1. Introduction
Babies start to understand their first words at around 6–9 months (Bergelson & Swingly, Reference Bergelson and Swingley2012), before they are able to produce words – at around one year of age (Fenson et al., Reference Fenson, Dale, Reznick, Bates, Thal and Pethick1994). By the end of the first year, infants can understand between 10 and 68 words, with comprehension expanding with age (Harris et al., Reference Harris, Yeeles, Chasin and Oakley1995). Research has shown that language comprehension predicts later language development (Fisher, Reference Fisher2017), as well as language skills such as phonological awareness and reading comprehension (Duff et al., Reference Duff, Reen, Plunkett and Nation2015), the development of cognitive abilities (Marchman & Fernald, Reference Marchman and Fernald2008), and additional academic achievements (Ferrer et al., Reference Ferrer, Shaywitz, Holahan, Marchione, Michaels and Shaywitz2015).
In different languages, words are classified into syntactic categories (or parts of speech). The distinction between nouns and verbs is probably the oldest distinction in linguistic categorization and remains central today. Definitions of these categories rely primarily on semantic and syntactic criteria. The semantic criterion appears more intuitive and potentially universal: nouns typically refer to concrete objects with referential functions, while verbs tend to designate events, actions, states, mental actions, or attitudes (Bassano, Reference Bassano2000). However, nouns sometimes also denote states, events, or even actions, and some words can be either nouns or verbs depending on sentence structure (e.g., “he is thinking” vs. “this is good thinking”). Thus, precise category definitions require syntactic criteria, where words from the same category occupy specific positions in sentences and/or feature distinctive morphological markers. This approach is complicated by cross-linguistic variation in language structure (Jelinek, Reference Jelinek1981).
There is considerable interest in the acquisition of nouns and verbs, particularly regarding cross-linguistic developmental variation. This interest stems partly from how these categories help investigate universal versus language-specific processes in early language development, as well as connections between lexical and syntactic development (Bassano, Reference Bassano2000).
The question of which syntactic category develops first has generated significant research interest, especially regarding noun and verb acquisition (Waxman et al., Reference Waxman, Fu, Arunachalam, Leddon, Geraghty and Song2013). Many studies show that in early language acquisition, children tend to produce nouns first, before expanding to other syntactic categories as vocabulary grows (Bates et al., Reference Bates, Marchman, Thal, Fenson, Dale, Reznick and Hartung1994). This observation led to the coining of the term noun bias.
Whether noun bias is universal in early language development is still debated. One perspective, the universalist approach, argues that noun bias results from an innate tendency linked to children’s natural ability to learn names of objects, people, and animals (Gentner, Reference Gentner and Kuczaj1982; Gleitman et al., Reference Gleitman, Cassidy, Nappa, Papafragou and Trueswell2005; Golinkoff & Hirsh-Pasek, Reference Golinkoff and Hirsh-Pasek2008). The alternative perspective claims that noun bias varies according to language characteristics, with different languages showing preferences for nouns or verbs based on their syntactic and cultural structures (Tardif et al., Reference Tardif, Gelman and Xu1999; Bornstein et al., Reference Bornstein, Cote, Maital, Painter, Park, Pascual and Vyt2004).
Several explanations have been suggested to support the possibility of a universal noun bias. One major explanation concerns conceptual differences between nouns and verbs – nouns, which universally tend to designate objects, appear more stable and fixed conceptually and perceptually, making them easier to acquire than verbs, which designate actions and events requiring understanding of relationships between objects and more complex processing (Gentner, Reference Gentner and Kuczaj1982). Verbs present greater difficulty because they cannot be learned through direct observation or pointing alone, but require understanding how language marks movement and action (Mandler, Reference Mandler2022). Additionally, nouns tend toward greater concreteness than verbs (Gentner, Reference Gentner, Hirsh-Pasek and Golinkoff2006; McDonough et al., Reference McDonough, Song, Hirsh‐Pasek, Golinkoff and Lannon2011).
