The Acquisition of Negation in Modern Standard Arabic by Bilingual Moroccan Children

Introduction. The present study reports results from grammaticality judgment and picture description tasks with 24 bilingual child learners of Modern Standard Arabic (MSA) (and Moroccan Arabic), aged between 5;1 and 9;1. The study investigates the learners’ comprehension and production of the anaphoric la: ‘no’, verbal la, lan, lam, ma: ‘not’, and non-verbal lajsa ‘is not’. It also explores whether the complex MSA negation delays the age of acquisition of the anaphoric, verbal, and non-verbal structures. Results. Results suggested that the learners’ comprehension of negation was more robust compared to production, indicating processing difficulties associated with production of the complex MSA negation. The analysis of error patterns showed that children overutilized the anaphoric la: ‘no’ to avoid more complex structures. Interestingly, the development of negation was non-linear, displaying a U-shaped learning curve. Conclusion. The current study examined negation in the comprehension and production of Moroccan child learners of MSA. First, in line with prediction, the complex structure of MSA negation was found to delay the development of the targeted negation structures. Second, the participants’ comprehension of MSA negation was more robust compared to production. This could be explained by difficulties in the implementation of grammatical knowledge. Third, reporting individual results, the study suggested that the child learners of MSA showed a non-linear pattern of negation acquisition, displaying a U-shaped learning curve. Finally, further research is needed to investigate MSA properties because it would provide valuable test cases for various important questions in linguistic theory and language acquisition, including our understanding of the nature, development, representation, and processing of MSA.


Introduction
The examination of children's negation acquisition has sparked a wider interest in acquisitional studies.This resulted from its wider usage in the world languages, in addition to the grammatical complexity that it manifests in these languages.The present study investigates the comprehension and production of negation in Modern Standard Arabic (MSA).The latter has complex negation structures.The study further contributes new child data from a hitherto understudied studied language.The available studies focus more on the Arabic varieties.MSA can offer valuable insights to L1 acquisition, given the prestigious status it enjoys in the Arab world.The Arab speakers regard it as 'The Arabic Language' (Albirini, 2015).

Background Literature
The Acquisition of Negation in English and Other Languages Children of different ages are reported to go through developmental stages in their development of negation in different languages like English (Batet & Grau 1995;Bloom, 1970;Cameron-Faulkner, Lieven & Theakston, 2007;Choi, 1988;Deprez & Pierce, 1993;Drozd, 2002;Klima & Bellugi, 1966), German (Wode, 1977), Danish, Norwegian, and Swedish (Plunkett & Strömqvist, 1990).Most of these L1 acquisition studies confirm a fixed acquisitional order.Children usually start from the holophrastic negation (one word negation) stage, producing equivalent of no, to incomplete multi-word negation stage, and end up attaining successful multi-word negation.Although they disagree on the number and specification of these stages, all the studies propose the existence of different developmental stages in children's acquisition of negation.
Interestingly, the acquisitional stages appear to be related to each language's formal negation structures, in addition to children's cognitive development (Austin et al., 2014;Choi, 1988;Feiman et al., 2017;Reuter et al., 2018).Austin et al. (2014), Feiman et al. (2017), and Reuter et al. (2018) suggest that child acquisition of negation is affected by their cognitive development, and their comprehension ability is affected by syntactic features.To this end, children are argued to fully reach adult-like production of negation by the age of 4;6 (Pea, 1980;Dimroth, 2010;Reuter et al., 2018;Thornton, 2020).
The development of negation has been studied from a pragmatic perspective, differentiating between anaphoric and non-anaphoric negation.The early presence of negation in child language is attested in a lot of studies, with the equivalent of the anaphoric no being developed in the holophrastic stage (Brink, 2017;Cameron-Faulkner et al., 2007;Plunkett & Strömqvist, 1990;Wode, 1977).
Early studies (e.g.Bellugi, 1967;Choi, 1988;Vorster, 1982) base their acquisitional arguments on spontaneous and/or elicited production of negation, drawing on longitudinal studies on child negation development.Attention to children's production and comprehension asymmetries in child development of negation has culminated in two views.The first one holds the assumption that comprehension precedes production (Bates et al., 1988;White et al., 2023).The second view argues that production abilities precede comprehension in German (Wojtecka et al., 2011) and Turkish (Ünal & Papafragou, 2016) children.
To this end, it is important to consider both comprehension and production data to draw conclusions on child development of negation.

