A Developmental Study of Lexico‐Phonological Distance in Vocabulary Learning: Evidence From Typically Developing and Learning‐Disabled Arabic‐Speaking Children

ABSTRACT

The present study examines the role of lexico‐phonological distance in shaping vocabulary acquisition within Arabic diglossia, drawing on 299 Arabic‐speaking children with and without learning disabilities across Grades 1–6. Performance followed a consistent gradient: identical words were easiest; unique words hardest; cognates fell between. This hierarchy held across both receptive and expressive tasks and remained stable over development; the Grade × Word Type interaction was non‐significant, indicating that distance‐related constraints remain stable with schooling. More telling was the Group × Word Type interaction: the performance gap between typically developing (TD) and learning‐disabled (LD) children widened disproportionately for unique words, where the demands of representational construction meet reduced phonological processing resources. A modality asymmetry further qualified the picture: expressive vocabulary proved more sensitive to diglossic distance than receptive vocabulary, consistent with the heavier retrieval demands production places on underspecified representations. Theoretically, the findings are interpreted within MAWRID, which treats linguistic distance as a constraint on representational quality and the Lexical Quality Hypothesis, which links representational precision to lexical access efficiency. These findings implicate lexico‐phonological distance as a structural, developmentally persistent constraint on lexical acquisition in Arabic diglossia, with disproportionate consequences for children with compromised phonological resources.

Practitioner Points

• Assess both receptive and expressive vocabulary to fully capture lexical development in Arabic diglossic learners.

• Prioritise identical and cognate words to ease vocabulary acquisition in early grades.

• Gradually introduce unique Standard Arabic words with structured, scaffolded support.

• Focus on expressive vocabulary difficulties among children with learning disabilities.

• Design vocabulary interventions aligned with children's spoken Arabic to bridge diglossic gaps.

1. Introduction

Arabic is a diglossic language, characterised by the coexistence of two varieties of the same language that are functionally distributed across different social contexts and communicative purposes (Ferguson 1959) and for different purposes and functions (Albirini 2016).

The Spoken Arabic (SpA) is acquired early and used in daily interactions, whereas Standard Arabic (StA) is the formal register taught in schools and used in written texts, public speech and official domains (Ferguson 1959). This linguistic gap has been shown to complicate language development and literacy acquisition for Arabic‐speaking children. A substantial body of psycholinguistic research has systematically examined this challenge through the lens of linguistic distance between SpA and StA (e.g., Tallas‐Mahajna et al. 2022; Saiegh‐Haddad 2003, 2004, 2007; Saiegh‐Haddad and Schiff 2016; Saiegh‐Haddad and Haj 2018; for a comprehensive review, see Saiegh‐Haddad 2022).

The effects of diglossia on language and literacy development in Arabic have attracted sustained empirical attention, pursued most systematically within a psycholinguistic research programme that treats diglossia not as a categorical condition but as a gradient of measurable linguistic distance between SpA and StA. This operationalization, first introduced by Saiegh‐Haddad (2003, 2004), reframed diglossia as an empirically tractable variable, one whose magnitude could be assessed at the level of individual linguistic units rather than assumed uniformly across the language system. From that foundation, the programme has expanded considerably, tracing the consequences of lexico‐phonological distance across multiple developmental domains: phonological processing, morphological awareness and reading development have each been examined in turn, yielding a cumulative picture of how structural divergence between the two Arabic varieties shapes children's acquisition of language and literacy skills.

Recent studies have begun to examine vocabulary acquisition in Arabic diglossia through the lens of linguistic distance. Tarabi‐Juma'a et al. (2026) tested the role of linguistic distance in expressive and receptive vocabulary knowledge of nouns and verbs by manipulating lexical and phonological distance across identical, cognate and unique word types, demonstrating graded effects in both modalities with stronger constraints in production. Ghawi‐Dakwar and Saiegh‐Haddad (2025) showed that phonological distance directly constrains word learning, with larger effects in production and in children with developmental language disorder. These studies establish that lexico‐phonological distance is a determinant of vocabulary performance. While these studies establish that lexico‐phonological distance is a key determinant of vocabulary performance, they do not examine how these effects develop across the full elementary school years, nor whether they differ as a function of learning status (TD vs. LD).

Studies have further differentiated among word types based on their distance between SpA and StA (namely, identical, cognate and unique forms) as defined and tested by Saiegh‐Haddad and Spolsky (2014). Building on this research programme, the present study aims to address two key gaps: (1) the developmental trajectory of sensitivity to lexical and lexico‐phonological distance across Grades 1–6 and (2) the role of learning status (TD vs. LD), theoretically motivated by evidence that LD is associated with phonological processing and representational vulnerabilities that are predicted to amplify distance‐based constraints on vocabulary.

2. Theoretical Framework

Two complementary theoretical frameworks guide the interpretation of the present study. The Lexical Quality Hypothesis (LQH; Perfetti 2007) provides the general architecture: it holds that the efficiency of word recognition and comprehension depends on the quality and integration of phonological, orthographic and semantic representations, with the phonological component playing a particularly central role. The MAWRID model (Saiegh‐Haddad 2018) applies this logic specifically to Arabic diglossia, treating lexico‐phonological distance as the primary constraint on how lexical representations consolidate over development. Together, these frameworks generate a coherent set of predictions for vocabulary acquisition in a diglossic context. These dynamics, referring to the impact of lexico‐phonological distance on representational quality and lexical access, are examined within the framework of the Lexical Quality Hypothesis (LQH; Perfetti 2007), which posits that the efficiency of word recognition and comprehension is determined by the quality and integration of phonological, orthographic and semantic representations. High‐quality lexical representations depend critically on the precision and stability of their phonological component. When these representations are underspecified or inconsistently activated, lexical access becomes slower and less reliable.

