hebrew verbal passives in later language development · pdf filehebrew verbal passives in...

Hebrew verbal passives in Later LanguageDevelopment the interface of register and verb

morphology

DORIT RAVID AND LIZZY VERED

Tel Aviv University Israel

(Received May ndashRevised August ndashAccepted September )

ABSTRACT

The current study examined the production of Hebrew verbal passivesacross adolescence as mediated by linguistic register and verbmorphology Participants aged eight to sixteen years and a group ofadults were asked to change written active-voice sentences intocorresponding passive-voice forms divided by verb register (neutral andhigh) binyan pattern (Qal Nifrsquoal Hifrsquoil Hufrsquoal and Pirsquoel Pursquoal)and verb tense (past and future tense) Results showed that Hebrewpassive morphology is a very late acquisition almost a decade laterthan in other languages that passivizing neutral-register verbs wasless challenging than high-register verbs and that past tense verbswere easier to passivize than future tense verbs An order ofacquisition was determined among the three binyan pairs The paperprovides an account of these findings grounded in the event-tellingrole of Hebrew passives in discourse and the spurt of abstractlexically specific vocabulary in Later Language Development

INTRODUCTION

Passive voice is among the most prominent and widespread perspective-changing constructions in the worldrsquos languages (Abraham amp Leisiouml Givoacuten Keenan amp Dryer ) The acquisition of passive formsby children has been the controversial topic of research studies acrossthe last decades (Baldie Messenger Branigan amp Mclean )Passive constructions have been described as delayed in some languages(Gordon amp Chafetz Prat-Sala Shillcock amp Sorace ) and asearly acquisitions in others (Allen amp Crago Aschermann Gulzow ampWendt Demuth Moloi amp Machobane ) Syntactic semanticdiscourse cognitive and usage factors as well as elicitation conditions have

[] Address for correspondence Dorit Ravid Tel Aviv University ndash School of EducationRamat Aviv Tel Aviv Israel e-mail doritrposttauacil

J Child Lang Page of copy Cambridge University Press doiS

available at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0305000916000544Downloaded from httpwwwcambridgeorgcore TEL AVIV UNIVERSITY on 19 Nov 2016 at 094941 subject to the Cambridge Core terms of use

been invoked in explaining why young children struggle in comprehendingand producing passive forms (Brooks amp Tomasello Messenger Pinker Lebeaux amp Frost )

The current study aimed to examine the acquisition of passive forms inHebrew in a new light First focus here is on the MORPHOLOGY of thepassive verb As general theoretical motivation for this verb-focusedinvestigation we follow Haspelmathrsquos ( p ) comment that ldquotheverbal morphology associated with a passive construction is an essentialpart of the construction whose properties are worthy of study in their ownright Indeed the passive can be regarded as first and foremost a verbalmorphological category whose meaning implies certain changes in theclause structurerdquo From a Hebrew-specific typological perspectivestudying passive verb morphology in acquisition is key as Hebrew passiveforms are part and parcel of the verb-pattern binyan system a set of sevenconjugations which provide the vocalic form of the verb and itssyntactico-semantic category Second previous research on passiveacquisition ndash in Hebrew as in other languages ndash examined spoken passiveconstructions in young children mostly in preschoolers In contrast thecurrent study elicited passive forms in writing from school-going childrenand adolescents compared with adults during the period of LaterLanguage Development (Berman Nippold ) This frameworkenabled us to contextualize the acquisition of passive verbs as part of a setof Hebrew morpho-lexical classes whose learning is delayed toadolescence ndash a time of when socio-cognitive linguistic and literacyabilities connect together in the complex expression of communicativegoals (Berman amp Ravid )

The main goals of the current study were as follows (i) to provide anaccount of the development of Hebrew passive verb morphology across theschool years based on an experimental task and (ii) to determine the roleof derivational and inflectional binyan morphology in this development

Passive constructions

Passive voice seems to be a universal phenomenon nonetheless it has beendescribed as an elusive category deserving of sound conceptual foundations(Shibatani ) Several conclusions come to mind when reviewing thelinguistic literature on passive constructions and their functional rolesacross different languages which can be useful in laying the backgroundand generating expectations for the current developmental study ofHebrew passives

First passives are noted among the most prominent and widespreadperspective-changing constructions in the worldrsquos languages (Givoacuten) The general view of passive constructions as backgrounding the

RAVID AND VERED


agent role while highlighting the patient role is shared among linguists(Myhill ) We adopt here Keenan and Dryerrsquos () analysis of thepassive as essentially different from syntactically restricted inversionphenomena such as topicalizing and dislocating constructions unlike theformer passive forms are integral to the grammar of languages inoccupying unmarked positions and the case marking of NPs in requiringno specific intonation patterns and in undergoing all major syntacticoperations According to Keenan and Dryer the most distinctive propertyof passives is the formation of the passive verb phrase (Haspelmath )This supports our focus on the morphology of Hebrew passive verbs ininvestigating their acquisitional path

A second observation from the cross-linguistic literature indicates that thenotion of lsquopassiversquo often serves as an umbrella term for a cluster ofconstructions sharing the perspective of the patient Across languagesthese include middle voice and impersonal constructions (Abraham ampLeiss ) in addition to get-passives medial passives and adjectivalpassives (Mitkovska amp Bužarovska Rathert ) all participatingin different configurations of agent and patient roles in the event Viewedfrom a developmental psycholinguistic perspective passives constitute oneform of voice construction whose function is to depict a non-canonicalview on events and situations This can generate the expectation that alanguage learner would need to gain experience not only with variegatedcommunicative events and contexts but also with the array ofagent-demoting constructions and discourse types typically associated withtheir expression in her language

A third conclusion drawn from the literature is the strong relationshipbetween transitivity and voice on the one hand and between aspectand voice on the other (Abraham amp Leisiouml Shibatani )Middle voice constructions tend to express a generic and imperfectiveoutlook eg these books sell well Passive voice in contrast prefers atransitive construction that expresses the patient argument in an objectNP with an agentive volitional grammatical subject and a concreteactional verb affecting the object as in the alternation John built the IKEAchest the IKEA chest was built by John No less importantly expressionof canonical passive voice is associated with realis concreteagentivetransitivity and perfective temporality (Foley ) In a language suchas Hebrew which does not mark aspectual distinctions in verb inflectionalmorphology this means a strong association of passive forms with thenarrative past tense ndash unlike telic past which strongly associates withmiddleinchoative semantics (Berman amp Slobin ) This generates twoexpectations that passivization of abstract less agentive and object-affecting Hebrew verbs should be more challenging to children and thatfuture tense passive forms should be more difficult to passivize than past

HEBREW PASSIVE MORPHOLOGY IN DEVELOPMENT


tense verbs The relationship between resultative predicates and passiveverbs has been shown to benefit learning (Lee amp Lee ) but the tensehypothesis to our best knowledge has not been tested before Theacquisition literature (see below) has long known that actional passiveconstructions are comprehended earlier than non-actional ones (MaratsosFox Becker amp Chalkley ) but testing passive formation in latechildhood adolescence makes it possible to examine a much richer array ofcognitive and abstract verbs

Acquisition of passive constructions

The acquisition of passive forms by children has been the controversial topicof numerous research studies across the last decades regarding the timing ofthe acquisition of passive forms and the reasons thereof (Budwig Crain Thornton amp Murasugi ) Much of the investigation of passiveacquisition has focused on English where it was often described asdelayed to age six years in both comprehension and production (Brooks ampTomasello Marchman Bates Burkardt amp Good ) Childrenwere shown to experience particular difficulty when full passives ndash that isthe whole syntactic structure including the by-phrase ndash were involved (Foxamp Grodzinsky ) The realization that the passive is a complexphenomenon with different parts and facets dictates a more nuanced viewof passive learning as a drawn-out process of learning Many studies haveshown that young children start by attending to semantically restrictedprototypical passives ndash that is irreversible passives with actional verbs andanimate subjects (Ferreira Pinker et al ) Children initiallyprefer less patient-oriented agent-demoting devices such as the getpassive or middle voice (Gaacutemez Shimpi Waterfall amp Huttenlocher )Only later on beyond age six and even later does childrenrsquos knowledge gobeyond the canonical passive (Messenger et al )

Frequency of passive forms and their discourse contexts in the input haveoften been invoked as determining factors in childrenrsquos acquisition ofpassives (Tomasello Brooks amp Stern ) Thus the rarity of agentlessand passive constructions in English input to children (Gordon amp Chafetz) could explain the discrepancy in reports on timing of acquisitionInvestigations of the early emergence of passive forms (around age two) inlanguages with prevalent passive constructions such as Bantu languagesInuktitut and Quiche Mayan have ruled out a universal maturationalconstraint on passive learning (Alcock Rimba amp Newton Allen ampCrago Demuth et al Pye amp Quixtan Poz ) Butchildren acquiring languages with lower frequencies of passive forms thanEnglish such as Catalan or Hebrew are reported to continue strugglingwith passive production even beyond age ten (Prat-Sala et al

RAVID AND VERED


Ravid ) At the same time exposure of child and adult participants topassive constructions of all types has been shown to enhance their passiveproductions (Gaacutemez et al ) Thus the literature suggests thatlearners are sensitive to the way the immediate and general linguistic inputemploys passive constructions in the depiction of different agent-demotingscenes

Passive voice in Hebrew

Syntactically the activepassive alternation in Hebrew is rather similar tothat of English Consider the pair of Hebrew sentences in () whichexpress active (a) and passive (b) voice perspectives on the same event

