Abstract
Bavin and Stoll’s book offers a comprehensive overview of current research on the early acquisition of ergative morphology. Two preliminary texts show the complexity of ergativity from the linguistic and developmental perspectives, followed by nine acquisition studies which analyze child (and adult) naturalistic data of 15 languages from widespread locations around the world: America (7), Asia (3), Australia (1), Europe (1) and Papua New Guinea (3).
The editors introduce the reader to the content of the book with an instructive account of the term ergative. In contrast to (nominative-)accusative languages such as English (among many others), which treat the sole argument of intransitive verbs (S) and the agent of transitive verbs (A) similarly, ergative languages mark the agentive subject of transitive verbs differently from the sole argument of intransitive verbs (S), which is marked similarly to (O)bject- or (P)atient- arguments of transitive verbs. As a general finding of the studies included in the volume, Bavin and Stoll highlight the (often) weak effect of factors such as frequency of ellipsis, salience, consistency and social context in the acquisition of ergativity.
In the second introductory chapter, Comrie focuses on the different degrees of consistency in ergativity observed across categories, structures or even language levels cross-linguistically. Terms such as (semantic vs. syntactic, symmetric vs. asymmetric) alignment, and hierarchy may help to identify some but not all subtypes of ergative languages. Moreover, some facts such as the loss of anti-passive structures found in a few ergative languages are identified as a symptom of the decay of syntactic ergativity.
Austin investigates Basque, an isolated European language. She analyzes case and argument–verb agreement morphology produced by (Spanish-)Basque bilingual and monolingual children aged between 2;0 and 3;4 and a third group of bilingual adults living in Northern Spain. Her data confirm the results of previous child studies concluding that: (a) the ergative case is acquired later than the absolutive and dative cases; (b) the target-deviant ergative case marking contrasts with the virtually target-like ergative verb agreement during the earliest stages; (c) the phonological context cannot explain the target-deviant production; and (d) the amount of exposure affects the rates of target-like productions.
Allen analyzes ergativity in Inuktitut, an Eskimo-Aleut language spoken in Canada. Data obtained across three different children’s and adults’ production samples of Northern Quebecquian Inuktitut are compared with data obtained in a fourth study on child West Greenlandic. The low presence of transitive structures as well as of ergative-marked arguments in spontaneous speech might suggest that ergativity is not yet acquired at age 3. However, older children’s and adults’ preference for intransitive structures such as passives and antipassives in third person agent contexts observed in the narratives in Inuktitut, together with the increasingly productive use of ergative (pro)nominal structures and transitive constructions by children speaking the West Greenlandic variety, indicate that 2- to 5-year-old children’s developmental patterns are closely related to the adult use of the specific ergative features in each variety under study. Finally, these data confirm the existence of dialectal diversity among Eskimo-Aleut languages, some of which are shifting to a nominative-accusative system.
Bavin’s chapter deals with Warlpiri, a Pama Nyungan language spoken in central Australia. The first sample of 2- to 4-year-old children’s conversations reveals a productive use of ergative subjects by age 3, whereas some target-deviant use of allomorphs is attested at age 4. The second sample of elicited narratives by older children (4;8–12;4) and adults shows a generalized absence of ergative arguments in the narratives of children under 6, which seems to develop inversely to the increase in complexity of the narratives of 10-year-old or older participants. In contrast to previous studies, the target-like use found in all groups in contexts of focus, emphasis or disambiguation reveals young children’s sensitivity towards the morpho-lexical and pragmatic rules governing the ergative marking.
Rumsey et al.’s collaborative chapter compares Kaluli, Ku Waru and Duna, three non-Austronesian languages spoken in Papua New Guinea. Data analyzed in the samples of children aged between 1;8 and 3;11 in spontaneous conversations with adults reveal an early presence of ergative marking after an initial stage of case and argument omission in the three languages. In general, children correctly apply the marker to nominals and personal pronouns without overextending such marking to objects. The authors conclude that some kinds of interactional routines associated with a high incidence of ergative marking may provide structural and situational templates for the earlier use of ergative markers.