Within the Arabic diglossic context, converging evidence points toward a consistent conclusion: linguistic distance between SpA and StA systematically compromises the quality of phonological representations. Phonological structures that have no counterpart in the spoken vernacular are, as early work showed, systematically harder for children to access and manipulate (Saiegh‐Haddad et al. 2011). The difficulty was traced to the representations themselves, less stable, less fully specified and less firmly consolidated than those anchored in familiar spoken forms. Saiegh‐Haddad and Haj (2018) pursued this further, finding that the effect is not simply present or absent but graded: representational accuracy declines as the distance between SpA and StA forms increases and this gradient holds across development rather than attenuating with age.

The theoretical machinery for interpreting this pattern is provided by MAWRID (Saiegh‐Haddad 2018). The model conceptualises diglossia as a source of lexical and lexico‐phonological distance that constrains how phonological representations consolidate over time and, by extension, limits the automaticity with which words are recognised. Distance, on this account, is not merely a structural property of the language system; it is a processing burden with measurable developmental consequences.

These theoretical and empirical considerations converge on a set of testable predictions for the present study. If representational quality degrades systematically as lexico‐phonological distance increases (as demonstrated by Saiegh‐Haddad and Haj (2018) across phonological tasks), then a parallel gradient should emerge in vocabulary acquisition: identical words, whose representations are consolidated through spoken language experience, should be acquired most readily; cognate words, which partially overlap with spoken forms, should show intermediate performance; and unique words, which require building representations without any spoken scaffolding, should prove most difficult. This prediction follows directly from the representational quality account embedded in MAWRID, rather than from exposure frequency alone. Critically, because Saiegh‐Haddad and Haj (2018) found that this representational gradient persisted across development without attenuating (remaining stable from kindergarten through Grade 6 despite increasing exposure to StA), the MAWRID account predicts that the word‐type hierarchy in vocabulary acquisition will likewise remain stable across the elementary school years, rather than narrowing as children accumulate literacy experience. Furthermore, because children with LD exhibit documented weaknesses in phonological encoding and representational stability, the MAWRID‐based account predicts that distance‐related effects will be amplified in this group, not merely that they will perform lower overall, but that the performance gradient across word types will be steeper relative to their typically developing peers.

3. Lexico‐Phonological Distance and Lexical Processing in Arabic

Arabic is a classic case of diglossia, characterised by the coexistence of two linguistically distinct forms: SpA and StA (Ferguson 1959). This sociolinguistic situation presents a unique developmental challenge: children acquire SpA as their native spoken language, whereas StA is primarily encountered through literacy and formal instruction and is rarely used in spontaneous everyday speech. The two varieties differ significantly, though not uniformly, across phonological, morphological, syntactic and lexical domains (Asadi, Atila, and Saleh 2024; Asli‐Badarenh and Leikin 2019; Saiegh‐Haddad and Henkin‐Roitfarb 2014).

The phonological divergence is particularly consequential. StA contains phonemes (among them the voiceless interdental fricative /θ/, the voiced interdental fricative /ð/ and the emphatic voiced interdental /ðˤ/) that are absent from many spoken vernaculars. Phonemes with no representation in a child's spoken system are, predictably, less accurately encoded and harder to retrieve. The two varieties also differ in permissible syllable structures: certain phonotactic patterns found in spoken dialects are unattested in StA and vice versa (Saiegh‐Haddad and Henkin‐Roitfarb 2014). The practical consequence is that children approaching literacy in StA must process phonological structures their prior spoken language experience has left only partially prepared.

To understand how diglossia specifically affects vocabulary acquisition, it is essential to examine the role of lexico‐phonological distance, given that linguistic distance between SpA and StA has been shown to undermine the quality of phonological representations (Asadi and Asli‐Badarneh 2026; Saiegh‐Haddad et al. 2011; Saiegh‐Haddad and Haj 2018) and that the precision of these representations is central to vocabulary acquisition within the Lexical Quality framework (Perfetti 2007). Examining the role of lexical and lexico‐phonological distance is therefore critical for understanding vocabulary development in Arabic diglossia. This perspective is further supported by MAWRID, which identifies linguistic distance as a key factor shaping the development of lexical representations in Arabic literacy (Saiegh‐Haddad and Haj 2018).

Lexico‐phonological distance refers to the extent of overlap or divergence in both the lexical and phonological forms of words across the two varieties. Saiegh‐Haddad and Spolsky (2014), based on a corpus of 4500 word types used by kindergarteners, proposed a tripartite classification of words: (1) identical words (comprised around 21.2% of the word types in the corpus examined), which share both lexical and phonological form in SpA and StA (e.g., /karaz/ ‘cherry’); (2) cognate words (40.6%), which are paired lexical items that share the same form and meaning, although this form is only partially phonologically overlapping (e.g., StA /qalb/ – SpA /ʔalb/ (‘heart’)); and (3) unique words (38.2%), which are entirely different in their lexical and phonological forms between StA and SpA (e.g., /StA dalew/ – /SpA satel/ ‘pail’).

In a recent study, Asli‐Badarneh et al. (2023) looked at how diglossia influences the vocabulary choices children make when telling stories. They focused on Arabic‐speaking immigrant children aged 7 to 12 living in Canada. The results showed a clear preference for SpA: 59.46% of the word types came from this variety. This included identical words (21.55%), SpA cognates (25.73%) and words unique to SpA (12.18%). StA, by contrast, was used less often, making up 38.59% of the vocabulary, split between StA cognates (19.97%) and words unique to StA (18.62%). One of the standout findings was that children used more SpA cognates than StA ones, pointing to the greater accessibility of SpA forms of words in their mental lexicon. These results highlight how familiar, everyday language is more easily accessible in word retrieval during narrative tasks, especially when compared to the more formal register of StA.