The Acquisition of Negation in the Arabic Varieties
A lot of studies investigate the development of negation in the Arabic varieties, but research on MSA is still rudimentary.As a frame of reference, this study includes findings from Egyptian Arabic (Omar, 1973), Jordanian Arabic (Smadi, 1979), Moroccan Arabic (Aftat, 1982), Qatari Arabic (Al-Buainain, 2003), Kuwaiti Arabic (Al-Jenaie, 2008), and Najdi Arabic (Binturki, 2015).Their findings are reported for context.
Several acquisitional patterns are attested in these studies.The anaphoric negation in all the Arabic varieties is developed before other negation structures.The non-discontinuous negation emerges before the discontinuous negation.The difficulty of the discontinuous structure could be explained by syntactic complexity (e.g., movement and merger) or morphological complexity (two-part form).Though the discontinuous morpheme appears in the negative existential structure ma-fi:-ꭍ 'there isn't/aren't,' it is not problematic for Arabic-speaking children.This may either be because it is acquired as a single unit and/ or it is frequent.The early occurrence of negative imperatives in child language, compared to other forms of verbal negation, may be due to the frequency of imperatives in parents' input.Thus, the development of negation in the Arabic varieties appears to be affected by syntactic, morphological, and frequency factors.

MSA Structures The Anaphoric, Verbal, and Non-verbal Negation in MSA
There are three types of negation in MSA: anaphoric, verbal, and non-verbal.The anaphoric la: 'no' can be used to negate yes/no questions as in (1) below.The negators /la:/, /lam/, /lan/, and /ma:/ 'not' negate verbal sentences in MSA (ma: can be used in non-verbal sentences as well).They occupy a preverbal position in the sentences which they negate.However, these negators differ from one another in terms of their contexts of use as well as the changes they bring into the sentences they negate.The major differences between them are summarized below.
The verbal negator la: 'not' is used to negate imperative verbs (2a), the imperfective/present verbs (2b), and pseudo verbs (2c).The verb following la: is marked for the indicative mood (2b).The negator lam 'not' negates the perfective/past tense form of verbs, and it assigns the jussive mood to the verb following it (3).The negator lan 'not' negates the imperfective/future tense form of verbs, assigning the subjunctive mood to the verb following it (4).The negator ma: 'not' negates the perfective/past (5a), habitual imperfective/present tense sentences (5b), and non-verbal sentences with nominal, adjectival, and prepositional predicates as well as tropicalized noun phrases and prepositional phrases (5c).
The non-verbal negator lajsa 'is not' has multiple agreement paradigms (table 5).It agrees with the subject in person and gender in the VSO order (6a), and in person, number, and gender in the SVO order (6b).
(6) a. lajsa l-walad-u fi l-bajt-i is.not the-boy-NOM in the-house-GEN 'The boy is not in the house.' b. ʔal-bint-u lajs-at fi l-madrasat-i the-girl-NOM is.not-F in the-school-GEN 'The girl is not in the school.'It is argued beforehand that the complexity of negation structures poses difficulties on the comprehension and production of these structures (White et al., 2022;Ünal & Papafragou, 2016;Wojtecka et al., 2011).MSA displays complex negative structures that interact with verbal morphology.This results in complex realizations of negation structures that requires advanced cognitive maturity among children.Assuming this analysis, verbal negators in MSA have different contexts of use.We would predict the complexity of the realization of negation in MSA to pose difficulty on the age of acquisition of the anaphoric, verbal, and non-verbal negation by Moroccan child learners of MSA.

Participants
Participants included twenty-four Moroccan child learners of MSA as an L1.Their age ranged between 5;1 and 9;1.They belonged to four different school levels: (1) preschool, (2) first grade, (3) second grade, and (4) third grade.They studied MSA at the same Moroccan public school in Casa-Settat Region.The participants were native speakers of Moroccan Arabic.The latter was the variety spoken at home, and the variety of day-to-day interaction.All the participants were exposed to MSA (L1) from birth through television (especially cartoons) and children's literature.They also went to religious schools before they attended formal schooling.The profile of the participants as well as their mean length of utterance (MLU) was shown in table 6.