() a ha-texnay hiklit et ha-rersquoayonthe-technician recorded ACC the-interviewlsquoThe technician recorded the interviewrsquo

b ha-rersquoayon huklat al-yedey ha-texnaythe-interview was-recorded by the-technicianlsquoThe interview was recorded by the technicianrsquo

The active version (a) has a grammatical subject depicting the agentha-texnay lsquothe technicianrsquo at sentence-initial position and a direct objectpatient ha-rersquoayon lsquothe interviewrsquo marked by accusative et following thetransitive verb In (b) the patient serves as subject at sentence-initialposition followed by the passive verb The by-phrase is a crucial test ofpassive voice in Hebrew since only truly passive (as opposed to medialpassive) constructions can take it The scope of the Hebrew passive isnotably less broad than in English as passivization is generally restrictedto transitive constructions with direct sometimes oblique objects UnlikeEnglish Hebrew does not allow passivization of indirect objects inditransitive constructions such as Mary handed the flowers to John Johnwas handed the flowers by MaryTwo language-specific structural and functional characteristics render

the Hebrew passive very different from that of English providing groundsfor further hypotheses regarding acquisition One is its Semitic verbmorphology another is the prevalence of generic subjectless constructionsin Hebrew

Passive verb morphology Like all Hebrew verbs passive verbs are formedwithin the binyan system by the Semitic non-linear affixation of root andpattern morphemes Semitic roots are discontinuous morphemes consistingof three or four radicals which constitute the structural and semantic coreof Hebrew words in general and verbs in particular eg root g-d-l lsquogrowrsquo

or root t-q-n lsquofixrsquo All Hebrew verbs are constructed by combining a rootwith one of seven verb conjugations termed binyanim (literally



lsquobuildingsrsquo) ndash traditionally named Qal Nifrsquoal Hifrsquoil Hufrsquoal Pirsquoel Pursquoaland Hitparsquoel (based on the root p-rsquo-l lsquoactrsquo) Binyan conjugations providethe main vocalic structure of the verb (sometimes accompanied byprefixation) For example Pirsquoel provides the CiCeC template into whichthe root (marked by capital Cs) is inserted In addition each binyanconsists of a specific bundle of temporal patterns that combine with a rootto create its paradigm of temporal stems For example gidel megadel andyegadel serve as the respective past present and future tense stems oflsquoraisersquo in Pirsquoel

The binyan system not only dictates the morphological structure of verbsbut also how transitivity relations including passive voice are expressedsyntactically Thus binyan patterns are associated with higher or lowertransitivity values with correspondingly richer or poorer argumentstructures ndash eg high-transitivity Hifrsquoil which is often associated with twoor three arguments or low-transitivity Nifrsquoal mostly occurring insingle-argument structures (Berman ) In this respect Hebrew binyanconjugations differ from Romance verb conjugations which aremorphophonological in nature (Monachesi ) and are somewhatsimilar to Slavic verb formation which is also associated with aspect andAktsionsart (Svenonius ) Hebrew root-based verbs with differentbinyan patterns constitute semi-productive derivational families combininglexically specific meanings with Aktionsart values such as inchoativitycausativity reflexivity reciprocity middle and passive voice (see Bermanamp Nir-Sagiv p for a detailed table) For example root g-d-llsquogrowrsquo combines with binyan patterns to create a family of six differentverbs ndash two of which in passive voice basic gadal lsquogrowrsquo causative higdillsquoenlargersquo passive hugdal lsquobe enlargedrsquo causative gidel lsquoraisersquo passivegudal lsquobe raisedrsquo and middle-reflexive hitgadel lsquoaggrandize oneselfrsquoPassive verb formation takes place within this morphologically stringent

root-binyan verb system Hebrew passive morphosyntax is based on thethree transitive binyan patterns ndash Qal Hifrsquoil and Pirsquoel each associatedwith a dedicated passive counterpart Qal with Nifrsquoal Hifrsquoil with Hufrsquoaland Pirsquoel with Pursquoal (Table ) In the current context we investigateVERBAL passives that is passive verbs in the past and future tense ratherthan resultative adjectival passives which are based on the present tensestems of passive binyan patterns such as Pursquoal metukan lsquofixedrsquo or Hufrsquoalmufta lsquosurprisedrsquo (Berman ) Adjectival passives constitute a major

Verbs are presented according to the Hebraic tradition in the past tense third personmasculine singular form which corresponds to the binyan name

See Ravid et al () for an account of the two binyan subsystems occupied by Qal HifrsquoilPirsquoel and their passive counterparts

RAVID AND VERED


structural source of Hebrew adjectives learned early on around four years ofage (Berman )

In one sense passive morphology is the most predictable of all binyanfunctions as choice of passive binyan is always entailed by its activecounterpart (Table ) However the three patterns expressing passivevoice are not uniform ndash rather they fall into two distinct groups strictpassives (Pursquoal and Hufrsquoal) and Nifrsquoal Pursquoal and Hufrsquoal passives shareseveral unique features First the passive voice is their only functionalrole so that their existence is predicated on that of a correspondingtransitive Hifrsquoil or Pirsquoel verb Structurally they share the vowel u acrosstheir temporal paradigms and unlike all other binyan patterns they do nothave infinitive forms (Table ) Again unlike all other patterns Pursquoal andHufrsquoal form their action nominals by attaching the abstract suffix -ut totheir adjectival present tense stems eg muxan-ut lsquoreadi-nessrsquo ormersquoorav-ut lsquoinvolve-mentrsquo (Ravid amp Avidor ) Importantly theirpresent tense forms express both passive participial and resultativemeanings so that mefursam in Pursquoal can be interpreted as both lsquois beingpublishedrsquo and lsquofamousrsquo Nifrsquoal in contrast has several characteristics thatmark its special status (Schwarzwald ) First in addition to its role asthe passive counterpart of Qal it has most of the middle functions ofHitparsquoel serving as the inceptive and inchoative counterpart of Hifrsquoil(Berman ) Moreover and again unlike the strict passives presenttense participial Nifrsquoal participates in a tripartite system with Qal and theresultative pattern CaCuC as in Qal kotev lsquois writingrsquo CaCuC katuvlsquowrittenrsquo Nifrsquoal nixtav lsquois being writtenrsquo Structurally too Nifrsquoal doesnot have the typical passive u vowel of the strict passives and againunlike them it has an infinitival stem and a derived action nominal likeall other non-passive binyan patterns Finally Nifrsquoal is the only binyan(or for that matter any derivational pattern in Hebrew) starting with

TABLE The three activepassive binyan pairs illustrated with three roots(l-m-d lsquolearnrsquo s-b-r lsquoexplainrsquo and t-p-l lsquotake care ofrsquo) across past presentand future tenses (in third person masculine singular) and the infinitive form

Voice Active Passive Active Passive Active Passive

Binyan Qal Nifrsquoal Hifrsquoil Hufrsquoal Pirsquoel PursquoalPast tense lamad nilmad hisbir husbar tipel tupalPresent tense lomed nilmad masbir musbar metapel metupalFuture tense yilmad yilamed yasbir yusbar yetapel yetupalInfinitive lilmod lehilamed lehasbir mdash letapel mdash

Side by side with derivation from present tense stems like its strict passive counterpartsCompare hipakdut lsquostate of being countedrsquo (regular action nominal with pattern



prefixal n- with a further peculiarity of having identical past and presentstems and phonologically distinct future and infinitive stems (Ravid )

Hebrew passive formation is firmly embedded in Hebrew verbmorphology using the same structural devices employed for theexpression of other binyan functions There is nothing special about themorphophonology of either the strict passives or Nifrsquoal the developmentalliterature shows that present tense adjectival passives with u are found inchild speech (eg mekulkal lsquoout of orderrsquo hafux lsquoupside downrsquo) Moreovertelic Nifrsquoal verbs such as nishpax lsquospilledrsquo or nirtav lsquogot wetrsquo are amongthe earliest past tense forms to occur in child speech with medial passiveHitparsquoel forms such as hitparek lsquofell apartrsquo soon following in their steps(Berman ) Therefore morphological structure alone cannot be theculprit in the very late acquisition of passive voice in Hebrew We arguethat the problem lies elsewhere at the interface of syntactic constructionsand event structure in the ambient language

Passive versus generic subjectless constructions Hebrew has severalsubjectless constructions that serve to express a non-agent oriented outlookon events (Berman ) Two prominent examples are predicate-first ()and impersonal () constructions which are prevalent in everydayinteractions child speech and input to children (Dromi amp Berman )

() a mutar lexa laleacutexetAllowed to-you to-golsquoYou may gorsquo

b xam pohot herelsquoitrsquos hot herersquo

() a bonim po gesherbuildingPL here bridgelsquoA bridge is under construction herersquo

b lo yimkeru lexa kan glidanot will-sellPL to-you here ice creamlsquoThey wonrsquot sell you ice cream herersquo

Constructions such as those in () and () share a general often modaldiscourse stance (Berman ) Such subjectless often verblessimpersonal constructions usually anchored in the present tense areprevalent in everyday communication and written discourse thusoccupying the preferred slot for the expression of habitual generic statesscenarios and situations in Hebrew Hebrew passive constructions are verydifferent Like their active counterparts and in direct contrast to

hiCaCCut) with nifkadut lsquogoing AWOLrsquo (based on the present tense stem) both based onthe root p-q-d

RAVID AND VERED


subjectlessverbless constructions they require a grammatical subject and arebased on lexical verbs that take all temporal inflections favoring theperfective past tense By itself passive syntactic structure cannot beregarded as the cause of delayed passive acquisition in Hebrew as childrenproduce the required grammatically integral SV(O) structures (Keenan ampDryer ) fairly early including many with non-agentive subjects andunaccusative verbs (Berman ) Relating different structures shouldnot be challenging to Hebrew-speaking children who are early on exposedto and produce sentences with pragmatically alternating word orders suchas () (Ravid )

