How templatic is Arabic input to children?:The role of child-directed-speech in the acquisition of Semitic morpho-phonology

Semitic languages such as Hebrew and Arabic are known for having a non-concatenative morphology: words are typically built of a combination of a consonantal root, typically tri-consonantal (e.g., k-t-b ‘related to writing’ in Modern Standard Arabic (MSA)), with a prosodic template. Research on Hebrew language development suggests early sensitivity to frequently occurring templates (Segal et al., 2015). For the Arabic dialects, little is known about whether implicit sensitivity to non-concatenative morphology develops at a young age through exposure to speech, and how templatic the spoken language is in comparison to MSA. We focus on Lebanese Arabic. We hypothesised that prolonged contact with French and English may have ‘diluted’ the salience of roots and patterns in the input. We used 3 different corpora of adult-directed-Speech (ADS), child-directed-speech (CDS) and child speech. We analysed the root and pattern structures in the 50 most frequent Lebanese Arabic word types in each corpus. We found fewer words with templatic patterns than expected among the most frequent words in ADS (35/50), even fewer in CDS (23/50) and still fewer in the children’s target words (15/50). In addition, only a minority contain 3 root consonants in their surface forms: 22 in ADS, 15 in CDS and only 7 in words targeted by the children. We conclude that Semitic structure is less evident in either input to children or words targeted by children aged 1-3 than has been assumed. We discuss implications for the development of sensitivity to templatic structure among Lebanese-acquiring children.