Complexity and density of Hebrew verbs in preschool peer talk

Abstract

Hebrew verbs were analyzed in the peer talk produced by 36 Hebrew-speaking children in two age/schooling groups (4;0–5;0 and 5;0–6;0 years), and from two socio-economic backgrounds (SES), mid-high and low. Each of the four age/SES groups consisted of nine children in three triads, where each triad was recorded for 30 minutes while playing. The interface of lexical and morphological growth was demonstrated in the developing organization of verbs in terms of roots, binyan conjugations and derivational families. SES was found the major source of variation in all measures, indicating a smaller and less specific verb lexicon in the low SES groups. Network analyses, a novel methodological approach, revealed the internal structure of the verb category in each age/SES cell, pointing to a scarce and less complex verb lexicon of the low SES groups. These measures also accounted for the growth potential of the network, increasing from the younger low SES group at one pole and peaking in the older mid-high SES at the other pole. These quantitative and qualitative differences in the morphological make-up of the verb lexicon and its usage patterns in preschool peer talk have implications for the impact of SES on verb learning in Hebrew.

Patterns of Adaptation in Child-Directed and Child Speech in the Emergence of Hebrew Verbs

“…Finally, another factor that is common to both children is the effect of the CS network density on degree level, such that degree levels decrease with higher network density. This is a manifestation of the growth potential interpretation of the density measure, proposed by Levie et al (2019), since lower density levels indicate a higher potential for the network to grow, realized here as higher degree levels.…”

“…These results demonstrate the manifestation of adaptation: both parties in the dyad adapt their network structure to that of their interlocutor, taking into account their current and the previous states (with some individual differences). Going from the statistics to morphology, we interpret density as growth potential (Levie et al, 2019). Thus, we can conclude that the potential to expand the network, forming new links between existing roots and inflected patterns, is a function of more than just the age of the child, with the current and previous morphological network structures of the other party in the developmental tango also having an effect.…”

“…This is a (usually) tri-literal consonantal string that constitutes the formal and semantic core of many Hebrew words, and most especially, of all Hebrew verbs (Laks, 2013;Kastner, 2019;Ravid, 2019). Many studies point to the Semitic root as the most accessible Hebrew morpheme in spoken and written language development and usage (Ravid and Bar-On, 2005;Gillis and Ravid, 2006;Schiff et al, 2012;Ben-Zvi and Levie, 2016;Deutsch and Kuperman, 2019), including contexts of language disability or environmental deprivation (Ravid and Schiff, 2006;Schiff and Ravid, 2007;Levie et al, 2017Levie et al, , 2019. Young Hebrewspeaking children demonstrate an early ability to extract roots from familiar words and use them in novel forms (Berman, 1985(Berman, , 2000(Berman, , 2012Ravid, 2003).…”

“…With age, children are exposed to larger root-pattern networks that highlight the shared vocalic structure of verbs, making it possible for binyan conjugations and their syntactic-semantico values to be learned (Levie et al, 2020). The increase in number, size and complexity of networks of root-related derivational verb families is a clear indicator of a growing verb lexicon (Ravid et al, 2016;Levie et al, 2019).…”

“…As such, somewhat counter-intuitively, a sparse network is taken here to indicate a higher level of potential productivity: In a sparse network, there are more root-and pattern-nodes which are not linked to each other, compared to a dense network in which most of the nodes are already linked. Hence, the potential to link two nodes that have not been linked before, thus creating new verb wordforms, is higher in a sparse network, compared to a dense network (Levie et al, 2019). That is, a sparse network means that the pool from which one can choose how to put experience into words, specifically verbs (by linking a root and a pattern) is not exhausted, and new verb wordforms can be created: new links between roots and inflected patterns, which refer to more fine-grained aspects of experience.…”