Representational quality is, in this respect, the crux of the matter. Saiegh‐Haddad and Haj (2018) examined this issue by assessing the quality of phonological representations for identical, cognate and unique word types across four developmental stages: senior kindergarten and Grades 1, 2 and 6. The results revealed a consistent gradient pattern, whereby identical words were represented with the highest accuracy, cognate words showed intermediate levels of precision and unique words, which are more distant from spoken forms, yielded the lowest accuracy across all age groups. Notably, this pattern remained stable even among sixth‐grade children, despite several years of exposure to Standard Arabic through literacy. This finding suggests that increased exposure alone does not fully compensate for the representational constraints associated with lexico‐phonological distance. Instead, linguistic distance appears to exert a persistent influence on the precision and stability of phonological representations across development. These findings suggest that linguistic distance imposes a long‐term constraint on the establishment and refinement of phonological representations, even for familiar lexical items. Within the framework of the Lexical Quality Hypothesis (Perfetti 2007), imprecise representations are not a neutral condition: they carry direct consequences for vocabulary knowledge and the efficiency of lexical access.

Diglossic effects on reading are by now well attested across a range of measures and populations (Asadi and Asli‐Badarneh 2023; Asli‐Badarneh and Asadi 2023; Asadi, Atila, and Saleh 2024; Asadi and Kawar 2024; Saiegh‐Haddad 2003; Saiegh‐Haddad and Schiff 2016; Schiff and Saiegh‐Haddad 2017; Taha and Saiegh‐Haddad 2017). What these studies collectively reveal is not a uniform reading impairment but a patterned one: performance varies with linguistic distance rather than declining across the board. Saiegh‐Haddad and Schiff (2016) documented this directly: fluency was higher for forms closer to the spoken vernacular and lower for those that diverge from it. Reading efficiency is less a function of print exposure than of the quality and accessibility of the phonological representations that reading must draw upon.

Dyslexic readers make the point more forcefully still. Schiff and Saiegh‐Haddad (2017) found that children with developmental dyslexia read StA with markedly lower accuracy and fluency than SpA, in both voweled and unvoweled conditions. SpA performance, moreover, predicted StA reading fluency, suggesting the two are not processed in isolation: reading in the standard variety draws on whatever representational resources the spoken variety has established. Where those resources are thin, StA reading suffers accordingly.

These findings suggest that the difficulties observed in reading StA are not solely orthographic in nature, but reflect underlying weaknesses in phonological representations, particularly for forms that are distant from the spoken language. This pattern is expected to be more pronounced in children with learning disabilities, who are known to exhibit reduced efficiency in phonological processing and representational stability.

4. Vocabulary Development in Diglossic Arabic

Vocabulary acquisition is a foundational component of language development and a reliable predictor of academic success and literacy achievement (He and Deocampo 2023; Johnson 2018). In Arabic, however, the story is more complex. From an early age, children speak a local dialect (SpA) which is rich and fully functional in everyday life, yet the variety they encounter in books and classrooms (StA) comes later, mainly through schooling (Asadi and Abu‐Rabia 2021; Saiegh‐Haddad and Spolsky 2014). This sociolinguistic context creates a clear disparity in vocabulary exposure between the two forms.

Research on diglossia and vocabulary development can be organised along two methodological lines. A first body of work invokes diglossia as a broad contextual factor to account for observed vocabulary gaps, without directly manipulating linguistic distance as an independent variable. A second, more recent line of research moves beyond contextual accounts by experimentally operationalizing lexico‐phonological distance, comparing performance across word types that differ systematically in their degree of overlap between SpA and StA (e.g., Ghawi‐Dakwar and Saiegh‐Haddad 2025; Tarabi‐Juma'a et al. 2026). The present study builds on this latter approach. The contextual line of research is reviewed first.

Fedda and Oweini (2012) examined Lebanese preschoolers and students in grades 1 through 5 using a standardised picture vocabulary test and found that Arabic vocabulary scores often fell below expected levels, particularly among older children. The researchers attributed this shortfall mainly to diglossia and limited exposure to StA vocabulary in everyday contexts. Supporting this, a longitudinal study by Oueini et al. (2020) found that oral proficiency in StA did not show significant gains until around the fourth grade, by which time many children had already become more fluent in English, their second language. Together, these studies establish that diglossic exposure patterns constrain vocabulary development, but they do not identify the specific linguistic mechanism responsible.

4.1. Evidence From Intervention Research

A mechanistic account is supported by intervention studies that have directly targeted the representational demands imposed by diglossia. Saiegh‐Haddad (2023) implemented a diglossia‐centred intervention in which metalinguistic awareness and linguistic representations were first established in SpA and only then extended to StA, reflecting the assumption that learning in StA depends on the availability and quality of underlying spoken‐based representations. Stronger causal evidence comes from the ADMIN programme (Haj, Ghawi‐Dakwar, et al. 2026), a diglossia‐centred multi‐domain intervention that produced significant gains not only in receptive and expressive vocabulary but in the quality of lexico‐phonological representations themselves. Children in the intervention group showed improved access to and production of StA lexical items, more accurate phonological encoding across all three distance types: identical, cognate and unique and stronger metalinguistic skills. Complementing this, Shahbari‐Kassem et al. (2024) demonstrated that gradual shifts in lexical distance, from identical to cognate to unique forms, facilitate learning, suggesting that linguistic distance constrains the accessibility and learnability of lexical items. Collectively, this intervention evidence supports a representation‐based account: lexical development in Arabic is constrained by linguistic distance, but can be facilitated when instruction directly targets that constraint.

4.2. Evidence From Experimental Studies of Distance and Modality

A parallel line of research has tested the role of linguistic distance more directly, by comparing vocabulary performance across experimentally controlled word types. Haj, Saiegh‐Haddad, et al. (2026) showed that listening comprehension in StA is strongly predicted by linguistic and lexical skills in SpA, as well as by the quality of lexico‐phonological representations. Most directly relevant to the present study, Tarabi‐Juma'a et al. (2026) demonstrated a consistent graded distance effect: identical > cognate > unique in both receptive and expressive vocabulary, with distance‐related difficulties more pronounced in production. This modality asymmetry reflects the greater demands of phonological retrieval in expressive tasks: within a representational quality framework, underspecified phonological representations disproportionately impair production relative to recognition. Ghawi‐Dakwar and Saiegh‐Haddad (2025) extended this logic to word learning, showing that phonological distance constrains the acquisition of new lexical items, with amplified effects in children with developmental language disorder whose phonological retrieval resources are already strained.