Tasks
The development of negation in MSA by Moroccan children was investigated using a grammaticality judgment task (GJT) and a picture description task (PDT).The participants were given a contrastive judgment task.Ungrammatical sentences that involve violations of aspects of the targeted negative structures were included to test the learners' sensitivity to such violations (7).The rationale behind using the GJT was to examine the participants' comprehension of MSA negation.
The GJT consisted of 16 sentence pairs, testing participants' knowledge of MSA negation (the GJT tests are included in Appendix B).The participants completed a picture description task (PDT).The PDT elicited the less frequent negation structures.Further, it helped us control the meaning to be associated with the targeted utterances.The participants completed the tasks in the same classroom where they met regularly for their MSA classes, and, thus, the setting and the linguistic environment were familiar to them.The participants were required to produce utterances using various MSA negation structures, and they were presented with 15 picture slides in random order via PowerPoint presentation (see Appendix B for task items).A sample of the picture slides used for the elicitation of negation in MSA (item # 14) appeared in Figure 1.

Figure 1
Test item № 14 from the PDT Note.The stimulus shown here is a picture slide used to elicit non-verbal negator lajsa 'is.not' in the PDT.This question ĥal ĥa:ꭍa: walad?'Is this a boy?' was used in the picture slide to elicit the targeted structure: 1) ĥa:ꭍa: lajsa walad 'This is not a boy'.

Data Collection and Coding
The GJT and the PDT were administered over three different days in a Moroccan public school in Casa-Settat Region, in May 2023, in a familiar setting, using familiar methods (pen, paper, laptop, iPad, and voice recorder).This was to avoid possible task environment effect on the participants' performance (Backman, 1990).The participants were tested individually, and the instructions were presented in MSA.Both tasks contained vocabulary within the participants' level, and they were told that they could ask about the meaning of ambiguous words to ensure that the tasks were not affected by learners' vocabulary knowledge.Before the experiment, the participants were given two practice test trials.The purpose of conducting the two tasks was explained to the participants to ensure they provided valid and reliable answers (i.e., they were told that the experiment was part of a study).
For preschoolers and first graders, both the GJT and PDT data were recorded on a digital audio recorder.For the second and third graders, the GJT data was written on paper, and the PDT data was recorded on a digital audio recorder.Then, the recordings were transcribed verbatim on Word document.The numbers of target and non-target responses were counted.The participants' productions went through a selection process before the data coding.Relevant negation structures were selected for further data analysis while the non-targeted structures (e.g., single nouns or other types of phrases) were eliminated and not included for further data analysis.The tasks, therefore, had control of the structures to be produced by the participants.The patterns of errors across types were coded.To this end, the present study reported individual data rather than group results to provide more robust grounds to examine arguments about language acquisition (Eckman, 2007;Christiansen et al., 2018).
Further, data was coded using implicational scaling, and a 70% criterion was used.* Each sentence was coded, depending on the targeted structure, for correct or incorrect use of a negation particle.For both tasks, the status of negation particle use was presented using '+' or '-'.For each participant '+' marked the use of a negation particle 70% of the time or more.'-' indicated the incorrect use of a negation particle as indicated by producing it less than 70% of the time.