Concrete objects and names (e.g., people, animals) may be acquired earlier due to clear perceptual mapping to individual concept categories, while words perceived as non-individual concepts (actions, prepositions, spatial words) require broader linguistic knowledge, delaying their acquisition (Gentner & Boroditsky, Reference Gentner, Boroditsky, Bowerman and Levinson2001). Another perspective attributes universal noun bias to differences in linguistic knowledge required for different syntactic categories – understanding a specific verb, for instance, depends on knowing the objects or persons to which it refers. Children may need to establish a repertoire of nouns, highlighting their sentential roles, before acquiring other syntactic categories like verbs and adjectives (Waxman & Lidz, Reference Waxman, Lidz, Kuhn and Siegler2006).
Joint attention between children and adults provides another explanation. When adults point to objects and label them, children easily connect words to concrete objects. Connecting words to actions presents greater challenges since actions are dynamic, changing, and often brief (Mandler, Reference Mandler, Marchetti, Benedetti and Alharbi2015; Tomasello & Todd, Reference Tomasello and Todd1983). During interactions, children more easily connect words to static environmental objects than to rapid actions requiring immediate contextual understanding (Mandler, Reference Mandler2022).
However, some studies have also challenged the universality of noun bias, suggesting that syntactic category development depends on language structure. According to this view, some languages show noun bias while others favour verbs. Evidence for a noun bias appears in languages like English and Italian (Caselli et al., Reference Caselli, Bates, Casadio, Fenson, Fenson, Sanderl and Weir1995), but comparisons show variations in strength – both English and Korean first words contain more nouns than verbs, but Korean speakers acquire more verbs than English speakers (Kim et al., Reference Kim, McGregor and Thompson2000). Other studies found no noun bias in German, Chinese, and Korean (Bloom et al., Reference Bloom, Tinker and Margulis1993; Choi & Gopnik, Reference Choi and Gopnik1995; Kauschke & Hofmeister, Reference Kauschke and Hofmeister2002; Tardif, Reference Tardif1996), arguing that initial lexicon composition reflects language structure, culture, and child-directed speech characteristics rather than conceptual preferences. In some languages, noun bias emerges only in advanced lexical-development stages – one study showed that with vocabularies under 50 words, children prefer nouns over adjectives and other parts of speech but not over verbs, with a significant noun bias emerging only after acquiring the first 50 words (Bornstein et al., Reference Bornstein, Cote, Maital, Painter, Park, Pascual and Vyt2004).
While most studies focus on syntactic category acquisition in language production, relatively few examine syntactic categories in comprehension. Since comprehension develops before production (Harris et al., Reference Harris, Yeeles, Chasin and Oakley1995; Fenson et al., Reference Fenson, Dale, Reznick, Bates, Thal and Pethick1994), the ability to understand words from different syntactic categories might show less or more category preference. A comparative study examining noun, adjective, verb, and function word acquisition in comprehension and production across languages in 8–18 -month-old children found that in most languages, percentages of nouns, adjectives, and verbs in receptive vocabulary were similar. However, some languages showed noun bias while others showed relative predicate bias (Frank et al., Reference Frank, Braginsky, Yurovsky and Marchman2021).
Studies on English show noun preference in both comprehension and production, with children understanding some nouns from 6–9 months (Bergelson & Swingly, Reference Bergelson and Swingley2012) and some verbs only from 10–13 months (Bergelson & Swingly, Reference Bergelson and Swingley2013). Research on German found that infants understand some verbs around 10 months, with some verbs identified even earlier (Nomikou et al., Reference Nomikou, Rohlfing, Cimiano and Mandler2019). Based on these findings, noun bias appears not necessarily universal but dependent on language characteristics, culture, and developmental contexts.
It is important to note that comparisons between syntactic categories typically rely on adult speech categories. Infants and children may not understand or use words as adults do, creating challenges in early-stage word categorization (Hagihara & Sakagami, Reference Hagihara and Sakagami2020; Bloom et al., Reference Bloom, Tinker and Margulis1993). Infants may produce words with different intended meanings – Bates et al., (Reference Bates, Marchman, Thal, Fenson, Dale, Reznick and Hartung1994) described a child saying “hot” (adjective) to mean “oven” (noun), or “up” (adverb or preposition) when requesting to be picked up (verb). Other research showed that words appearing as nouns in children’s lexicons up to 21 months often describe undifferentiated whole events strongly influenced by related routine actions. For example, “shoes” might describe the entire shoe-wearing event or even going for a walk. With age and vocabulary growth, words gradually start to resemble adult categories (Hagihara & Sakagami, Reference Hagihara and Sakagami2020).