Converging evidence from these contextual, interventional and experimental lines therefore supports a unified account in which linguistic distance constrains vocabulary acquisition by limiting the quality, accessibility and integration of lexical‐phonological representations. Words that are lexically and phonologically distant from the spoken vernacular are less efficiently learned and accessedת and this constraint operates with particular force in production and among children with processing vulnerabilities.

In conclusion, Tarabi‐Juma'a et al. (2026) demonstrated that lexico‐phonological distance constrains both receptive and expressive vocabulary knowledge, with stronger effects in production. The present study extends this work in two theoretically motivated ways: it traces these effects across the full elementary school span (Grades 1–6) to examine whether representational constraints are developmentally persistent and it examines whether the effects are amplified in children with LD, whose known phonological processing deficits are predicted to compound the representational consequences of diglossic distance.

5. Vocabulary Acquisition in Typical and Atypical Development

Vocabulary development shows clear differences between typically developing (TD) children and those with learning disabilities (LD), especially in linguistically complex environments such as Arabic diglossia. Typically developing children tend to build their vocabulary steadily throughout the school years, supported by increased exposure, reading experience and metalinguistic growth (Gray 2004; Jackson et al. 2021).

In contrast, children with LD often experience persistent difficulties in vocabulary acquisition, which have been linked to weakness in phonological processing and working memory, which consequently made the formation of lexical representations less stable (Elbro and Jensen 2005; Gathercole and Baddeley 1990; Mengisidou and Marshall 2019). It is important to note, however, that LD is a heterogeneous category encompassing different profiles (e.g., dyslexia, poor comprehenders), which may involve distinct underlying mechanisms. In the present study, LD is treated as a broad educational classification based on school diagnoses and thus the findings should be interpreted at this general level rather than as reflecting specific subtypes.

These challenges are expected to be amplified in diglossic contexts. Arabic‐speaking children with learning difficulties must acquire StA, a variety to which they have relatively limited exposure, while simultaneously contending with weaknesses in phonological processing. Crucially, how diglossia constrains lexical development is, at this point, reasonably well understood at the mechanistic level. Saiegh‐Haddad and Haj (2018) showed that lexico‐phonological distance between SpA and StA bears directly on representational quality: identical words were encoded most accurately, cognates fell in the middle and unique forms, the most distant from spoken experience, yielded the lowest accuracy. This gradient held across development, not just at a single age point, establishing phonological distance as a systematic determinant of how precisely lexical‐phonological representations consolidate in Arabic.

Ghawi‐Dakwar and Saiegh‐Haddad (2025) pushed this logic into word learning directly, demonstrating that phonological distance constrains the acquisition of new lexical items in measurable ways. Children performed significantly worse on phonologically distant forms and the effect surfaced in both comprehension and production. Children with developmental language disorder (DLD) (whose difficulties with phonological encoding, storage and retrieval are well documented) showed not only lower overall performance but a more stubborn sensitivity to distance, particularly on production tasks where phonological retrieval demands are highest. Saiegh‐Haddad and Ghawi‐Dakwar (2017) had already pointed in this direction: repetition accuracy in children with DLD dropped sharply when target words contained phonological patterns absent from SpA, suggesting that the spoken system's gaps translate directly into representational instability under production pressure.

What emerges from this body of work is a coherent picture. Vocabulary acquisition in Arabic diglossia is shaped by the intersection of two sources of vulnerability (linguistic distance and phonological representational weakness) and the two are not independent. Phonologically distant StA forms are harder to acquire than closer ones, with performance following the now‐familiar gradient: identical words outperform cognates, which outperform unique forms. Among children with learning difficulties, this gradient is expected to steepen, most visibly on expressive tasks that place the heaviest demands on precise phonological retrieval.

6. Research Gap

The evidence reviewed above leaves two gaps that the present study is designed to address. Tarabi‐Juma'a et al. (2026) documented lexico‐phonological distance effects on vocabulary, but the developmental arc of those effects was not their focus; whether the representational constraint associated with word type holds steady, intensifies or erodes as children move through elementary school remains unanswered. Grades 1 through 6 represent a period of substantial literacy growth and it is far from obvious that a gradient documented at one point in development will look the same at another.

The second gap is more pointed. No study to date has directly tested whether lexico‐phonological distance effects on receptive and expressive vocabulary differ between TD children and those with LD across development. The comparison is not merely descriptive in motivation. Children with LD exhibit phonological processing and memory deficits (Elbro and Jensen 2005; Gathercole and Baddeley 1990; Mengisidou and Marshall 2019) that both MAWRID and the Lexical Quality Hypothesis predict will amplify representational degradation precisely where lexico‐phonological distance is highest, making LD status not a covariate to control for, but a theoretically informative variable in its own right.

The present study pursues both lines. It examines how differences in representational quality tied to lexical and lexico‐phonological distance translate into differences in receptive and expressive vocabulary across Grades 1–6 and it tests whether learning status moderates those effects, asking, in short, whether the burden of diglossic distance falls harder on children whose phonological representations and phonological memory are already compromised.

7. Current Study

Building on this body of research, the present study examines how lexical and lexico‐phonological distance shapes receptive and expressive vocabulary in Arabic diglossia across Grades 1–6 and whether these effects differ between typically developing (TD) children and children with learning disabilities (LD). In addition to these primary research questions, the study explores an asymmetry between receptive and expressive vocabulary as a function of lexico‐phonological distance. Although modality comparison was not a primary focus of the study, prior work suggests that production places heavier demands on phonological retrieval than recognition does (Ghawi‐Dakwar and Saiegh‐Haddad 2025; Tarabi‐Juma'a et al. 2026), raising the question of whether distance‐related effects are amplified in expressive relative to receptive vocabulary, particularly among children with LD whose phonological retrieval resources are already strained.