() a ha-tik nafal nafal ha-tikthe-bag has-dropped has-dropped the-baglsquothe bag has droppedrsquo

b kvar axalti et ha-agas et ha-agas kvar axaltialready ateSTSG ACC the-pear ACC the-pear already ateSTSGlsquoI already ate the pearrsquo

We argue that the FUNCTIONAL ROLE reserved in Hebrew to passive verbconstructions is the gist of the problem In direct contrast to genericpresent tense state-oriented subjectless constructions verbal passivesfunctionally pinpoint highly transitive perfective EVENTS either realis orirrealis as in ()

() a af baacuteyit lo shupats be-maharsquolax ha-tkufano house not renovated in-the-course the-periodlsquoNot a single house was renovated during this periodrsquo

b im tarsquoase kax lo tenuke me-ashmaIf will-doNDSG that not will-be-cleanedNDSG from-guiltlsquoIf you do that you will not be cleared of guiltrsquo

Passive constructions thus require the expression of specific perfectiveevents through an abstract and distanced agent-demoting stance whichdoes not characterize early child language interaction This restrictedfunctional role predicts a prolonged period of learning passive voiceenabled by socio-cognitive changes in adolescence during Later LanguageDevelopment (Blakemore amp Choudhury ) Learning passiveconstructions in Hebrew is based on gaining extensive experience with theappropriate communicative contexts of event- and story-telling and thepassive constructions associated with them and the ability to perceivemultiple perspectives as well as familiarity with literate written languagestyles that prefer such forms of expression (Berman amp Ravid )

We take irrealis as an umbrella term covering non-indicative less-than-real modalityfunctions (Timberlake )



Learning Hebrew passives Most research on Hebrew passive acquisition todate has focused on its distributions in child adolescent and adult corporaVerbal passives of Pursquoal and Hufrsquoal were virtually absent in spoken motherndashchild interactions child-directed speech child speech childrenrsquos peer talk(Berman Ravid et al ) and in childrenrsquos spokenpersonal-experience story-telling (Berman amp Slobin ) They were alsonegligible (under ) in the written narrative and expository texts ofHebrew-speaking high-schoolers and even university-educated adults(Berman amp Nir-Sagiv ) Past tense (no future tense) Pursquoal and Hufrsquoalverbs constituted under of verb types and tokens in childrenrsquos story-booksand early school texts Passive Nifrsquoal past and future tense forms in the samecorpora were as sparse However passive verb usage was noted as a prominenthigh register marker in the expression of detached abstract discourse stance inadolescent and adult discourse production (Berman amp Ravid Ravid ampBerman ) with special concentration in adult narrative writing (Ravid ampChen-Djemal ) This supports our hypothesis regarding the special rolepassive constructions occupy in Hebrew event-telling and the drawn-out routeto learning their usage contexts in the language

Ravid () reviewed several small-scale Hebrew-language experimentalstudies where past tense passive constructions were elicited in school-agedchildren They all showed that by age ten syntactic errors in passivesentences were negligible but correct production of passive morphologywas still at A number of studies that elicited passive forms inmorphosyntactic tasks had the same results (Ravid amp Geiger )showing correct production of passive verb morphology at ceiling only bylate adolescence (Ravid amp Saban ) Relatedly Ravid and EpelMashraki () reported strong correlations between passive productionprosodic reading and reading comprehension in nine-year-olds Across allof these studies Nifrsquoal had the highest correct scores and attracted themost errors leading us to assume that it constituted the bridge leadingtowards strict passives given the prominence of intransitive telic andchange-of-state Nifrsquoal forms in early childhood (Berman )To sum up corpora studies indicated the marked absence of passive

constructions from spoken and written Hebrew texts except for thespecific adult preference for narrating events from a distanced discoursestance Correspondingly passive verb production in experimentalconditions outlined a learning path starting very late around age ninereaching command only by late adolescence Given childrenrsquos commandof Hebrew verb morphology and argument structure in early childhoodwe hypothesized that the delay in learning Hebrew passives does notderive from syntactic nor morphological factors but rather from the rareencounters with Hebrew passive constructions that are the direct outcomeof their specific narrative role coupled with a detached and general stance

RAVID AND VERED


Linguistic register A critical component of the current study is the notionof LINGUISTIC REGISTER characterized by Ferguson ( p ) as ldquothelinguistic differences that correlate with different occasions of userdquo Thismeans that acquisition of register involves gaining command of the rangeof expressive options available in the target language and being able tomap relevant linguistic forms in accordance with communicative context(Biber ) Moreover register-sensitive usage is more aligned with thedisplaced writing mode which allows for more planning and monitoringhence more complex linguistic features (Halliday ) Accordinglyverbal passives occupied a prominent part of the elevated Hebrew registerin Ravid and Bermanrsquos () analysis of text production across adolescence

Hypotheses in the current study

Against this background the current study was the first systematiclarge-scale dedicated study of Hebrew passive elicitation in sententialcontext to engage school-going populations ndash children and adolescents ndashcompared with adults The passivization task included two new variablesstudied for the first time suitable for examining the language of literateparticipants during the period of Later Language Development First thevariable of linguistic REGISTER that is language level Register was used asa measure of lexical specificity and degree of abstractness of the verbfollowing the criteria established in Hebrew by Ravid and Berman ()A second new variable was past versus future verb tense as against allprevious experimental studies on Hebrew passive production whichinvolved past tense the default form of Hebrew passives

We had four hypotheses following the literature reviewed above ()Correct performance on the passive task was expected to increase from theyoungest age group to adulthood across the period of later languagedevelopment () Sentences with active verbs in higher register were expectedto incur lower correct scores than those with verbs in neutral register () WeexpectedNifrsquoal to lead correct passive performance that is to have the highestscores starting from the lowest age group We had no expectations regardingthe ordering within the two strict passives Pursquoal and Hufrsquoal as previousstudies had shown conflicting results () As an irrealis temporal form futuretense is rarely used in referring to events Accordingly we expected higherperformance on the canonical past tense verbs

METHOD

This study was a structured elicitation task testing the production of Hebrewpassive voice constructions in writing PARTICIPANTS were typicallydeveloping monolingual native Hebrew-speaking children adolescents andadults with no diagnosed language or learning disorders They were all of



middle SES (socio-economic status) as determined by the Strauss CultivationMeasure (Strauss Ensminger amp Fothergill ) All participants livedin the same region in the south of Israel Participants were in seven ageschooling level groups eight- to nine-year-olds (M= ) in third grade(henceforth designated eight-year-olds) nine- to ten-year-olds (M= ) infourth grade ten- to eleven-year-olds (M= ) in fifth grade eleven-to twelve-year-olds(M= ) in sixth grade thirteen- to fourteen-year-olds (M= ) in eighth grade sixteen- to seventeen-year-olds (M=) in eleventh grade and adult university students aged ndash

Materials

Participants were asked to change written active-voice sentences intocorresponding passive-voice sentences Materials consisted of tasksentences divided according to three variables verb register binyan andverb tense (Table )

Register Half of the sentences () were in neutral register and the otherhalf in high register Verbs with neutral register corresponded to Ravid andBermanrsquos () level ndash colloquial everyday usage eg shalax lsquosentrsquo ndash whilehigh-register verbs corresponded to level ndash the standard written usage ofeducated monolingual speakers eg lexically specific abstract hexerimlsquoconfiscatedrsquo Task verbs were checked against school texts to ensure theirsuitability for eight-year-olds the youngest age group Sententialarguments (agent and patient nouns) were adjusted to the verbrsquos registerlevel based on Ravidrsquos () Noun Scale For example the agent andpatient for neutral-register send were Ron and letters respectively (Ron sent theletters) while the agent and patient for high-register confiscate were theCustoms and merchandise respectively (the Customs confiscated the merchandise)The register division was designed to determine whether lexical properties ofthe active verb and its context were helpful or detrimental in learning Allmaterials were piloted in a corresponding population to ensure that childrenunderstood the meanings of the sentences and their components

Binyan The three transitive verb patterns were given equal representationof sixteen sentences each Qal (targeting passive Nifrsquoal as in shadadnishdad

TABLE Structure of the Passive Task (N = items)

Qal Nifrsquoal N= Pirsquoel Pursquoal N= Hifrsquoil Hufrsquoal N=

Neutral register N = Neutral register N = Neutral register N =

Past tense Future tense Past tense Future tense Past tense Future tenseN= N= N= N= N= N=

High register N= High register N= High register N=


RAVID AND VERED


lsquorobbed was robbedrsquo) Pirsquoel (targeting passive Pursquoal as in pizerpuzarlsquoscattered was scatteredrsquo) and Hifrsquoil (Hufrsquoal as in yaklityuklat lsquowillrecord will be recordedrsquo) This division enabled us to test the role ofbinyan morphology in learning to passivize

Tense Half of the task sentences had past tense verbs and half future tenseverbs The most compelling reason for this was that present tense passiveforms (eg meluxlax lsquodirtyrsquo in Pursquoal) comprise an entirely differentlinguistic and psycholinguistic domain in Hebrew acquisition (Berman Persquoer ) Past and future tense verbs agree with theirgrammatical subjects in person gender and number while present tenseverbs carry gender and number (but no person) agreement marking likeadjectives The division into past and future verb tense enabled us to testthe role of the specific temporal patterns of each binyan in learning thepassive forms and to ask whether past tense passives were easier toacquire All task verbs were in third person eg ha-talmid tersquoer et ha-rivlsquothe-student described ACC the-fightrsquo so as to be narrative in tone