Studies arguing against a universal noun bias have examined two factors potentially affecting lexical development and syntactic category acquisition: (1) language structure and (2) parental input. The first suggests that a language’s structure explains lexicon composition and preference for nouns, verbs, or adjectives, accounting for similarities across languages from different cultures. Researchers argue that typological factors (morphology, word order, etc.) influence early word acquisition (Tardif et al., Reference Tardif, Shatz and Naigles1997). For example, Arabic verbs have more inflections than noun inflections, potentially making them more complex to acquire.
The second factor suggested is cultural and social differences between societies or within a society, manifested in parents’ linguistic behaviours with their children (Chai et al., Reference Chai, Low, Wong, Onnis and Mayor2021). For example, parents’ tendency to point to and name objects (rather than actions) may lead to a noun bias. Researchers argue that reference patterns vary culturally, affecting lexicon composition – individualistic cultures emphasize objects, associated with nouns, while collectivistic cultures emphasize connections between entities (including people), typically described by verbs (Nisbett & Miyamoto, Reference Nisbett and Miyamoto2005).
This explanation accounts for lexicon composition differences between languages or within the same language spoken by people with different sociological backgrounds. Tardif et al., (Reference Tardif, Gelman and Xu1999) found that Mandarin-speaking mothers used more verbs than English-speaking mothers (who used more nouns), with Mandarin-speaking toddlers also using more verbs and English-speaking toddlers using more nouns. Research on bilingual children found that English receptive vocabulary in Mandarin-English speakers included more verbs than that of Malay–English speakers, supporting connections between culture and receptive vocabulary, and suggesting that lexicon composition is influenced by factors beyond the characteristics of English (Chai et al., Reference Chai, Low, Wong, Onnis and Mayor2021). Casillas et al., (Reference Casillas, Foushee, Méndez Girón, Polian and Brown2024) found some noun bias in Tzeltal spontaneous production, but very large variance between children.
The current study examined noun bias in the Palestinian-Arabic receptive lexicon during early language development, considering two main possibilities for this preference – language structure and maternal input – to explain lexicon composition and syntactic category preferences in Palestinian Arabic in Israel.
Arabic is a Semitic language characterized by roots and morphological patterns (Holes, Reference Holes1995; Saiegh-Haddad & Henkin-Roitfarb, Reference Saiegh-Haddad, Henkin-Roitfarb, Saiegh-Haddad and Joshi2014). As a Semitic language, Arabic acquisition includes unique features absent in most other languages. For example, Arabic has dual noun forms with unique endings, while English uses the same endings for all plural forms.
Arabic morphology is characterized by lexical items based on mostly 3–4 consonant roots combined with patterns to create verbs and nouns. Root–pattern combinations create uniquely meaningful items associated with specific syntactic categories (Saiegh-Haddad & Henkin-Roitfarb, Reference Saiegh-Haddad, Henkin-Roitfarb, Saiegh-Haddad and Joshi2014). Arabic has eight different verbal patterns, which are conjugated by gender, tense, and number. Similarly, patterns for creating nouns undergo morphological inflections for number and gender.
Arabic also features diglossia – regional dialects for everyday informal speech alongside standard language for writing and formal communication, also known as Modern Standard Arabic (MSA). Children encounter MSA early, through television, prayer ceremonies, books, etc. (Saiegh-Haddad & Henkin-Roitfarb, Reference Saiegh-Haddad, Henkin-Roitfarb, Saiegh-Haddad and Joshi2014). Differences between these language forms are reflected at the phonological, morphological, semantic, and syntactic levels (Saiegh-Haddad, Reference Saiegh-Haddad2003), potentially affecting lexicon development.
Dialectical Arabic varies by geographical area, religion, and socio-economic status, creating linguistic variation among Arabic-speaking communities (Holes, Reference Holes1995). In Israel, the Palestinian dialect prevails among Palestinian Arab populations (Henkin, Reference Henkin2000), with many sub-dialects differing in phonological, lexical, and even morphological characteristics. Israel has three main Palestinian Arabic dialects (southern, central, northern) and a Druze dialect with distinctive phonological characteristics. However, this division can be misleading, as large differences also exist between Rural and Urban dialects, and even between different villages or neighborhoods. Dialect differences stem from lifestyle, residence, religion, social level, education, gender, etc. (Jarrar et al., Reference Jarrar, Habash, Alrimawi, Akra and Zalmout2017).
