the author on the proofs are suggestions only and may be disregarded at the discretion of the editor. Changes that affect entire pages will not be accepted. ABSTRACT: All manuscripts should be accompanied by an abstract of approximately 100-125 words. OFFPRINTS AND COPIES: Authors will receive 10 free copies of the work. Additional copies will be available at discounts of 30% for a single order of 5 copies or more, and 40% for a single order of 20 copies or more (a single order is to be billed and shipped to the same address). MANUSCRIPT PREPARATION MANUSCRIPT: All material must be typed, double-spaced throughout on non-erasable and non-onionskin bond; photocopies or xeroxcopies are however accepted. Isolated corrections may be entered by hand, but should be printed. FOOTNOTES: All notes are to be typed, double-spaced, on separate pages with running numeration. TABLES AND CHARTS: Tables and charts must be laid out as desired to appear in print, with explanations in the margin when the typed manuscript does not show clearly the intended format. TITLES: In manuscripts divided into sections, all sections and subsections should be titled and numbered using the following model: 1. (Major heading) 1.1. (Sub-heading) 1.2. (Sub-heading) 1.2.1. (Second level sub-heading) 1.2.2, (Second level sub-heading) 2. (Major heading) BIBLIOGRAPHY: For alll references please include (a) Full name of author(s), (b) place of Publication and name of publisher, (c) page numbers of article.AFROASIATIC LINGUISTICS AAI, includes conteibutions in linguistics within the vast domain of Afroasi Articles of general, theoretical interest using Afro: studies are included. (Hamito-Semitic) languages. je material, descriptive, historical and comparative Editors: Robert Hetzron (698 Zink Av., Santa Barbara, Ca. 93111, U.S.A.) Russell G. Schuh (15337 Hart St., Van Nuys, Ca. 91406, U.S.A.) Volume One Articles by P. Newman, R. G. Schuh, J. L. Malone, R. Hetzron, T. Givon, T. M. Johnstone, B. W. Andrzejowski, and Il. Minkoff. Volume Two Articles by D. R. Cohen, C. D. Johnson, A. Barnea, R. Nir, C. T. Hodge, G. Janssens, S. Segert, J.B, Callender, J. L. Malone, T. Givin, and A. D. Corrs. ‘Volume Three Articles by R. G. Schuh, G, Buccellati, R. Hetzron, J. Saib, R. Steiner, D. Boyarin, I. Avinery, A. Zaborski, E, Rubinstein, and P. Abboud. Volume Four 1, Neil Skinner, North Bauch Chadic Languages: Common Roots, 49 pp. Wd, ‘Flv’ (Noun) and Mouth’ in Afroasiatic, 12 pp. $5.15. 2. Lewis Glinert, umber Switch: A Singular Feature-change Rule in Modem Hebrew, 38 pp.. $3.20. 3. Russell G. Schuh, Bade/Ngizim Determiner System, 74 pp. $6.25. 4. Joshua Blau, The Beginnings of the Arabic Diglossia. A Stuly of the Origins of Neoarabic, 28 pp., $2.35. Volume Five 1. Paul Newman, Chadlie Classification ard Reconstructions, 42 pp.. $3.55. |. D. L. Appleyard, A Comparative Approach to the Amharic Lexicon, 67 pp.. $5.65. 3. Shmuel Bolozky, Word Formation Strategies int the Hebrew Verb System: Denominative Verbs, 26 pp., $2.20. 4. J. Rosenhouse, On the Complexity of Some Types of Complex Sentences in Urban Morocean Arabic and Some Other Arabic Dialects, 15 pp.. $1.30. 5. Janet I. Johnson, Remarks on Egyptian Verbal Sentences, 20 pp., $1.70. Subscriptions to one volume of about 200 pp.: $15.00. Individual issues available separately. Descriptive flyers, with complete list of abstracts and prices for individual issues, are free on request. Write to: UNDENA PUBLICATIONS, P.O. Box 97, Malibu, California 90265, U.S.A.