Procedure

Testing took place in writing in the class forum that is during the schoolday in the classroom Participation was voluntary students who did notwish to participate were exempted and given other tasks in a differentlocation on the school premises Each student received one of tworandomized versions of the target forty-eight sentences on two pagesStudents sitting next to each other received two different versions to ensureindividual work Each target sentence in active voice was followed by aresponse line starting with the new grammatical subject ie theactive-sentence object NP which served as a prompt to the passive voiceconstruction that participants were asked to complete For exampleha-moxer yishkol et ha-sxora lsquothe-vendor will-weigh the-merchandisersquo wasfollowed by a line starting with ha-sxora lsquothe merchandisersquo Followingpiloting the instruction at the top of the first page was as followsldquoFollowing below are sentences Re-write them without changing themeaning of the sentence nor its tenserdquo Given the age range of theparticipants only one example was given Yossi axal et ha-tapuacuteax lsquoYossi ateACC the-applersquo followed by ha-tapuacuteax nersquoexal al yedey Yossi lsquoThe-apple waseaten by Yossirsquo (passive verb and by-phrase underlined in the given example)

Scoring

As the patient grammatical subject was used as the prompt syntactic errorswere virtually absent Scoring thus focused on the morphological changefrom active to passive verb Responses were categorized into six levelsfrom to Level designated a correct passive form in the required



binyan with further categorization indicating inflectional errors in tenseLevel indicated incorrect passive binyam responses based on the correctroot eg erroneous Hufrsquoal hushdad for correct Nifrsquoal nishdad lsquowasrobbedrsquo For each of the three binyan patterns in the test there were twopossible erroneous alternatives ndash the other two binyan patterns ThusNifrsquoal or Pursquoal passives were possible Level errors for a target Hufrsquoalpassive form Level involved two passive-related errors () present-tenseresultative adjectives eg mexudad lsquosharpenedrsquo for target future tenseyexudad lsquowill be sharpenedrsquo and () medial-passive Hitparsquoel egerroneous hitkanes for target Nifrsquoal niknas lsquowas finedrsquo Level errorsconsisted of morphophonologically non-felicitous forms with some passiveindication such as u-vowels (eg tusuman for correct tesuman lsquowill bemarkedrsquo) Level errors consisted of non-passive responses usuallyfocusing on the inflectionall markers of the cue active form For exampleplural hegifu lsquothey shutteredrsquo for hegifa lsquoshe shutteredrsquo where the targetpassive form should have been hugfu lsquowere shutteredrsquo Level errorsconsisted of non-passive semantic and syntactic alternatives (eg kibel knaslsquoreceived a finersquo for niknas lsquowas finedrsquo) irrelevant answers and empty slots

RESULTS

As register was the only non-morphological variable we first report correctresponses (Level responses only converted into percentages) in twoseparate tables by register Table presents correct responses for the

TABLE Percentages and standard deviations (in parentheses) of correctpassive responses to the neutral-register verbs by ageschooling group binyanand verb tense

AgeSchoolinggroup

Qal Nifrsquoal Hifrsquoil Hufrsquoal Pirsquoel Pursquoal

Pasttense

Futuretense

Pasttense

Futuretense

Pasttense

Futuretense

rd grade middot middot middot middot middot middot

ndash-year-olds (middot) (middot) (middot) (middot) (middot) (middot)th grade middot middot middot middot middot middotndash-year-olds (middot) (middot) (middot) (middot) (middot) (middot)th grade middot middot middot middot middot middot


ndash-year-olds (middot) (middot) (middot) (middot) (middot) (middot)th grade middot middot middotndash-year-olds (middot) (middot) (middot) (middot) (middot) (middot)Adults middot

ndash-year-olds (middot) (middot) (middot) (middot) (middot) (middot)

RAVID AND VERED


neutral-register verbs and Table presents correct responses for thehigh-register verbs A four-way ANOVA of correct responses in () ageschooling groups (eight- nine- ten- eleven- thirteen- sixteen-year-oldsand adults) times () binyan verb patterns (Qal Nifrsquoal Hifrsquoil Hufrsquoal Pirsquoel Pursquoal) times () verb tense (past tense future tense) times () register (neutralhigh) was performed on the data in Tables and

This analysis yielded an effect for register (F() = middot plt middotη= middot) ndash verbs with neutral register scored higher (M= middot) thanverbs with high register (M= middot) A four-way interaction of ageschooling group binyan pattern verb tense and register was found(F() = middot p lt middot η= middot) with subsequent two three-way andthree two-way interactions involving register These results confirmed ourinitial hypothesis regarding the critical importance of linguistic register inpassivization We thus turned to two separate analyses in neutral- andhigh-register activepassive verbs

Correct passivization of verbs in neutral register

A three-way ANOVA of correct (Level ) responses in () ageschoolinggroups times () binyan verb patterns times () verb tense was performed on thedata in Table Correct responses increased (F() = middot p lt middotη = middot) from middot in eight-year-olds to middot in thirteen-year-oldsFurther Bonferroni pairwise comparisons showed three clusters of

TABLE Percentages and standard deviations (in parentheses) of correctpassive responses to the high-register verbs by ageschooling group binyan andverb tense

AgeSchoolinggroup


Pasttense

Futuretense

Pasttense

Futuretense

Pasttense

Futuretense




ndash-year-olds (middot) (middot) (middot) (middot) (middot) (middot)th grade middot middot middot middot middot middotndash-year-olds (middot) (middot) (middot) (middot) (middot) (middot)Adults middot middot middotndash-year-olds (middot) (middot) (middot) (middot) (middot) (middot)



ageschooling groups ndash eight- and nine-year-olds ten-and eleven-year-oldsand the three oldest groups The effect for binyan verb pattern (F() =middot p lt middot η= middot) and further Bonferroni comparisons showed thatHufrsquoal responses (M= middot) were higher than Pursquoal (M= middot) andNifrsquoal (M= middot) responses which did not differ from each otherVerbs in past tense scored higher (M= middot) than verbs in future tense(M= middot) (F() = middot plt middot η= middot)Three two-way interactions of age group with binyan (F() = middot

p = middot η = middot) and with verb tense (F = () = middot p = middot η= middot)and of binyan and verb tense (F = () = middot p lt middot η= middot) werefound as well as a three-way interaction of age group binyan and verbtense (F() = middot p lt middot η= middot) Figure shows that futuretense Nifrsquoal had the lowest scores in eight-year-olds and the shallowestgrowth curve while future tense Hufrsquoal verbs and past tense Nifrsquoal verbshad the highest scores Verbs in past tense Hufrsquoal and both Pursquoal tensepatterns were in the middle

Correct passivization of verbs in high register

A three-way ANOVA of correct responses in () ageschooling groups times ()binyan verb patterns times () verb tenses was performed on the data inTable Correct responses increased with age (F() = middot plt middotη= middot) from middot in eight-year-olds to middot in sixteen-year-olds

Fig Interaction of age group binyan verb pattern and verb tense in correct passiveresponses of neutral register

RAVID AND VERED


Further Bonferroni comparisons showed the same three clusters of ageschooling groups as in neutral register ndash eight- and nine-year-olds ten-and eleven-year-olds and the three oldest groups The effect for binyanverb pattern (F() = middot p lt middot η= middot) and Bonferronicomparisons showed that all three binyan patterns significantly differedfrom each other Hufrsquoal responses (M= middot) were highest followed byPursquoal (M= middot ) and then Nifrsquoal (M= middot) responses Verb tensewas not significant

Two two-way interactions of age group with binyan (F() = middotp = middot η= middot) and of binyan and verb tense (F = () = middotp lt middot η= middot) were found as well as a three-way interaction of agegroup binyan and verb tense (F() = middot p lt middot η = middot)Figure shows two distinct groups of patterns in acquisition the threehigher-scoring patterns were past tense Pursquoal and the two Hufrsquoal tensepatterns the three lower-scoring patterns were past tense and future tenseNifrsquoal and future-tense Pursquoal

Error analysis

Non-morphological errors were very few and did not permit statisticalanalysis They occurred only in eight- and nine-year-olds and mostlyconsisted of providing syntactic alternatives in the form of subordinatedclauses eg ha-shulxan zaz biglal she-ha-mora heziza oto lsquothe desk moved

Fig Interaction of age group binyan verb pattern and verb tense in correct passiveresponses of high register



because the teacher moved itrsquo instead of changing heziza lsquoshe moved TRrsquo tohuzaz lsquowas movedrsquo as prompted The overwhelming majority of errors weremorphological Accordingly we focused on Level errors ndash that isresponses that used erroneous passive binyan morphology Tables and

present the percentages of erroneous passive binyan responses out of thetotal number of responses Three-way ANOVAs of erroneous passivebinyan responses in () ageschooling groups times () binyan verb patterns times() verb tenses were performed on the data in Tables and

Neutral register Erroneous passive responses declined with age (F() =middot plt middot η= middot) with a cut-off between the younger ageschoolinggroups (M= middot in eight-year-olds M= middot in nine-year-olds) andthe rest of the groups (under in ten- and eleven-year-olds dwindlingto in the older groups virtually absent in the adults) Regardingbinyan (F() = middot p lt middot η= middot) Nifrsquoal had the most passivebinyan errors (M= middot) followed by Pursquoal (M= middot) and Hufrsquoalresponses (M= middot) which did not differ Verb tense was alsosignificant (F() = middot plt middot η= middot) with more passive binyanerrors (M= middot) in future than in past tense (M= middot) Threetwo-way interactions of age group with binyan (F() = middot p = middotη= middot) age group and verb tense (F() = middot plt middot η= middot)and binyan and verb tense (F() = middot p lt middot η= middot) werefound as well as a three-way interaction of age group binyan and verb