The current study focused on two dialects: northern triangle and northern. Like other dialects, these differ phonologically – for example, \k\ is pronounced as \ts’\ in the triangle. Additional differences appear in morphology (gender inflections) and the lexicon (Palva, Reference Palva1984).
Arabic language structure – featuring roots, derivation, and varying noun, verb, and adjective forms – may affect infants’ comprehension when acquiring different syntactic categories, with verb morphological complexity posing special challenges. Studies on Arabic worldwide have examined differences in verb, noun, and adjective acquisition primarily regarding production. Palestinian Arabic appears to show a noun bias (Saiegh-Haddad & Henkin-Roitfarb, Reference Saiegh-Haddad, Henkin-Roitfarb, Saiegh-Haddad and Joshi2014; Rosenhouse, Reference Rosenhouse2000).
The current study investigated whether factors beyond the structure of Arabic relate to the receptive lexicon composition in early language development, particularly exploring connections between culture, the input children receive, and noun bias. Research describes the Arab culture in Israel in terms of authority and extended family structure, characterizing it as collectivistic and emphasizing interpersonal relationship importance (Shoshana & Shchada, Reference Shoshana and Shchada2018). The Arab society in Israel shows great internal diversity, comprising several subgroups (including different religious groups), potentially affecting child-rearing approaches. However, almost no research addresses connections between culture and lexicon composition in early Arabic language development.
As noted earlier, parents’ ways of speaking to their infants influence language development, potentially causing differences between languages or within the same language. Rosenhouse (Reference Rosenhouse2000) attempted to explain differences in Arabic development across dialects (Palestinian Arabic in Israel, Egyptian Arabic, MSA, Jordanian Arabic), referencing social, environmental, and cultural differences. Differences appeared between dialectical Palestinian and Egyptian Arabic in language production, with some characteristics acquired earlier in Palestinian Arabic (certain phonemes, syntactic characteristics, morphological categories of colours and numbers). These differences stemmed from dialect structure and cultural environment, demonstrating culture’s relation to language acquisition even within the same language across dialects.
Additional parental input aspects studied in Arabic that may be related to language development and lexicon composition include shared reading. Reading to and with children is an important interaction for language development and learning (Phillips, Reference Phillips2000; Sénéchal, Reference Sénéchal, McCardle and Berninger2010). When reading in Arabic, parents encounter challenges stemming from diglossia – desiring to read books from an early age while recognizing that books usually use MSA. According to Rosenhouse, (Reference Rosenhouse2000), some mothers read stories verbatim so children will eventually understand MSA, while others “translate” words into dialectal Arabic. The effects on children’s understanding regarding vocabulary composition remain unknown.
In summary, Arabic language structure may promote a noun bias, but the society’s collectivistic values may moderate this preference if mothers use more verbs, similar to other collectivistic cultures. Additionally, societal heterogeneity allows examination of maternal input differences related to lexical development within a diverse society sharing a common language.
The current study thus continues the discussion of noun bias, focusing on Palestinian Arabic in Israel in the receptive vocabulary of infants. Few studies examined the receptive lexicon composition of Arabic-speaking children, particularly under age two. We also examine relationships between maternal input and receptive lexicon composition in Palestinian Arabic-speaking toddlers, as well as between parents’ demographic characteristics and cultural values and maternal input.
We ask: (1) Does the Palestinian Arabic receptive vocabulary of infants show a noun bias? (2) Is there correspondence between the usage of syntactic categories in maternal input and the noun bias in children’s receptive lexicon? (3) Is there a relationship between mothers’ cultural values and relative verb/noun usage in maternal input? (4) Is there a relationship between maternal education level and relative verb/noun in maternal input?
We predicted that: (1) Palestinian Arabic would show noun bias in receptive vocabulary, based on previous studies showing noun production bias in early language development (Al-Jabali, Reference Al-Jabali2003; Hashoul-Essa, Reference Hashoul-Essa2017); (2) A correlation would exist between noun and verb usage in maternal input and noun bias in infants; (3) A relationship would exist between the maternal cultural values questionnaire and verb/noun usage in the input, with collectivist mothers using relatively more verbs; (4) We explore the relationship between maternal education and relative verb/noun usage in the input.