The Research Questions are as follows:

1. How does lexical and lexico‐phonological distance (identical, cognate, unique) affect receptive and expressive vocabulary performance among Arabic‐speaking children?

2. How do these effects vary across developmental stages (Grades 1–6)?

3. To what extent are these effects modulated by learning status (typically developing vs. learning disabilities)?

8. Method

8.1. Research Design

The present study used a cross‐sectional quantitative design to explore how different lexico‐phonological distance of words (identical, cognate and unique) affect receptive and expressive vocabulary among Arabic‐speaking children. Participants were drawn from six elementary school grade levels (1st through 6th), enabling a developmental perspective on the relationship between word type and vocabulary performance. The study compared the performance of two groups: typically developing (TD) children and children with learning disabilities (LD). The independent variables were word type (Identical, Cognate and Unique), grade level (1st–6th grade) and group (TD vs. LD). The dependent variables were receptive and expressive vocabulary scores. Data were analysed using a multivariate analysis of variance (MANOVA) to explore main effects and interactions among the variables.

8.2. Participants

The sample comprised 299 Arabic‐speaking children attending elementary schools in Israel, ranging from first to sixth grade. 150 were classified as typically developing (TD), based on school records indicating typical academic progress and the absence of identified learning difficulties, while the remaining children (n = 149) were identified by the school system as having learning difficulties.

It is important to clarify that the term learning disabilities (LD) in the present study refers to school‐identified learning difficulties, primarily in reading and academic domains and does not correspond to a clinical diagnosis of developmental language disorder (DLD).

Group classification was based on official school records and professional evaluations conducted by qualified personnel (e.g., school psychologists and learning disability specialists), either within the school system or through external services. These evaluations included assessments of academic achievement, cognitive processes related to learning and teacher reports. The research team was not involved in the diagnostic process.

While detailed diagnostic profiles were not available for all participants, the LD group predominantly included children identified with reading‐related learning difficulties rather than primary language impairments.

All children in both groups were within the normal range of cognitive ability, as indicated by the standardised assessments included in the diagnostic process.

It is important to note that in the lower grades, particularly in Grade 1, group classification was based on school‐based identification procedures, including teacher evaluations and early indicators of learning difficulties, rather than a confirmed clinical diagnosis.

Each of the six grade levels included an equal number of participants from both groups, with 25 typically developing children and 25 children with learning disabilities in each grade, resulting in 50 children per grade, except for 49 children in first grade.

The sample was balanced in terms of gender and drawn from a range of schools representing lower to middle socio‐economic backgrounds. Inclusion criteria included native Arabic speakers with no hearing, visual or neurological impairments. Informed parental consent was obtained for all participants. The study was approved by the ethics committee of the Arab Academic College in Haifa and the Israeli Ministry of Education's Chief Scientist.

8.3. Measures

8.3.1. Vocabulary Tests

Two purpose‐built vocabulary assessments were employed: a Receptive Vocabulary Test and an Expressive Vocabulary Test, each comprising a total of 90 target items. In both tests, the items were evenly divided among three word‐type categories to reflect differing levels of lexico‐phonological distance (according to the lexical categories introduced in Saiegh‐Haddad and Spolsky 2014): 30 identical words (lexically and phonologically the same in both SpA and StA), 30 cognate words (which differ yet remain related in their phonological forms across the two varieties, sharing some phonological characteristics while differing, often systematically, in others; e.g., StA /miqlama/ – SpA /miɁlami/) and 30 unique words (words that exist solely in StA, with no overlap with SpA). Examples of unique words include: /dalew/ (pail; StA) vs. /satel/ (SpA equivalent) and /ðahab/ (gold; StA) vs. /dahab/ (SpA). The stimulus set comprised both nouns and verbs (e.g., nouns: /burtuqal/ ‘orange’, /kitab/ ‘book’; verbs: /qafaza/ ‘jumped’, /rasama/ ‘drew’), distributed equally across the three word‐type categories, with 15 nouns and 15 verbs per category. Grammatical class was not treated as an independent variable but was controlled through balanced distribution across conditions.

All items were presented in StA. The selection of target words for both tests was guided by well‐defined developmental and linguistic criteria. Specifically, words were selected from elementary school textbooks commonly used in the participating schools, ensuring that the items were familiar and appropriate for the targeted grade levels. Items were evaluated against the following criteria and only words meeting all criteria were retained: (a) age‐appropriateness, verified against curriculum‐grade expectations; (b) frequency of occurrence in the Al‐Fanus educational corpus of 70 Ministry‐of‐Education‐recommended storybooks; (c) word length (2–4 syllables); (d) syllabic structure (CV, CVC, CVCV patterns); (e) phoneme inventory, unique StA items were required to include at least one phoneme absent from the spoken dialect; and (f) morphological transparency. Items were matched across word‐type categories on syllable count and length to ensure that any observed performance differences reflect lexico‐phonological distance rather than surface‐level complexity. Items were systematically selected to ensure comparability across categories in frequency, length and syllabic complexity, thereby minimising potential confounds unrelated to lexico‐phonological distance.

Each word was clearly classified into one of the three lexico‐phonological categories based on comparisons between the children's spoken dialect and StA, thereby capturing varying levels of linguistic distance.

The test items were presented in a fixed order reflecting the intended gradation of lexico‐phonological distance across word types. All children received the full 90‐item test (30 items per word‐type category).

The stimulus set included both nouns and verbs. These were distributed across the three word‐type categories; however, grammatical class (noun vs. verb) was not treated as an independent variable in the analysis but was controlled for through balanced representation across conditions.

Raw scores were calculated as the number of correct responses out of 30 for each word‐type category and were converted into proportional accuracy scores (0–1) per word type.