TABLE Percentages and standard deviations (in parentheses) of erroneouspassive binyan responses out of the total of responses in neutral register by ageschooling group binyan and verb tense

AgeSchoolinggroup


Pasttense

Futuretense

Pasttense

Futuretense

Pasttense

Futuretense




ndash-year-olds (middot) (middot) (middot) (middot) (middot) (middot)th grade middot middot middot middot middot middotndash-year-olds (middot) (middot) (middot) (middot) (middot) (middot)Adults middot middot middot middot middot middotndash-year-olds (middot) (middot) (middot) (middot) (middot) (middot)

RAVID AND VERED


tense (F() = middot plt middot η = middot) Figure shows that future tenseNifrsquoal had the most passive binyan errors declining from close to ineight-year-olds to a virtual absence in adults Of the erroneous passivebinyan responses to target future tense Nifrsquoal () were in Hufrsquoaland the rest in PursquoalHigh register Erroneous passive responses declined with age (F() =

middot p lt middot η= middot) but the cut-off was between eight- andeleven-year-olds with over such errors (M= middot in eight-year-olds)and the two oldest groups with under errors with thethirteen-year-olds lying in between (M= middot) differing only fromadults Regarding binyan (F() = middot p lt middot η= middot) Nifrsquoal hadthe most passive binyan errors (M= middot) followed by Pursquoal (M=middot) and Hufrsquoal responses (M = middot) which did not differ Verb tensewas not significant Three two-way interactions of age group with binyan(F() = middot p lt middot η= middot) age group and verb tense (F() =middot p= middot η = middot) and binyan and verb tense (F() = middotp lt middot η= middot) were found as well as a three-way interaction of agegroup binyan and verb tense (F() = middot p lt middot η = middot)Figure shows that future tense and past tense Nifrsquoal had the mostpassive binyan errors declining from about in eight-year-olds tovirtual absence in adults Of the erroneous passive binyan responses ontarget future tense Nifrsquoal () were in Hufrsquoal and the rest in Pursquoal

TABLE Percentages and standard deviations (in parentheses) of erroneouspassive binyan responses out of the total of responses in high register by ageschooling group binyan and verb tense

AgeSchoolinggroup


Pasttense

Futuretense

Pasttense

Futuretense

Pasttense

Futuretense







of the Level responses on target past tense Nifrsquoal () were inPursquoal and the rest in Hufrsquoal

DISCUSSION

This study elicited passive verbs in writing in a sentential context fromHebrew-speaking school-aged children adolescents and adults focusingon the variables of neutral vs high register activepassive binyan patternpairs (QalNifrsquoal HifrsquoilHufrsquoal PirsquoelPursquoal) and past vs future tense

Fig Interaction of age group binyan verb pattern and verb tense in erroneous passivebinyan responses in neutral register

Fig Interaction of age group binyan verb pattern and verb tense in erroneous passivebinyan responses in high register

RAVID AND VERED


Passive as a very late acquisition development

Our first two hypotheses regarding the protracted process of acquisition andthe impact of register were confirmed The current study found that learningpassive verb morphology in Hebrew is a very late acquisition Active verbs inneutral register with human subjects and concrete objects (eg tsavalsquopaintedrsquo tadpis lsquowill type uprsquo pizer lsquoscatteredrsquo) reached correctresponses only by age thirteen to fourteen almost a decade later thanattested in other languages with late passives In our view this delay ismostly due to the drawn-out process of learning to reconcile the role ofHebrew passives in the expression of event-oriented dynamic and specific(rather than concept-oriented generic and static) content with using thedetached distanced mode preferred by adult narrators (Berman Ravid amp Chen-Djemal )

Passive learning in Hebrew demonstrates the interface of discourseabilities with grammatical acquisition While Hebrew-speaking eight-year-olds have gained full and automatic command of verb morphology(Berman ) their ability to productively match verb forms todiscourse functions is just emerging (Berman ) School-goingten-year-olds use distinct grammatical devices for narrative and expositoryagent demotion (Ravid ) being familiar with the conventionalchild-oriented medial passive and adjectival passive devices prevalent instory-books (Ravid amp Levie ) Having encountered school texts theyare also familiar with the use of subjectless often verbless constructionsfor the expression of routines ideas and factual information (Berman Ravid Ravid amp Zilberbuch ) These two classes ofgrammatical devices are however kept apart for distinct genre expression

Learning verbal passives in Hebrew requires the flexibility brought onby adolescent socio-cognitive development (Blakemore amp Choudhury Nippold ) coupled with enhanced exposure to diverse communicativeevents and written texts of different genres (Berman amp Ravid Christie amp Derewianka ) By late adolescence these developmentsenable Hebrew users to cross the strict genrendashstructure association andadopt the mature preference for the specialized usage of event-orientedverbal passives in taking a distanced generic outlook on specific events

Passive as a very late acquisition register

As predicted again learning verbal passives was found to be mediatedby linguistic register In the current study abstract lexically specifichigher-register active verbs (kansa lsquofinedrsquo tanxe lsquowill lead the ceremonyrsquonimek lsquomotivatedrsquo) yielded lower passive scores than neutral verbs andreached only by sixteen to seventeen years of age Outside ourexperimental design high-register verbal passives are the norm rather than



the exception in Hebrew usage and moreover they usually co-occur withother abstract mental and irrealis high-register lexical items typical ofadult expression (Ravid amp Chen-Djemal ) This is exemplified in ()a September Hebrew posting on Facebook

() huvhar li she-haseacuteret yukran ha-eacuterevwas-clarified to-me that-the-film will-be-screened this-eveninglsquoIt was made clear to me that the film would be screened this eveningrsquo

As register is a lexical feature in Hebrew (Ravid amp Berman ) verbalpassives are lexically rather than syntactically challenging They are part ofthe lexical explosion that doubles adolescent vocabulary ushering in alsoderived nominals (Ravid Ravid amp Avidor ) and denominaladjectives (Ravid amp Zilberbuch ) categories with restricted semanticndashpragmatic distributions and simpler alternatives used by younger speakers

Learning the Hebrew passive morphology and register

We now turn to the morphological aspects of passive acquisition Recall theremaining hypotheses regarding the primacy of Nifrsquoal as the bridge topassive learning the indeterminate order of acquisition and the challengeposed by future tense passive forms The findings revealed a clear thoughunexpected order of acquisition among the three passive verb patternsinteracting differently with temporal patterns and introducing syntacticand semantic considerations to the developmental picture

Counter to previous findings across development in both registers andparticularly in the high register past and future tense Hufrsquoal verbs faredthe best in correct responses Pursquoal verbs followed however past waseasier than future tense especially in the high register Finally Nifrsquoalverbs behaved differently in the two registers future tense was extremelychallenging in both registers but past tense Nifrsquoal verbs were as easy asHufrsquoal in neutral register and as difficult as future tense Nifrsquoal in highregister Moreover target Nifrsquoal verbs entailed four times as many passivebinyan errors than both the other passive patterns with Hufrsquoal occupyingall of Nifrsquoal error space in neutral register and all future tense Nifrsquoalerror space in high register Pursquoal attracted almost all past tense Nifrsquoalerrors Explaining these results requires the examination of the propertiesof each activepassive binyan pair as mediated by the lexical properties ofregister

Hufrsquoal passives With the highest correct passive scores across all agegroups and as the major attractor of passive binyan errors Hebrewlearners clearly preferred Hufrsquoal as the default verbal passive Thishighlights HifrsquoilHufrsquoal as the most syntactically semantically andphonologically regular and transparent activepassive pair in Hebrew Of

RAVID AND VERED


the three active binyan patterns Hifrsquoil is the most transitive with mostHifrsquoil verbs taking the accusative direct object (Dattner ) henceconstituting the prototypical syntactic source for passive formationPhonologically too Hufrsquoal not only has the hallmark passive u-vowel butis also completely uniform the only passive binyan with the same vocalicpattern across all temporal stems Semantically Hufrsquoal passives are the mostinflection-like in nature solely dedicated to the regular virtually automaticpassive modulation on the hifrsquoil verb meaning This is supported by arecent analysis of Hebrew resultative adjectives in childrenrsquos peer talkwhich showed that the category of derivational lexically specific Hufrsquoaladjectives was the smallest in the corpus (Persquoer )Pursquoal passives Pursquoal passives had an interim status in the study In neutral

register they lay in between the high-scoring Hufrsquoal past tense Nifrsquoal andthe low-scoring future tense Nifrsquoal In high register past tense Pursquoal verbsclustered with Hufrsquoal while future tense Pursquoal verbs shared a steeptrajectory with Nifrsquoal verbs These results reflect the specific properties ofthe PirsquoelPursquoal pair

Syntactically Pirsquoel is less transitive than Hifrsquoil with more oblique (ratherthan accusative) objects (Dattner ) and therefore fewer opportunitiesfor passive formation Phonologically although Pursquoal shares the u-voweland the strict passive morphology with Hufrsquoal it is less uniform absentthe typical past tense h-prefix Morphologically Pirsquoel and Pursquoal arestrongly linked to Hitparsquoel (Ravid et al ) offering a medial-passiveescape hatch for the younger age groups Thus Pursquoal had nine times asmany Level medial-passive errors () than Nifrsquoal and Hufrsquoal (under) in the younger groups ndash all of them in Hitparsquoel for example hitpazrulsquoscatteredrsquo for correct puzru lsquowere scatteredrsquo and yistakem lsquowill add up torsquofor correct yesukam lsquowill be summarizedrsquo