Stoll and Bickel’s study on Chintang, a Sino-Tibetan language spoken in Nepal, reveals a quantitatively and functionally similar use of ergatives between four children (2;0–4;4) and their (child and adult) interlocutors from the first recording sessions. The increase of lexical variability, together with the more autonomous production of ergative marking suggests that children have knowledge of ergativity at around age 3.
Narasimhan investigates Hindi, an Indo-European language spoken in Northern India. The conversations of children aged 1;7–3;9 with their caretakers reveal an extremely low frequency of ergative arguments (below 2%) in both groups’ productions among which pronominal forms are the most frequent nominal category. The adult-like scarcity of ergative arguments and the absence of overuse of ergative marking confirm children’s sensitivity to ergative morphology despite its low presence in the input. Hindi data converge with children’s resort to multiple probabilistic cues in the acquisition of language-specific categories.
Mahalingappa analyzes Kurmanji Kurdish spoken in Turkey. Both the naturalistic and elicited data of children (1;6–3;6) and their caretakers confirm the existence of an ergative split in children’s and adults’ productive systems. Moreover, the presence of DIR-DIR and OBL-OBL patterns together with traditional patterns in the younger caretakers’ and children’s speech suggests an ongoing linguistic change towards a nominative-accusative pattern. However, the variability attested among adults makes it difficult to determine to what extent patterns observed in children’s production reveal a developmental production or rather reflect the inconsistency of a changing system.
Brown and colleagues develop a comparative study of the acquisition of four Mayan languages spoken in Mexico and Guatemala which code ergativity only on verb morphology: K’iché, Tzeltal, Tzotzil and Yucatec. Data from children aged 2;0–3;6 reveal a gradual development of case agreement (absolutive > ergative) without overgeneralization errors. The high accuracy shown even in split systems like Yucatec indicates that transitivity plays a major role and that young children are aware of constraints on the use of ergative and absolutive morphology from the earliest phases. Results indicate that phonological features (prevocalic > preconsonantal, syllabic > non-syllabic morphemes) as well as position (suffixed > prefixed) play a role in the acquisition of morphological markers.
Finally, Pye and colleagues compare the acquisition of three Mayan languages spoken in Guatemala and Mexico, which ‘extend’ the ergative marker to cross-reference subjects of monovalent predicates (Q’anjob’al and Yucatec) or to cross-reference the direct object (Mam). Data of three children (1;9–3;1) and one parental adult collected for each language indicate that children use such extended marking regardless of its frequency in the particular language. The absence of overgeneralization of the extended marker to other verbs suggests children’s awareness of the language-specific rules that govern the use of ergative marking.
This volume is interesting for numerous reasons. First, the novelty of the languages collected, some of them (minority) languages whose production in general, or ergativity in particular, has received little, if any, attention in the acquisition research, makes this a valuable compilation for investigators of cross-linguistic phenomena. Second, the homogeneity of the data obtained from (pseudo-)spontaneous productions of 2- to 4-year-olds and adults offers useful empirical material for comparative purposes. Third, the virtually adult-like distinction among A, S and O around age 3, together with the unpredicted absence of overgeneralization of ergative marking to objects observed in all the languages analyzed, suggests an early and uniform development which contrasts with the variability in morpho-syntactic markers attested cross-linguistically. These results provide important fuel for debate about universalities of content and/or processes in language acquisition. Fourth, the inclusion of both children’s and adults’ production data offers a great opportunity for the reader to discern the developmental ‘distance’ of children’s productions from their target and/or input language. Fifth, the updated information regarding the linguistic and historic context of the languages studied, most of them in long-standing contact with non-ergative languages, provides new and useful material for researchers interested in areas that go beyond language acquisition, such as language typology, historical linguistics and sociolinguistics.