Results
Data was analyzed quantitatively for the comprehension of MSA negation on the GJT (see appendix A for a display of quantitative raw data calculations on the GJT).The results of the judgments of negation were displayed in table 7 below for all the participants.
The top row of the implicational scaling tables on both tasks showed negative particles.The left-most column in each table presented the participants.The cells in each participant's row presented the results describing each participants' performance.The order of the data did not follow the participants' number.The items ordered across the top row were ranked from easy to difficult based on the participants' judgments.The students were ranked with the best participant at the bottom and the lowest at the top.The line marked the division between "-" and "+".
Table 7 reported the GJT results.Generally, individual differences were attested.Particularly, the participants exhibited deviations from the ideal matrix to the right and the left of the dividing line.P6 and P5 developed the negator verbal la: 'not' before the anaphoric la: 'no'.The other items were not yet developed.P4 and P3 acquired the negator lam 'not' only, contradicting the prediction based on their position in the matrix.The negator lam 'not' appeared to be the most difficult to acquire by all the participants.While P3 acquired the anaphoric la: 'no', P4 did not.P13 and P17 also acquired lam 'not' before other items.P15 and P14 developed lajsa 'is not' at an early stage, contradicting the predictions.P17 appeared to lack knowledge of lan 'not', while he developed other items that were predicted to be more difficult.Further, P10, P8, P7, P11, P12, and P3 acquired the anaphoric la: 'no' only.P1, P9, P16, P13, and P15 developed both the anaphoric la: 'no' and the verbal la: 'not'.P18, P23, and P14 showed knowledge of the anaphoric la: 'no', the verbal la: 'not', and lan 'not'.P2, P19, and P17 acquired the anaphoric la: 'no', the verbal la: 'not', lan 'not', and ma: 'not'.Finally, P20, P21, P22, and P24 developed the anaphoric la: 'no', the verbal la: 'not', lan 'not', ma: 'not', and the nonverbal lajsa 'is not'.The negator lam 'not' appeared to be the most difficult, and it was not yet developed by the most proficient learners.
Table 7 showed that the Moroccan learners of MSA developed the negation items in a non-linear pattern.They also showed graded difficulty, and the items were developed as follows.
Anaphoric la: > la: > ma: > lan > lajsa > lam The PDT results were presented in table 8. Interestingly, the participants' performance on the PDT appeared the worst compared to the GJT.Violations of ideal matrix were registered.Specifically, P2 and P8 produced the verbal negator la: 'not' before lajsa 'is not' and the anaphoric la: 'not'.P13 produced lam 'not' at an early stage.P19 and P20 failed to produce la: 'not'.While P14 did not produce lam 'not', P21 failed to produce lan 'not'.Further, all the participants produced the anaphoric la: 'not'.While P17 and P18 produced la: 'not' and lajsa 'is not', P19 and P20 produced lam 'not' and lajsa 'is not' only.P14, P15, P16, P22, P23, and P24 produced lan, lam, la: 'not', and lajsa 'is not'.P9 produced all the items.
In table 8, it was shown that the Moroccan learners of MSA displayed graded difficulty in the development of the negation items.The latter were produced as follows.
la: > lajsa > lam > lan > ma:  Figure 2 reported error patterns from the PDT.Most of the participants overgeneralized the anaphoric la: 'no' to contexts of other negators.The other participants overutilized rare instances of other negators.For example, P15 and P22 used the verbal negator la: 'not'.P9 and P14 utilized lan 'not'.P17, P19, P20, P21 used lam 'not'.P9 and P16 produced lajsa 'is not'.
The investigation of the error production by the participants revealed some conclusions.The participants overgeneralized the anaphoric la: 'no' to contexts where the other negators were needed.They did not overgeneralize the other negators.Instances of overgeneralizing la: 'no' are presented in examples (1), (2), and (3) respectively.In these examples, P1, P10, and P12 overgeneralized the anaphoric la: 'no' to contexts where the negators lam, lan 'not', and lajsa 'is not' were needed.(1) Erroneous production (P1) la: no 'no' Correct production lam ta-qt ʕ if Sara kull-a t-tuffa:ĥ-i not.past3SG.F.IMPFV-take.JUSS Sara all-ACC the-apple.PL-GEN 'Sara did not take all the apples.' (2) Erroneous production (P10) la: no 'no' Correct production ħa:da: lajsa kalb-a-n this.M is.not dog-ACC-INDEF 'This is not a dog' (3) Erroneous production (P12) la: no 'no' Correct production lan ja-ʔkul-a θ-θaʕlab-u l-ʁura:b not.future3SG.M.IMPFV-eat-SBJV the-fox-NOM the-raven 'The fox will not eat the raven.'On a careful comparison of participants' comprehension and production (see figures 1-A, 2-A, 3-A, 4-A, 5-A, and 6-A in Appendix A), comprehension of MSA negation was better compared to production.The participants performed better on the GJT but worse on the PDT.They showed increased sensitivity to ungrammatical sentences, but they exhibited difficulties at the production level.