Receptive Vocabulary Test. In this task, each child was presented with 90 spoken words, evenly distributed across the three lexical categories: 30 identical, 30 cognate and 30 unique items. For each item, the child was asked to select the corresponding picture from a set of four options. The response options consisted of one target image and three distractors. Distractors were semantically related to the target item (e.g., belonging to the same category) while differing clearly in lexical identity, in order to minimise random guessing and ensure that successful performance required accurate lexical knowledge rather than perceptual matching. Care was taken to ensure that distractors were comparable in visual complexity and familiarity across items. This format assessed the child's recognition and understanding of the word's meaning. Each response was scored dichotomously (1 = correct, 0 = incorrect) and separate scores were calculated for each word type. The analysis Cronbach's alpha value was 0.91.

Expressive Vocabulary Test. In this task, children were presented with 90 individual pictures and asked to name each depicted item aloud. The test comprised an equal distribution of word types, including 30 identical, 30 cognate and 30 unique words, maintaining parallel structure with the receptive vocabulary assessment. This task aimed to examine how well the child could produce a word when presented with a visual picture. Children were explicitly instructed to respond in StA. A score of 1 was assigned to responses that matched the target word in StA, while a score of 0 was given if the response was either incorrect or not provided. Responses produced in SpA were not considered correct, as the task specifically targeted lexical retrieval in the standard variety. Phonologically distorted responses were considered incorrect if they altered the target form in a way that affected word identity or resulted in a non‐standard or unrecognisable form. Minor phonetic variations that did not affect lexical identity were accepted as correct. Scores for identical, cognate and unique words were recorded separately to enable meaningful comparisons across word types. The analysis Cronbach's alpha value was 0.88.

8.3.2. Procedure

Testing was conducted individually in quiet rooms within the school environment. Graduate students in linguistics and learning disabilities, fluent in the children's dialect, administered the tasks. The vocabulary tasks were given in one session lasting approximately 30 min per child. The order of tasks was counterbalanced across participants to minimise order effects.

8.4. Ethical Considerations

All ethical procedures were strictly followed. Participants' confidentiality and anonymity were maintained by assigning identification codes and securely storing all collected data. Written informed consent was obtained from the children's parents and verbal assent was given by the children themselves. The study was conducted in accordance with established ethical standards for research involving minors.

9. Results

The central question was whether word type (Identical, Cognate or Unique) shapes receptive and expressive vocabulary performance differently across TD children and those with LD and whether these patterns shift across six grade levels. A multivariate analysis of variance (MANOVA) was used to examine these effects. Significant main effects emerged for group, word type and grade level; several interactions further specified the nature of group differences, particularly for children with LD.

Preliminary analyses indicated that neither gender nor socioeconomic status exerted significant effects on vocabulary performance across any of the measures (p > 0.05) and both variables were excluded from subsequent analyses.

Table 1 presents the means and standard deviations for receptive and expressive vocabulary across groups and word types. The table is organised by word type (the primary independent variable) within each group to reflect the theoretical focus of the study on lexico‐phonological distance. TD children consistently achieved higher scores than children with learning disabilities, with Identical words eliciting the highest performance and Unique words proving the most difficult.

TABLE 1.

Means and standard deviations for receptive and expressive vocabulary by group and word type.

Group	Word type	Receptive vocabulary (M, SD)	Expressive vocabulary (M, SD)
Typically developing	Identical	90.2 (6.1)	86.7 (6.3)
	Cognate	83.5 (7.3)	78.4 (7.9)
	Unique	72.8 (8.5)	66.3 (9.3)
Children with reading disabilities	Identical	75.4 (7.9)	69.8 (8.3)
	Cognate	68.3 (8.6)	61.4 (9.2)
	Unique	56.7 (9.1)	49.5 (10.1)

Note: Values represent percentage accuracy (0–100).

Bartlett's test of sphericity revealed statistically significant intercorrelations among the dependent variables (χ ² (9, N = 299) = 165.43, p < 0.001), confirming the appropriateness of MANOVA. Box's M test (value = 52.17, F(50, 121,092.32) = 1.08, p = 0.39) and Levene's test supported the assumptions of equality of variance–covariance matrices and univariate homogeneity of variance, respectively.

MANOVA results revealed significant main effects of developmental group (TD vs. Children with disabilities), word type and grade level on receptive and expressive vocabulary scores. There was a statistically significant main effect of group, F(1, 298) = 81.25, p < 0.001, η ² = 0.22, indicating that TD children significantly outperformed children with learning disabilities across all dependent measures. Additionally, a significant main effect for word type was observed, F(2, 297) = 49.34, p < 0.001, η ² = 0.25. Identical words yielded the highest performance, followed by Cognates, while Unique words were the most challenging for both groups. This pattern is also illustrated in Figure 1, which shows the gradient effect of word type across groups. The interaction between group and word type was statistically significant, F(2, 297) = 15.42, p < 0.001, η ² = 0.11, reflecting a larger performance gap between TD and children with reading disabilities for Unique words compared to Identical and Cognate words.

FIGURE 1.

Mean receptive vocabulary scores by word type (identical, cognate, unique) and group (typically developing vs. learning disabilities).

Grade level also had a significant main effect, F(5, 288) = 26.89, p < 0.001, η ² = 0.19. Descriptive trends demonstrated consistent improvements in receptive and expressive vocabulary across grades for both TD and children with learning disabilities groups. However, TD children showed consistently higher performance and somewhat greater gains compared to children with learning disabilities. As shown in Table 2, vocabulary scores increased steadily across grade levels for both groups.

TABLE 2.

Means and standard deviations for vocabulary scores by grade level and group.