Taken together these properties indicate the double function of PirsquoelPursquoal as the less regular and transparent strict passive pair side by sidewith Pursquoalrsquos central role as a derivational device generating twice as manyresultative adjectives than the other two passive patterns many of themlacking transitive Pirsquoel sources (Persquoer ) Pursquoal was also revealed as thecanonical Hebrew perfective passive accounting for the fact that past tensePursquoal passives attracted the overwhelming majority Nifrsquoal errors and bythe inferior results of future tense Pursquoal passives

Nifrsquoal passives Counter our predictions the QalNifrsquoal pair constitutedthe major challenge to passive learning as the least preferred option for theexpression of passive voice in participants up to adulthood Future tenseNifrsquoal verbs in neutral register and both tenses in high register had the

Even the small set of intransitive inchoative Hifrsquoil verbs such as hivri lsquoget wellrsquo or hexmirlsquogrow severersquo always have a causative interpretation



lowest scores in eight-year-olds anddistinctly steep trajectoriesThemaverickmorphophonological properties ofNifrsquoal so different from the strict passiveswould have led us to posit a suitable hypothesis but all previous studies onHebrew passive in acquisition had pointed at Nifrsquoal as the bridge to passivelearning Thanks to including the tense and register variables the currentstudy can now offer a solution to this apparent contradiction

Clearly passive voice is only one of the functions ofNifrsquoal It is the canonicalintransitive binyan in the QalNifrsquoalHifrsquoilHufrsquoal subsystem holding asimilar role to Hitparsquoel in the subsystem it shares with Pirsquoel and Pursquoal (Ravidet al ) The relationship between Qal and Nifrsquoal is often transitivemiddle as in shavarnishbar lsquobreakTRbreakINTrsquo often with a simultaneousinterpretation of both middle or passive as in taraknitrak lsquoslamTRslamINTsim be slammedrsquo Many Nifrsquoal middlepassives do not have Qalcounterparts eg nirtav lsquoget wetrsquo while others hold idiosyncraticrelationships with Qal eg both avadnersquoevad meaning lsquoget lostrsquo in high andneutral register respectively Qal itself the most structurally andsemantically irregular and variegated binyan is the least transitive of thethree passive-source binyanim designated as transitivity-neutral in Berman() This is also supported in that half of the adjectival passives inCaCuC the passive resultative adjective pattern corresponding to the QalandNifrsquoal paradigm do not match a transitive verb inQal (Persquoer )It was only the neutral register past tense Nifrsquoal verbs all conveying the

canonical perfective outlook that had early high scores in the currentstudy as in the previous studies and served as the early bridge to maturepassives Although most of them had the double middlepassiveinterpretation (eg neheras lsquobeget destroyedrsquo nishkal lsquobe weighed weighoneselfrsquo) there was no separate middle option attracting errors as Hitparsquoeldoes for Pursquoal In other words in a different binyan they would haveshown Level errors But when required to use the non-canonical futurefor the perfective outlook these atypical Nifrsquoal passives were rejected infavor of the default Hufrsquoal The high-register abstract Nifrsquoal passives sovery different from the early Nifrsquoal telics such as nikra lsquotorersquo were the lastto be acquired They were abstract clearly passive verbs (eg tibalem lsquowillbe restrainedrsquo niknesa lsquowas finedrsquo) but nonetheless systematically rejectedby younger participants in favor of Hufrsquoal (future tense) and Pursquoal (pasttense) errors For example Hufrsquoal tuvdak for correct Nifrsquoal tibadek lsquowillbe examinedrsquo or Pursquoal guzal for correct Nifrsquoal nigzal lsquowas usurpedrsquo It wasonly in late adolescence that participants were willing to forgo the strictpassive options and accept Nifrsquoal as a passive option of Qal

CONCLUSIONS

The very late acquisition of Hebrew verbal passives is an example of howspecific language typology interacts with the agent-demoting outlook of

RAVID AND VERED


passive voice Morphological passives ride piggyback on middlepassive pasttense Nifrsquoal forms to begin with but once the canonical Pursquoal and Hufrsquoalpassives are learned in grade school the Nifrsquoal bridge is burned Hebrewpassive morphology enters the developmental picture only when learnershave identified how verbal passives behave in the specifically perfective yetdetached communicative contexts where generic subjectless or adjectivalpassives will not do the job ndash most typically when conversant with theliterate written mode

Taking an overarching view of the domain the study of passive voiceacquisition has undergone several phases side by side with changes in theconstrual of the passive voice in linguistics Earlier studies viewed passivevoice (mainly through the prism of English) as a syntactic phenomenonwith the delay to about age five explained by the need to move NPs in thesentence (Borer amp Wexler Pierce ) or due to a universalmaturational delay preventing children from connecting the patient role tothe subject position (Horgan Maratsos et al ) Laterinput-driven analyses explained that English-speaking children under threedo not generalize the construction due to scarce exposure in earlychildhood (Brooks amp Tomasello ) This view was supported bytypological studies pointing to the early acquisition of both simple andcomplex forms of the passive in languages where passive constructions areprevalent (Allen amp Crago Demuth Gil Pye amp QuixtanPoz ) The current study was chaperoned by the shifting linguisticspotlight towards the verbal morphology of passives on the one hand andthe extension of developmental psycholinguistic investigation to the realmsof adolescence and written language on the other

REFERENCES

Abraham W amp Leisiouml E (eds) () Passivization and typology form and functionAmsterdam John Benjamins

Abraham W amp Leiss E () The impersonal passive voice suspended under aspectualconditions In W Abraham amp E Leisiouml (eds) Passivization and typology form andfunction ndash Amsterdam John Benjamins

Alcock K J Rimba K amp Newton C R J C () Early production of the passive in twoEastern Bantu languages First Language ndash

Allen S E M amp Crago M B () Early passive acquisition in Inuktitut Journal of ChildLanguage ndash

Aschermann E Gulzow I amp Wendt D () Differences in the comprehension of passivevoice in German- and English-speaking children Swiss Journal of Psychology ndash

Baldie B J () The acquisition of the passive voice Journal of Child Language ndashBerman R A () The case of an (S)VO language subjectless constructions in ModernHebrew Language ndash

Berman R A () Acquisition of Hebrew Hillsdale NJ Lawrence ErlbaumBerman R A () Marking of verb transitivity by Hebrew-speaking children Journal ofChild Language ndash



Berman R A () Formal lexical and semantic factors in acquisition of Hebrewresultative participles Berkeley Linguistic Society ndash

Berman R A () Between emergence and mastery the long developmental route oflanguage acquisition In R A Berman (ed) Language development across childhood andadolescence (pp ndash) Amsterdam John Benjamins

Berman R A () Introduction developing discourse stance in different text types andlanguages Journal of Pragmatics ndash

Berman R A () The psycholinguistics of developing text construction Journal of ChildLanguage ndash

Berman R A amp Nir-Sagiv B () Linguistic indicators of inter-genre differentiation inlater language development Journal of Child Language ndash

Berman R A amp Ravid D () Becoming a literate language user oral and written textconstruction across adolescence In D R Olson amp N Torrance (eds) Cambridgehandbook of literacy ndash Cambridge Cambridge University Press

Berman R A amp Slobin D I () Relating events in narrative a crosslinguisticdevelopmental study Hillsdale NJ Erlbaum

Biber D () Dimensions of register variation a crosslinguistic comparison CambridgeCambridge University Press

Blakemore S-J amp Choudhury S () Development of the adolescent brain implicationsfor executive function and social cognition Journal of Child Psychology and Psychiatry ndash

Borer H amp Wexler K () The maturation of syntax In T Roeper and E Williams(eds) Parameter setting (pp ndash) Dordrecht Reidel

Brooks P J amp Tomasello M () Young children learn to produce passives with nonceverbs Developmental Psychology ndash

Budwig N () The linguistic marking of nonprototypical agency an exploration intochildrenrsquos use of passives Linguistics ndash

Christie F amp Derewianka B () School discourse London ContinuumCrain S Thornton R amp Murasugi K () Capturing the evasive passive LanguageAcquisition ndash

Dattner E () Enabling and allowing in Hebrew a Usage-Based Construction Grammaraccount In B Nolan G Rawoens amp E Diedrichsen (eds) Causation permission andtransfer argument realisation in GET TAKE PUT GIVE and LET verbs ndashAmsterdam Benjamins

DemuthK () Subject topic and theSesothopassiveJournal ofChildLanguagendashDemuth K Moloi F amp Machobane M () -Year-oldsrsquo comprehension productionand generalization of Sesotho passives Cognition ndash

Dromi E amp Berman R A () Language-general and language-specific in developingsyntax Journal of Child Language ndash

EnsmingerMEampFothergillKE ()AdecadeofmeasuringSESwhat it tellsusandwhereto go fromhere InMHBornsteinampRHBradley (eds)Handbookof parenting socioeconomicstatus parenting and child development Vol ndash Mahwah NJ Erlbaum

Ferguson C A () Dialect register and genre working assumptions aboutconventionalization In D Biber amp E Finegan (eds) Sociolinguistic perspectives onregister ndash New York Oxford University Press

Ferreira F () Choice of passive voice is affected by verb type and animacy Journal ofMemory and Language ndash

Foley W A () A typology of information packaging in the clause In T Shopen (ed)Language typology and syntactic description Vol II Clause structure ndash CambridgeCambridge University Press

