Date of Award

May 2019

Degree Type


Degree Name

Doctor of Philosophy



First Advisor

Sandra Pucci

Committee Members

Hamid Ouali, Nicholas Fleisher, Fred Eckman


Heritage languages, Heritage speakers, Heritage speakers in France, Language acquisition, Morphology, The acquisition of nominal and verbal morphology


There are two major perspectives regarding heritage speakers’ (henceforth HS) ultimate attainment. Some researchers on HS in the U.S. conclude that HS have incomplete grammars (Benmamoun, Montrul & Polinsky, 2013). It is argued that heritage languages (henceforth HL) do not fully develop (Montrul, 2016), and they are not completely acquired because of shifting to a dominant language (Benmamoun et al., 2013). Other researchers argue that HS’ grammars are complete, but simply different as monolingual and HS experience different linguistic realities (Pascual y Cabo & Rothman, 2012). While there is abundant research on Arabic as a HL in the U.S., research on HS in Europe has been rather limited (Montrul, 2016). This dissertation focuses on Moroccan Arabic (henceforth MA) as a HL in France and aims at contributing to the understanding of the linguistic outcomes of the acquisition of Arabic as a HL in an immigrant context.

The current study investigates the acquisition of nominal morphology and verbal-derivational patterns by Moroccan HS in France. Nominal morphology was represented by plural and diminutive formation, and verbal derivations were represented by four patterns. Nominal morphology gives insights into both concatenative and non-concatenative morphological processes. Verbal derivational processes are characterized by the use of non-concatenative morphological processes, and semantic notions such as causativeness and reciprocity are lexicalized within the MA verb-pattern system. The studied patterns are the basic (P1), causative (P2), medio-passive (P5), and reciprocal (P6). 15 Moroccan-French participants took part in this study. The data were gathered through three production experiments. Experiment 1 investigated the acquisition of 3 sound morphemes and 14 broken plurals. In experiment 2, participants were tested in diminutive formation, exemplified by the six types of diminutives, and experiment 3 examined the acquisition of verbal derivations.

Experiment 1 revealed that participants’ plural system is mainly characterized by concatenative processes as just two sound plural morphemes were acquired by a significant number of participants. All the participants acquired the morpheme -at and 80% of the participants acquired the sound morpheme -in. Participants depend on overgeneralization and simplification of their plural system. The sound morphemes characterize the HL and were overgeneralized in broken plural targeted data. Additionally, the sound plural [-at] seems to be the underspecified default morpheme in the HL.

The findings of experiment 2 show that the mean percentage of source-like use of diminutive forms is 38%. The results revealed that just two patterns were acquired by a significant number of participants: CCiCa and CCiCjCjəC. Diminutive forms that do not require complex processes are acquired by a significant number of participants and the percentage of source-like use is high as well. In this study, a brief analysis of diminutive derivational processes was given, and participants’ patterns of acquisition correspond to the suggested continuum of complexity. Irregular stems present difficulties to HS as complex processes are applied. Non-source like data is rule-governed as 69% of the non-source-like data shows the use of either initial consonant cluster or insertion of the glide /-j/, and these are the main processes characterizing diminutive processes. Additionally, the requirement of having two syllables was met. Participants tend to regularize diminutive formation and show a preference to the following processes: initial constant cluster and insertion of the palatal glide. Generally, HS’s variety is mainly characterized by two patterns.

Experiment 3 reveals that the basic pattern (P1) was acquired by all the participants and 40% of the participants acquired the causative (P2). The medio-passive pattern (P5) and the reciprocal patterns (P6) were not acquired. ANOVA showed that there were statistically significant differences among the use of the four patterns. The main finding of this study is that semantic distinction realized by pattern alternation is neutralized in the HL. Specifically, the basic pattern (P1) and periphrastic constructions were used predominantly in P5 and P6 targeted data. It is likely that the basic pattern is used as a default morphological device because it unmarked.

The three experiments demonstrate that participants omit irregularities and non-source like forms are rule governed. Less complex and less marked morphological structures characterize the HL. Specifically, morphological aspects thought to be acquired earlier in language development are the ones characterizing the HL. Additionally, the findings of the experiments propose implicational hierarchies for the acquisition of the studied morphological structures. Adopting overgeneralization in nominal morphology, and neutralization in verb patterns showed that HS speak a variety that is reanalyzed. Accordingly, HS in France have a distinct variety that was shaped by their linguistic experience. Their variety is different, reanalyzed and does not comprise all the patterns attested in the source language.

Included in

Linguistics Commons