2. Method
2.1. Public material and data
All materials, data and analysis code are available through this OSF link: https://osf.io/vwxac/?view_only=4328f162828b4f05a177eaccbcb955f2.
2.2. Participants
The study included 31 Arabic-speaking infants aged 16–24 months (mean age = 20.81 months, SD = 2.12; 15 girls), speakers of two dialects: the Northern Triangle area dialect and the Rural Northern dialect. Most participants lived in the Northern Triangle area, including in Jatt, Baqa al-Gharbiya, Kafr Qara, Ar′ara, and Umm al-Fahm except for one family from the Northern region speaking the Rural Northern dialect. Each infant was tested in their mother’s dialect. All mothers spoke the local dialect of their residential area.
All toddlers were typically developing with no hearing or vision impairments, and were monolingual. Regarding maternal education, one mother completed high school, six had professional education, 17 had a bachelor’s degree, and seven had a master’s degree or higher.
Some participants were recruited through daycare providers, with requests sent to parents to sign consent forms. Other participants were recruited through social media. In total, 90 families from the Triangle and Northern regions consented and filled in a demographic questionnaire, but some families were not tested due to our exclusion criteria (dialect, developmental concerns, age, area of residence, and maternal dialect different from residential area). 33 families completed all research stages, but two were excluded from data analysis: one due to technical problems, and another was excluded since the mother helped with the CCT, contrary to instructions.
3. Tools
3.1. CCT for the Palestinian Arabic Dialect
The Computerized Comprehension Task (CCT; Friend & Keplinger, Reference Friend and Keplinger2003) is designed for direct assessment of language comprehension abilities in infants aged 16–24 months. The test uses a touchscreen to engage infants, effectively maintain infants’ attention, and provide more reliable comprehension assessment compared to other methods (Friend & Keplinger, Reference Friend and Keplinger2008; Friend et al., Reference Friend, Schmitt and Simpson2012).
The test was adapted to Arabic in the Palestinian dialect in two recent studies: one for the Northern Triangle area dialect (Jammal-Agbaria, Reference Jammal-Agbaria2024; validity ranging from 0.253 to 0.857 depending on administration form, reliability = 0.565) and another for the Rural Northern dialect (Zuabi, Reference Zuabi2025; validity and reliability were not examined). The test includes 41 word pairs (nouns, verbs, adjectives). At each trial, toddlers see pairs of pictures and hear instructions to touch one picture. The test begins a training phase comprising of 4 trials, to introduce the test and ensure understanding of instructions to touch the appropriate pictures after hearing the prompt. During training, mothers or research assistants could help, but were instructed to stop assistance once testing began. Verbs were presented in the present tense, simple form, and masculine singular form.
3.2. Demographic questionnaire
The demographic questionnaire included questions regarding toddlers’ age and gender, parents’ education level and profession, number of children in the family, child’s birth order, birth week and weight, exposure to additional languages, parents’ native language and dialect, and suspected developmental issues or hearing/vision problems.
3.3. Cultural values questionnaire
This questionnaire examined families’ perceptions of child individuality and how they encourage or limit independent personality development. Seven statements were chosen for the questionnaire, each rated on a 1–5 scale (1 = “strongly disagree,” 5 = “strongly agree”). Statement selection was based on previous research examining family values and child-rearing approaches across cultures, with a focus on the Arab society in Israel (Shoshana & Shchada, Reference Shoshana and Shchada2018; Camaioni et al., Reference Camaioni, Longobardi, Venuti and Bornstein1998; Bornstein et al., Reference Bornstein, Tamis‐LeMonda, Tal, Ludemann, Toda, Rahn and Vardi1992). For analysis, some statements were reverse-coded to create a uniform continuum distinguishing between mothers with more collectivist perceptions and those encouraging greater openness and individuality.