Fox D amp Grodzinsky Y () Childrenrsquos passive a view from the by-phrase LinguisticInquiry ndash

Gaacutemez P B Shimpi P M Waterfall H R amp Huttenlocher J () Priming aperspective in Spanish monolingual children the use of syntactic alternatives Journal ofChild Language ndash

RAVID AND VERED


Gil D () The acquisition of voice morphology in Jakarta Indonesian In N Gagarina ampI Guumllzow (eds) The acquisition of verbs and their grammar the effect of particular languagesndash Dordrecht Springer

Givoacuten T () Typology and functional domains Studies in Language ndashGordon P amp Chafetz J () Verb-based versus class-based accounts of actionality effectsin childrenrsquos comprehension of passives Cognition ndash

Halliday M A K () Spoken and written language Oxford Oxford University PressHaspelmath M () The grammaticalization of passive morphology Studies in Language ndash

Horgan D () The development of the full passive Journal of Child Language ndashKeenan E L amp Dryer M S () Passive in the worldrsquos languages In T Shopen (ed)Language typology and syntactic description Vol II Clause structure ndash CambridgeCambridge University Press

Lee K-O amp Lee Y () An event-structural account of passive acquisition in KoreanLanguage and Speech ndash

Maratsos M Fox D E Becker J A amp Chalkley M A () Semantic restrictions onchildrenrsquos passives Cognition ndash

Marchman V A Bates E Burkardt A amp Good A B () Functional constraints of theacquisition of the passive toward a model of the competence to perform First Language ndash

Messenger K () Syntactic priming and childrenrsquos production and representation of thepassive Language Acquisition ndash

Messenger K Branigan H P amp Mclean J F () Is childrenrsquos acquisition of the passivea staged process Evidence from six- and nine-year-oldsrsquo production of passives Journal ofChild Language ndash

Mitkovska L amp Bužarovska E () An alternative analysis of the Englishget-past participle constructions Is get all that passive Journal of English Linguistics ndash

Monachesi P () The verbal complex in Romance a case study in grammatical interfacesOxford Oxford University Press

Myhill J () Towards a functional typology of agent defocusing Linguistics ndashNippold M A () Later language development school-age children adolescents and youngadults rd ed Austin TX Pro-Ed

Persquoer M () Resultative passives in the Hebrew lexicon and in Hebrew-speaking childrenrsquospeer talk Unpublished MA thesis Tel Aviv University [in Hebrew]

Pierce A () The acquisition of passives in Spanish and the question of A-chainmaturation Language Acquisition ndash

Pinker S Lebeaux D S amp Frost L R () Productivity and constraints in theacquisition of the Passive Cognition ndash

Prat-Sala M Shillcock R amp Sorace A () Animacy effects on the production ofobject dislocated descriptions by Catalan-speaking children Journal of Child Language ndash

Pye C amp Quixtan Poz P () Precocious passives (and antipassives) in Quiche MayanPapers and Reports on Child Language Development ndash

Rathert M () Simple preterit and composite perfect tense the role of the adjectivalpassive In W Abraham amp L Leisiouml (eds) Passivization and typology form and functionndash Amsterdam John Benjamins

Ravid D () Language change in child and adult Hebrew a psycholinguistic perspectiveNew York Oxford University Press

Ravid D () Later lexical development in Hebrew derivational morphology revisited InR A Berman (ed) Language development across childhood and adolescence psycholinguisticand crosslinguistic perspectives ndash Amsterdam Benjamins

Ravid D () Emergence of linguistic complexity in written expository texts evidencefrom later language acquisition In D Ravid amp H Bat-Zeev Shyldkrot (eds) Perspectiveson language and language development ndash Dordrecht Kluwer



Ravid D () Semantic development in textual contexts during the school years NounScale analyses Journal of Child Language ndash

Ravid D Ashkenazi O Levie R Ben Zadok G Grunwald T Bratslavsky R amp GillisS () Foundations of the root category analyses of linguistic input to Hebrew-speakingchildren In R Berman (ed) Acquisition and development of Hebrew from infancy toadolescence (pp ndash) (TILAR (Trends in Language Acquisition Research) )Amsterdam John Benjamins

Ravid D amp Avidor A () Acquisition of derived nominals in Hebrew developmentaland linguistic principles Journal of Child Language ndash

Ravid D amp Berman R () Developing linguistic register in different text typesPragmatics amp Cognition ndash

Ravid D amp Chen-Djemal Y () Spoken and written narration in Hebrew a case studyWritten Language and Literacy ndash

Ravid D amp Epel Mashraki Y () Prosodic reading reading comprehension andlanguage skills in Hebrew-speaking

th graders Journal of Research in Reading ndashRavid D amp Geiger V () Promoting morphological awareness in Hebrew-speakinggradeschoolers an intervention study using linguistic humor First language ndash

Ravid D amp Levie R () Adjectives in the development of text production lexicalmorphological and syntactic analyses First Language ndash

Ravid D amp Saban R () Syntactic and meta-syntactic skills in the school years adevelopmental study in Hebrew In I Kupferberg amp A Stavans (eds) Languageeducation in Israel papers in honor of Elite Olshtain ndash Jerusalem Magnes Press

Ravid D amp Zilberbuch S () Morpho-syntactic constructs in the development ofspoken and written Hebrew text production Journal of Child Language ndash

Schwarzwald O () The special status of Nifrsquoal in Hebrew In S Armon-LotemG Danon and S Rothstein (eds) Current issues in generative Hebrew linguistics ndashAmsterdam John Benjamins

Shibatani M () On the conceptual framework for voice phenomena Linguistics ndash

Strauss S () Ministry of Education Chief Scientist report on the Cultivation MeasureJerusalem Ministry of Education [in Hebrew]

Svenonius P () Slavic prefixes inside and outside VP In P Svenonius (ed) NordlydTromsoslash Working Papers on Language and Linguistics () Special issue on Slavic prefixesndash Tromsoslash University of Tromsoslash

Timberlake A () Aspect tense mood In In T Shopen (ed) Language typology andsyntactic description Vol II Grammatical categories and the lexicon ndash CambridgeCambridge University Press

Tomasello M Brooks P J amp Stern E () Learning to produce passive utterancesthrough discourse First Language ndash

RAVID AND VERED


been invoked in explaining why young children struggle in comprehendingand producing passive forms (Brooks amp Tomasello Messenger Pinker Lebeaux amp Frost )

The current study aimed to examine the acquisition of passive forms inHebrew in a new light First focus here is on the MORPHOLOGY of thepassive verb As general theoretical motivation for this verb-focusedinvestigation we follow Haspelmathrsquos ( p ) comment that ldquotheverbal morphology associated with a passive construction is an essentialpart of the construction whose properties are worthy of study in their ownright Indeed the passive can be regarded as first and foremost a verbalmorphological category whose meaning implies certain changes in theclause structurerdquo From a Hebrew-specific typological perspectivestudying passive verb morphology in acquisition is key as Hebrew passiveforms are part and parcel of the verb-pattern binyan system a set of sevenconjugations which provide the vocalic form of the verb and itssyntactico-semantic category Second previous research on passiveacquisition ndash in Hebrew as in other languages ndash examined spoken passiveconstructions in young children mostly in preschoolers In contrast thecurrent study elicited passive forms in writing from school-going childrenand adolescents compared with adults during the period of LaterLanguage Development (Berman Nippold ) This frameworkenabled us to contextualize the acquisition of passive verbs as part of a setof Hebrew morpho-lexical classes whose learning is delayed toadolescence ndash a time of when socio-cognitive linguistic and literacyabilities connect together in the complex expression of communicativegoals (Berman amp Ravid )

The main goals of the current study were as follows (i) to provide anaccount of the development of Hebrew passive verb morphology across theschool years based on an experimental task and (ii) to determine the roleof derivational and inflectional binyan morphology in this development

Passive constructions

Passive voice seems to be a universal phenomenon nonetheless it has beendescribed as an elusive category deserving of sound conceptual foundations(Shibatani ) Several conclusions come to mind when reviewing thelinguistic literature on passive constructions and their functional rolesacross different languages which can be useful in laying the backgroundand generating expectations for the current developmental study ofHebrew passives

First passives are noted among the most prominent and widespreadperspective-changing constructions in the worldrsquos languages (Givoacuten) The general view of passive constructions as backgrounding the