Dependent measure	1st TD	1st LD	2nd TD	2nd LD	3rd TD	3rd LD	4th TD	4th LD	5th TD	5th LD	6th TD	6th LD
Receptive (M, SD)	78.5 (6.4)	65.3 (7.2)	82.1 (6.1)	68.6 (7.3)	85.7 (5.9)	72.4 (7.0)	89.2 (5.5)	75.6 (6.9)	92.4 (5.4)	79.1 (6.8)	94.6 (5.3)	81.5 (6.9)
Expressive (M, SD)	73.9 (6.7)	58.4 (7.9)	77.4 (6.5)	61.7 (8.2)	80.8 (6.2)	65.3 (7.8)	84.1 (5.9)	68.7 (7.6)	87.3 (5.8)	71.5 (7.4)	89.7 (5.6)	74.2 (7.3)

Univariate tests confirmed that word type significantly influenced both receptive (F(2, 297) = 36.78, p < 0.001, η ² = 0.21) and expressive vocabulary (F(2, 297) = 34.12, p < 0.001, η ² = 0.20). Post hoc pairwise comparisons (Bonferroni‐corrected) indicated that performance differed significantly across all three‐word types. Specifically, identical words yielded significantly higher scores than cognate words (p < 0.001) and cognate words yielded significantly higher scores than unique words (p < 0.001). This gradient pattern is also evident in expressive vocabulary, as shown in Figure 2. Post hoc comparisons were also conducted to interpret the significant Group × Word Type interaction. These analyses confirmed that the magnitude of group differences varied significantly across word types, with the largest differences observed for unique words, followed by cognate words and the smallest differences for identical words (all ps < 0.001).

FIGURE 2.

Mean expressive vocabulary scores by word type (identical, cognate, unique) and group (typically developing vs. learning disabilities).

The interaction between word type and grade level was not statistically significant (F(10, 290) = 1.34, p = 0.22, η ² = 0.04), suggesting that the effect of word type was consistent across grades. However, the interaction between group and word type was significant (F(2, 297) = 12.91, p < 0.001, η ² = 0.10), further emphasising the challenges children with learning disabilities face with unique words. Post hoc pairwise comparisons (Bonferroni‐corrected) were conducted to further examine this interaction. These analyses revealed that children with learning disabilities performed significantly lower than typically developing children across all word types (all ps < 0.001). Importantly, the magnitude of the group difference varied as a function of word type, with the largest differences observed for unique words, followed by cognate words and the smallest differences for identical words.

Further analysis showed that while TD children outperformed children with learning disabilities across all word types. As shown in Tables 3 and 4, the gap was particularly pronounced for unique words (ps < 0.001), compared to cognate and identical words.

TABLE 3.

Means for receptive vocabulary by word type, grade and group.

Grade	Group	Identical	Cognate	Unique
1	TD	0.917	0.829	0.715
	LD	0.743	0.676	0.564
2	TD	0.912	0.843	0.721
	LD	0.765	0.677	0.567
3	TD	0.903	0.852	0.716
	LD	0.747	0.664	0.579
4	TD	0.903	0.836	0.708
	LD	0.745	0.660	0.563
5	TD	0.907	0.828	0.733
	LD	0.764	0.651	0.571
6	TD	0.921	0.837	0.725
	LD	0.749	0.673	0.565

Note: Values represent proportional accuracy (0–1).

TABLE 4.

Means for expressive vocabulary by word type, grade and group.

Grade	Group	Identical	Cognate	Unique
1	TD	0.867	0.781	0.665
	LD	0.681	0.640	0.528
2	TD	0.845	0.803	0.703
	LD	0.715	0.647	0.548
3	TD	0.844	0.781	0.701
	LD	0.724	0.625	0.512
4	TD	0.839	0.792	0.691
	LD	0.672	0.648	0.511
5	TD	0.833	0.789	0.700
	LD	0.721	0.616	0.507
6	TD	0.856	0.801	0.668
	LD	0.701	0.623	0.539

Note: Values represent proportional accuracy (0–1).

To further clarify the pattern of performance across word types, grade levels and groups within each task, descriptive means are presented separately for receptive and expressive vocabulary in Tables 3 and 4.

As can be seen, performance was consistently highest for identical words, followed by cognates and lowest for unique words across all grades and both groups.

Analysis of the receptive‐expressive gap revealed a significant difference between groups (F(1, 298) = 18.76, p < 0.001, η ² = 0.06). Children with learning disabilities showed a larger discrepancy between their receptive and expressive vocabulary scores compared to TD children particularly for Unique words.

10. Discussion

The objective of the current study was to examine the impact of lexico‐phonological distance between StA and SpA on vocabulary acquisition among TD children and those with LD across grades 1–6. Specifically, it investigated how word type (identical, cognate and unique) impacts both receptive and expressive vocabulary performance.

Vocabulary performance followed a clear gradient of lexico‐phonological distance, with identical words yielding the highest scores, cognate words showing intermediate performance and unique words yielding the lowest scores. This pattern was consistent across both receptive and expressive tasks and across all grade levels. Although overall performance improved with age, the relative ordering of word types remained stable. The absence of a significant Word Type × Grade interaction indicates that the effects of linguistic distance did not vary across the elementary school years. The hierarchy is established early and, it appears, persists. Together, these findings point to a coherent pattern linking lexico‐phonological distance, representational quality and vocabulary performance. While these findings describe the observed performance patterns, their underlying explanation lies in previously established differences in representational quality. As demonstrated by Saiegh‐Haddad and Haj (2018), increasing lexico‐phonological distance erodes the precision and stability of phonological representations, with downstream consequences for lexical access efficiency. Identical words sit at the favourable end of this continuum: shared across SpA and StA, they draw on representations already consolidated through spoken language experience, making them more salient and more readily accessible. Unique words occupy the opposite pole; their acquisition requires building representations from scratch, without the scaffolding that spoken familiarity provides and the result is encodings that are less robust and harder to retrieve. The present findings map onto this hierarchy with some fidelity, pointing toward representational quality, rather than frequency of exposure, as the more proximal constraint on vocabulary performance.