The statements were: 1) “Your child has a separate and independent personality”; 2) “Your child is part of the family and society and depends on them”; 3) “You encourage your child to be independent and make their own decisions, regardless of your decisions and those of society”; 4) “The child should accept the decisions of the family and environment for them”; 5) “You feel connected to the importance of society and its perceptions regarding child-rearing (e.g., respect and obedience to adults)”; 6) “You prefer to encourage the child for personal successes that are not influenced by society”; 7) “You prefer that the family and society be pleased with the child’s behaviour and successes.”
3.4. Picture narratives
Three picture narratives from Wang et al., (Reference Wang, Hoehl, Sommer, Wertz, Schreiner, Kühn-Popp, Cowell, Frank, Hamlin, Scott and Kline2023) were used. These colourful illustrations depict people engaged in various activities, with objects in the background or in use by characters. For example, one picture shows a child sitting on a chair beside a table reading a book, with a girl nearby building a tower of colourful blocks, and a window, bookcase, plant, and carpet visible in the background. Pictures were displayed on participants’ home touch devices using PsychPy (Pierce et al., Reference Peirce, Gray, Simpson, MacAskill, Höchenberger, Sogo and Lindeløv2019) and Pavlovia (Morys-Carter, Reference Morys-Carter2021). Mothers read detailed instructions: “Now three pictures will be presented. You are asked to show the toddler each picture separately and describe to them what you see in your own words. At the end of describing each picture, press the blue arrow to move to the next picture.” Each picture narrative depicted actions describable using verbs like “reading,” “building,” “riding,” and a variety of objects describable using nouns like “child,” “blocks,” “bicycle” (see Figure 1).
Figure 1. Illustrations from Wang et al., Reference Wang, Hoehl, Sommer, Wertz, Schreiner, Kühn-Popp, Cowell, Frank, Hamlin, Scott and Kline2023 used in the study. These can be described by using both nouns (e.g., “boy,” “sun,” “bicycle”) and verbs (e.g., “riding,” “spinning,” “building”).
Coding. Narratives were transcribed by native speakers of participants’ dialect. From transcripts we recorded the number of noun types (Type) and the total number of nouns (Token). Verbs were counted similarly. Each inflection was included in total verb tokens but counted only once for type. For example, when mothers used “play” in different inflections – “playing” (feminine), “playing” (plural), “played” – three tokens were counted, but the verb was recorded only once as a type.
4. Procedure
The research was conducted in three stages. First, parents received consent forms and demographic questionnaires. Then, the CCT was administered using touch devices by researchers in daycare centres or by parents at home. Third, mothers received a link to the picture narratives to describe to their children in their own words at home. The research was approved by the Ethics Committee of the School of Psychological Sciences, University of Haifa, Israel.
4.1. Detailed research stages
First stage – consent and demographic questionnaire. Consent forms and questionnaires were distributed via Google Forms. For daycare-recruited participants, links were forwarded through the daycare staff to parents. Social media-recruited participants received questionnaires directly via phone messages.
Second stage – CCT test administration. In daycares, researchers sat with each infant separately in quiet daycare areas with touch devices displaying test pictures. Some toddlers required acquaintance games to encourage cooperation. At home, mothers were asked to find a quiet time and place for the infant to participate. In both the daycare and home settings, test administration began with training trials where picture pairs were shown to children, and with maternal or researcher mediation, they were asked to touch the correct pictures after hearing audio input. During training, mothers were asked to help toddlers understand requirements. Training could be repeated once if children failed to understand the task. After training, testing began with 41 picture pairs. In each trial, infants were asked to touch the correct image according to the prompt without the researcher or maternal mediation. Mothers were instructed to help infants touch only if intentions regarding which image to touch were clear (e.g., clear pointing), as in the original English CCT.
Third stage – Picture narrative description. Three picture narratives were sent to mothers’ home touch devices. Mothers were informed that the application recorded their voices during descriptions for later coding.
5. Results
Data analysis was done with R software (RStudio Team, 2020). We first present descriptive statistics of demographic variables and dependent and independent variables (Tables 1 and 2 and Figures 2 and 3). Then we present inferential statistics to test the research hypotheses.
Table 1. Descriptive statistics of toddlers’ responses in the CCT test by different categories (nouns, verbs, adjectives) and noun bias. Note. Noun bias was calculated as the sum of correct responses to nouns divided by the number of nouns in the test, divided by the sum of correct responses to nouns and verbs divided by the number of nouns and verbs in the test, for each child
Table 2. Descriptive statistics of transcription results from picture narrative input and values questionnaire administered to mothers
Abbreviations: TTR, type–token ratio.