RAVID AND VERED











RAVID AND VERED


















RAVID AND VERED



















RAVID AND VERED















RAVID AND VERED






METHOD





Materials










RAVID AND VERED




Procedure


Scoring





RESULTS



AgeSchoolinggroup


Pasttense

Futuretense

Pasttense

Futuretense

Pasttense

Futuretense






RAVID AND VERED







AgeSchoolinggroup


Pasttense

Futuretense

Pasttense

Futuretense

Pasttense

Futuretense












RAVID AND VERED




Error analysis









AgeSchoolinggroup


Pasttense

Futuretense

Pasttense

Futuretense

Pasttense

Futuretense





RAVID AND VERED





AgeSchoolinggroup


Pasttense

Futuretense

Pasttense

Futuretense

Pasttense

Futuretense








DISCUSSION




RAVID AND VERED

















RAVID AND VERED













CONCLUSIONS


RAVID AND VERED




REFERENCES
































RAVID AND VERED









































RAVID AND VERED











RAVID AND VERED


















RAVID AND VERED



















RAVID AND VERED















RAVID AND VERED






METHOD





Materials










RAVID AND VERED




Procedure


Scoring





RESULTS



AgeSchoolinggroup


Pasttense

Futuretense

Pasttense

Futuretense

Pasttense

Futuretense






RAVID AND VERED







AgeSchoolinggroup


Pasttense

Futuretense

Pasttense

Futuretense

Pasttense

Futuretense












RAVID AND VERED




Error analysis









AgeSchoolinggroup


Pasttense

Futuretense

Pasttense

Futuretense

Pasttense

Futuretense





RAVID AND VERED





AgeSchoolinggroup


Pasttense

Futuretense

Pasttense

Futuretense

Pasttense

Futuretense








DISCUSSION




RAVID AND VERED

















RAVID AND VERED













CONCLUSIONS


RAVID AND VERED




REFERENCES
































RAVID AND VERED









































RAVID AND VERED


















RAVID AND VERED



















RAVID AND VERED















RAVID AND VERED






METHOD





Materials










RAVID AND VERED




Procedure


Scoring





RESULTS



AgeSchoolinggroup


Pasttense

Futuretense

Pasttense

Futuretense

Pasttense

Futuretense






RAVID AND VERED







AgeSchoolinggroup


Pasttense

Futuretense

Pasttense

Futuretense

Pasttense

Futuretense












RAVID AND VERED




Error analysis









AgeSchoolinggroup


Pasttense

Futuretense

Pasttense

Futuretense

Pasttense

Futuretense





RAVID AND VERED





AgeSchoolinggroup


Pasttense

Futuretense

Pasttense

Futuretense

Pasttense

Futuretense








DISCUSSION




RAVID AND VERED

















RAVID AND VERED













CONCLUSIONS


RAVID AND VERED




REFERENCES
































RAVID AND VERED









































RAVID AND VERED



















RAVID AND VERED















RAVID AND VERED






METHOD





Materials










RAVID AND VERED




Procedure


Scoring





RESULTS



AgeSchoolinggroup


Pasttense

Futuretense

Pasttense

Futuretense

Pasttense

Futuretense






RAVID AND VERED







AgeSchoolinggroup


Pasttense

Futuretense

Pasttense

Futuretense

Pasttense

Futuretense












RAVID AND VERED




Error analysis









AgeSchoolinggroup


Pasttense

Futuretense

Pasttense

Futuretense

Pasttense

Futuretense





RAVID AND VERED





AgeSchoolinggroup


Pasttense

Futuretense

Pasttense

Futuretense

Pasttense

Futuretense








DISCUSSION




RAVID AND VERED

















RAVID AND VERED













CONCLUSIONS


RAVID AND VERED




REFERENCES
































RAVID AND VERED









































RAVID AND VERED















RAVID AND VERED






METHOD





Materials










RAVID AND VERED




Procedure


Scoring





RESULTS



AgeSchoolinggroup


Pasttense

Futuretense

Pasttense

Futuretense

Pasttense

Futuretense






RAVID AND VERED







AgeSchoolinggroup


Pasttense

Futuretense

Pasttense

Futuretense

Pasttense

Futuretense












RAVID AND VERED




Error analysis









AgeSchoolinggroup


Pasttense

Futuretense

Pasttense

Futuretense

Pasttense

Futuretense





RAVID AND VERED





AgeSchoolinggroup


Pasttense

Futuretense

Pasttense

Futuretense

Pasttense

Futuretense








DISCUSSION




RAVID AND VERED

















RAVID AND VERED













CONCLUSIONS


RAVID AND VERED




REFERENCES
































RAVID AND VERED









































RAVID AND VERED






METHOD





Materials










RAVID AND VERED




Procedure


Scoring





RESULTS



AgeSchoolinggroup


Pasttense

Futuretense

Pasttense

Futuretense

Pasttense

Futuretense






RAVID AND VERED







AgeSchoolinggroup


Pasttense

Futuretense

Pasttense

Futuretense

Pasttense

Futuretense












RAVID AND VERED




Error analysis









AgeSchoolinggroup


Pasttense

Futuretense

Pasttense

Futuretense

Pasttense

Futuretense





RAVID AND VERED





AgeSchoolinggroup


Pasttense

Futuretense

Pasttense

Futuretense

Pasttense

Futuretense








DISCUSSION




RAVID AND VERED

















RAVID AND VERED













CONCLUSIONS


RAVID AND VERED




REFERENCES
































RAVID AND VERED









































RAVID AND VERED



Materials










RAVID AND VERED




Procedure


Scoring





RESULTS



AgeSchoolinggroup


Pasttense

Futuretense

Pasttense

Futuretense

Pasttense

Futuretense






RAVID AND VERED







AgeSchoolinggroup


Pasttense

Futuretense

Pasttense

Futuretense

Pasttense

Futuretense












RAVID AND VERED




Error analysis









AgeSchoolinggroup


Pasttense

Futuretense

Pasttense

Futuretense

Pasttense

Futuretense





RAVID AND VERED





AgeSchoolinggroup


Pasttense

Futuretense

Pasttense

Futuretense

Pasttense

Futuretense








DISCUSSION




RAVID AND VERED

















RAVID AND VERED













CONCLUSIONS


RAVID AND VERED




REFERENCES
































RAVID AND VERED









































RAVID AND VERED




Procedure


Scoring





RESULTS



AgeSchoolinggroup


Pasttense

Futuretense

Pasttense

Futuretense

Pasttense

Futuretense






RAVID AND VERED







AgeSchoolinggroup


Pasttense

Futuretense

Pasttense

Futuretense

Pasttense

Futuretense












RAVID AND VERED




Error analysis









AgeSchoolinggroup


Pasttense

Futuretense

Pasttense

Futuretense

Pasttense

Futuretense





RAVID AND VERED





AgeSchoolinggroup


Pasttense

Futuretense

Pasttense

Futuretense

Pasttense

Futuretense








DISCUSSION




RAVID AND VERED

















RAVID AND VERED













CONCLUSIONS


RAVID AND VERED




REFERENCES
































RAVID AND VERED









































RAVID AND VERED



RESULTS



AgeSchoolinggroup


Pasttense

Futuretense

Pasttense

Futuretense

Pasttense

Futuretense






RAVID AND VERED







AgeSchoolinggroup


Pasttense

Futuretense

Pasttense

Futuretense

Pasttense

Futuretense












RAVID AND VERED




Error analysis









AgeSchoolinggroup


Pasttense

Futuretense

Pasttense

Futuretense

Pasttense

Futuretense





RAVID AND VERED





AgeSchoolinggroup


Pasttense

Futuretense

Pasttense

Futuretense

Pasttense

Futuretense








DISCUSSION




RAVID AND VERED

















RAVID AND VERED













CONCLUSIONS


RAVID AND VERED




REFERENCES
































RAVID AND VERED









































RAVID AND VERED







AgeSchoolinggroup


Pasttense

Futuretense

Pasttense

Futuretense

Pasttense

Futuretense












RAVID AND VERED




Error analysis









AgeSchoolinggroup


Pasttense

Futuretense

Pasttense

Futuretense

Pasttense

Futuretense





RAVID AND VERED





AgeSchoolinggroup


Pasttense

Futuretense

Pasttense

Futuretense

Pasttense

Futuretense








DISCUSSION




RAVID AND VERED

















RAVID AND VERED













CONCLUSIONS


RAVID AND VERED




REFERENCES
































RAVID AND VERED









































RAVID AND VERED







RAVID AND VERED




Error analysis









AgeSchoolinggroup


Pasttense

Futuretense

Pasttense

Futuretense

Pasttense

Futuretense





RAVID AND VERED





AgeSchoolinggroup


Pasttense

Futuretense

Pasttense

Futuretense

Pasttense

Futuretense








DISCUSSION




RAVID AND VERED

















RAVID AND VERED













CONCLUSIONS


RAVID AND VERED




REFERENCES
































RAVID AND VERED









































RAVID AND VERED




Error analysis









AgeSchoolinggroup


Pasttense

Futuretense

Pasttense

Futuretense

Pasttense

Futuretense





RAVID AND VERED





AgeSchoolinggroup


Pasttense

Futuretense

Pasttense

Futuretense

Pasttense

Futuretense








DISCUSSION




RAVID AND VERED

















RAVID AND VERED













CONCLUSIONS


RAVID AND VERED




REFERENCES
































RAVID AND VERED









































RAVID AND VERED






AgeSchoolinggroup


Pasttense

Futuretense

Pasttense

Futuretense

Pasttense

Futuretense





RAVID AND VERED





AgeSchoolinggroup


Pasttense

Futuretense

Pasttense

Futuretense

Pasttense

Futuretense








DISCUSSION




RAVID AND VERED

















RAVID AND VERED













CONCLUSIONS


RAVID AND VERED




REFERENCES
































RAVID AND VERED









































RAVID AND VERED





AgeSchoolinggroup


Pasttense

Futuretense

Pasttense

Futuretense

Pasttense

Futuretense








DISCUSSION




RAVID AND VERED

















RAVID AND VERED













CONCLUSIONS


RAVID AND VERED




REFERENCES
































RAVID AND VERED









































RAVID AND VERED



DISCUSSION




RAVID AND VERED

















RAVID AND VERED













CONCLUSIONS


RAVID AND VERED




REFERENCES
































RAVID AND VERED









































RAVID AND VERED

















RAVID AND VERED













CONCLUSIONS


RAVID AND VERED




REFERENCES
































RAVID AND VERED









































RAVID AND VERED













CONCLUSIONS


RAVID AND VERED




REFERENCES
































RAVID AND VERED









































RAVID AND VERED




REFERENCES
































RAVID AND VERED









































RAVID AND VERED
























RAVID AND VERED









































RAVID AND VERED









































RAVID AND VERED


hebrew verbal passives in later language development · pdf filehebrew verbal passives in...

Documents