Discussion
The current study sought to uncover typically developing bilingual Moroccan child learners' (of Moroccan Arabic and MSA) ability to comprehend and produce MSA negation.We predicted the complexity of the MSA negation to delay its development, compared to other languages like English.Interestingly, both the comprehension and production results revealed that the participants faced graded difficulties in the development of negation until the age of 9;1.The results was in line with White et al. (2023) who reported that the complexity of Afrikaans negation posed delaying effect on typically developing Afrikaans children's comprehension and production beyond the age of 5 years.Data from the Arabic varieties also reiterate the same results.For example, Omar (1973) confirmed that by the age of five years, child learners of Egyptian Arabic were still facing difficulties with negation structures.Al-Buainain (2003) also reported that child learners of Qatari Arabic continued to face difficulties in the development of negation by the age of seven years.However, the results contradict the findings from English speaking children in Pea's (1980) study who were argued to fully acquire negation by the age of 4;6.Regarding the question of whether comprehension precedes production or vice versa, the results showed that the participants' overall comprehension of MSA negation preceded production, contra Wojtecka et al. (2011Wojtecka et al. ( , 2013)), who claimed that production preceded comprehension.
The analysis of implicational scaling on the GJT and PDT left us with intriguing results.Moroccan MSA learners were found to develop negation in a non-linear order due to the attested violations of the ideal matrix, displaying characteristic of U-shaped learning curve (Pauls et al., 2013;Carlucci & Case, 2013).At the level of comprehension, the less proficient participants appeared to have acquired lam, lan, la: 'not', and lajsa 'is not' in the first stage while the more proficient participants performed worse on the same items, in the second stage.In the third stage, the most proficient participants fully acquired lan, la: 'not', and lajsa 'is not' with no violations of the ideal matrix.Similarly, at the level of production, the less proficient participants produced lan, la: 'not' in the first stage while the more proficient participants performed bad on lam, lan, la: 'not', in the second stage.In the third stage, the most proficient participants fully acquired all the items with no violations of the ideal matrix.

Conclusion
The current study examined negation in the comprehension and production of Moroccan child learners of MSA.First, in line with prediction, the complex structure of MSA negation was found to delay the development of the targeted negation structures.Second, the participants' comprehension of MSA negation was more robust compared to production.This could be explained by difficulties in the implementation of grammatical knowledge.The participants' production was affected by the computational or processing load.That is, Moroccan child learners of MSA had less working memory capacities available for them, and they showed a processing delay when they were required to produce the complex MSA negation.To avoid the complex negative structures that created computational load or required complex syntactic processing, the participants used the simplified anaphoric la: 'no'.This behavior could mean that the participants were following an 'economy principle' (Chomsky, 2005), avoiding more costly negative structures.Third, reporting individual results, the study suggested that the child learners of MSA showed a non-linear pattern of negation acquisition, displaying a U-shaped learning curve.Finally, further research is needed to investigate MSA properties because it would provide valuable test cases for various important questions in linguistic theory and language acquisition, including our understanding of the nature, development, representation, and processing of MSA.

ADHERENCE TO ETHICAL STANDARDS
Ethics Declarations.All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards.
Data Availability.The data supporting the findings of this study are available within the manuscript.Raw datasets generated during the current study are available from the corresponding author, Dr. Khalid Lahbibi, upon reasonable request.

Comparing the Mean Percentage Accuracy of Comprehension and Production
ʕ -t ʕ ifl-u not.PST 3SG.M.IMPFV-play-JUSS the-boy-NOM 'The boy did not play.'b. *lam laʕiba t ʕ -t ʕ ifl-u not.PST play.3SG.M.PFV the-kid-NOM 'The boy did not play.' Figure 2Participants' Error Types and Percentages in the PDT

Funding.
No funding was received to conduct this research.Conflicts of Interest.The authors declare no conflict of interest.Author contributions.Lahbibi K.: the idea, modeling of general design of the research, formulation of goals and objectives of the research, general organization of empirical research, writing abstracts to the article.Preparation of the initial version of the manuscript, planning and managing the implementation of experimental activity, selection of stimulus material, drawing up a list of sources used in the APA style.Data collection and analysis, interpretation of data of the research, preparation of tables and figures in their original form.Preparation of documents and collection of informed consent from potential participants of the experiment.Yeou M.: reviewing and editing Appendix A 1. Raw Numbers for Each Participant on the GJT and PDT of the Empirical Study.

Figure
Figure 1-A Mean Percentage Accuracy of Comprehension and Production of the Anaphoric /la:/ 'no'

Table 6
Information about the Participants

Table 7
Implicational Scale Showing the Development of Negation by Moroccan Learners of MSA on the GJT

Table 8
Implicational Scale Showing the Development of Negation by Moroccan Learners of MSA on the PDT

Table 1 -
A Number and Percentage of Correct Instances of Negative Particles in MSA out of the Sentence Pairs Judged by the Participants on the GJT

Table 2 -
A Number and Percentage of Correct Instances of Negative Particles in MSA out of the Sentence Pairs Judged by the Participants on the PDT