Developmental improvements were observed across grades for both groups, reflecting increased exposure and linguistic experience. However, performance on unique words remained consistently lower than on identical and cognate words at all grade levels, indicating that the representational constraints imposed by linguistic distance are persistent and not fully resolved through development. The absence of a significant Grade × Word Type interaction carries theoretical weight in its own right. It indicates that the hierarchical ordering of word types (identical above cognate above unique) did not shift, attenuate or intensify across the six elementary school years examined. This developmental stability is not a null finding in the trivial sense; it constitutes positive evidence that the representational constraints associated with lexico‐phonological distance are not resolved by accumulated exposure to Standard Arabic over time. This pattern converges with the findings of Saiegh‐Haddad and Haj (2018), who documented the same graded representational hierarchy across senior kindergarten through Grade 6 and interpreted its persistence as evidence that linguistic distance imposes a structural constraint on phonological encoding that schooling alone does not overcome. The present study extends that observation to vocabulary acquisition specifically, suggesting that the same mechanism—distance‐driven degradation of representational quality operates with comparable persistence across both phonological and lexical domains.

Both groups showed the distance gradient, but the group gap was substantially larger for unique words than for identical ones. Lexico‐phonological distance, the pattern suggests, does not operate in isolation; it interacts with phonological processing capacity. Children with LD exhibit well‐documented weaknesses in phonological encoding, storage and retrieval (Elbro and Jensen 2005; Gathercole and Baddeley 1990; Mengisidou and Marshall 2019) and these weaknesses already compromise lexical representation under ordinary conditions. When distance‐related degradation is added to that baseline vulnerability, the performance gap widens beyond what either factor would predict alone.

This interpretation finds support in earlier work. Saiegh‐Haddad and Ghawi‐Dakwar (2017) showed that children with DLD produce phonologically distant forms less accurately and Ghawi‐Dakwar and Saiegh‐Haddad (2025) found that distance constrains word learning more strongly when phonological processing is impaired. The present study examines LD rather than DLD, but the mechanism is the same: representational quality suffers under high linguistic distance and suffers more when the processing system is already strained. Developmental gains were more evident among TD children; in the LD group, progress was more limited, particularly for the word types furthest from SpA.

A consistent modality effect was also observed, with receptive vocabulary outperforming expressive vocabulary across groups. This asymmetry was more pronounced in children with LD and for lexically distant words. From a descriptive perspective, this reflects greater difficulty in expressive tasks; however, previous research indicates that this pattern is driven by the higher demands of phonological retrieval. Within a representational quality framework, underspecified phonological representation (particularly for distant StA forms) disproportionately impairs expressive performance. This account is consistent with findings from Tarabi‐Juma'a et al. (2026), who reported stronger distance effects in production than in comprehension.

The findings point in a consistent direction. Linguistic distance constrains vocabulary development through its effects on representational quality, a conclusion that sits comfortably within the MAWRID model (Saiegh‐Haddad and Haj 2018), which treats diglossia as a source of distance that limits how fully and how accessibly lexical representations consolidate. What the present study adds is a developmental and clinical dimension: the representational consequences of distance do not resolve across the elementary school years and they are measurably larger in children with learning difficulties than in their typically developing peers.

The findings also align with the Lexical Quality Hypothesis (Perfetti 2007), which posits that the efficiency of lexical access depends on the precision and integration of phonological representations. In the context of Arabic diglossia, lexico‐phonological distance systematically limits this precision, thereby constraining vocabulary acquisition.

The findings have both theoretical and educational implications. Theoretically, they reinforce the role of representational quality as the mechanism linking linguistic distance to vocabulary performance. Educationally, they suggest that instruction may benefit from addressing the representational demands imposed by diglossia. Evidence for this approach is provided by Saiegh‐Haddad (2023) and Haj, Ghawi‐Dakwar, et al. (2026), who demonstrated that structured, distance‐sensitive interventions can improve both vocabulary and the quality of lexical‐phonological representations. The present findings extend this evidence by showing that similar constraints operate in naturalistic development, particularly among children with learning difficulties.

Several limitations warrant acknowledgment. The study did not include measures of executive function or other cognitive variables that may independently shape vocabulary development and expressive responses were scored for accuracy without systematic analysis of error patterns, information that could have illuminated the nature of retrieval failures more precisely. The LD classification presents a further constraint: group membership was determined from school records rather than a uniform diagnostic protocol, meaning the LD sample almost certainly included children with heterogeneous profiles whose difficulties may rest on different underlying mechanisms. Whether the effects documented here differ across LD subtypes (dyslexia, poor comprehension or other profiles) remains an open question. So too does the mediating role of representational quality: whether it accounts for the relationship between linguistic distance and vocabulary performance, rather than merely co‐occurring with it, is a question the present design was not built to answer.

These limitations notwithstanding, the present study makes a coherent contribution. It demonstrates that lexico‐phonological distance constrains vocabulary acquisition in Arabic diglossia through a mechanism—the graded degradation of representational quality—that is developmentally persistent, operating with comparable force across all six elementary school years examined. It further establishes that this constraint interacts with learning status in theoretically predicted ways: children with LD, whose phonological encoding and representational resources are already compromised, bear a disproportionate burden at precisely the points of greatest linguistic distance. These findings extend the MAWRID account from phonological processing and reading into the domain of vocabulary acquisition and identify lexico‐phonological distance as a structural variable that shapes lexical development in Arabic with both breadth and persistence.

Funding

The author has nothing to report.

Ethics Statement

All procedures performed in this study involving human participants were in accordance with the ethical standards of the institutional and national research committee and with the comparable ethical standards. The study was approved by the ethics committee at the Arab Academic College in Haifa, Israel.

This article does not contain any studies with animals performed by any of the authors.

Consent

Informed consent was obtained from all individual participants included in the study.

Conflicts of Interest

The author declares no conflicts of interest.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author Dr. Abeer Asli‐Badarneh, upon request.