Figure 2. Distribution of current noun and verb responses on the CCT.
Note: Noun success = proportion of correct answers for nouns. Verb success = proportion of correct answers for verbs. The boxes represent the interquartile range for the proportion of correct answers for nouns (blue colour) and verbs (green colour). The black line inside each box is the median. The lines above and below the boxes represent the minimum and maximum values and show the total range of correct answers for each category. The success rate for nouns is higher than for verbs.
Figure 3. Distribution nouns and verbs (types, tokens, and TTR) in maternal input.
Note: The graph presents the use of nouns and verbs in mothers’ input, distinguishing between types, tokens, and TTR. The six boxes represent the interquartile range for each category. The black line inside each box represents the median. The lines above and below the boxes represent the minimum and maximum values of the data (without outliers) and show the total range for each category. The points outside mark extreme values or outliers. Nouns are more common in both types and tokens than verbs, and the TTR for verbs is higher than for nouns.
5.2. Inferential statistics
Hypothesis 1: There will be a noun bias in Palestinian-Arabic-learning infants’ receptive vocabulary.
To test the first hypothesis, we calculated for each infant the ratio of correct responses on noun pairs divided by the ratio of correct responses on verb and noun pairs taken together – a number greater than 1 indicates a noun bias. To determine significance, we conducted a single sample t-test. Results showed significant noun preference in the CCT (t(30) = 5.223, p < .0001).
We also examined whether noun bias existed in mothers’ picture narrative descriptions input (for which we did not specify a-priori hypotheses). Results showed significant noun bias in the total number of words in maternal input (t(30) = 13.736, p < .0001), in the number of types of words from different categories (t(30) = 17.198, p < .0001, see Figure 4), and the reverse in terms of TTR (t(30) = −2.847, p < .008) – verbs TTR is higher than noun TTR, and can be expected given the complexity of verb conjugations and inflections in Arabic.
Hypothesis 2: There will be a correlation between the use of nouns in maternal input and the noun bias in infants’ receptive lexicon.
Figure 4. Proportion of correct touches on noun trials by proportion of success on noun and verb trials.
Note: Dashed red line marks equality of proportion between the noun trials and noun + verb trial, such that any dot above this line is an infant with a noun bias. 26 out of 31 infants show a noun bias. Dot size marks infants’ age.
To test the second hypothesis, we conducted three linear regressions. The dependent variable in all cases was the noun bias in the CCT.
In the first regression, independent variables were the number of verb types used by mothers in picture narrative descriptions and the number of noun types, while statistically controlling for infants’ age. The results showed that these three factors together explained 14% of the variance (R2 = .145, F(3, 27) = 1.526, p = .23). We found a significant negative correlation between the number of verb types in the input and noun bias in the CCT (β = −0.677, p = .048), with more verb types used by mothers associated with less noun bias. No significant correlation was found between the number of noun types and noun bias on the CCT (β = 0.208, p = .143), and no significant age effect was found (β = 0.279, p = .594).
In the second regression, independent variables were the number of verb tokens used by mothers in picture narrative descriptions and the total noun tokens, while statistically controlling for infants’ age. Results showed these three factors together explained 17% of the variance (R2 = .17, F(3, 27) = 1.846, p = .163). We found a significant negative correlation between verb tokens and noun bias on the CCT (β = −0.308, p = .03), with more verbs used by mothers associated with less noun bias. A marginally significant positive correlation was found between the noun tokens and noun bias in the CCT (β = 0.14, p = .067), and no significant age effect was found (β = 0.258, p = .608).
In the third regression, independent variables were the TTR of verbs used by mothers in picture narrative descriptions and the TTR of nouns, while statistically controlling for infants’ age. Results showed these three factors together explained only 1% of the variance (R2 = .013, F(3, 27) = 0.1158, p = .95). We found no significant relationship between TTR of verbs in maternal output (β = −0.515, p = .767) or between TTR of nouns in maternal output and noun bias on the CCT (β = .307, p = .876), and no significant age effect (β = 0.262, p = .641).
Hypotheses 3: There will be a relationship between mothers’ degree of collectivism in the cultural values questionnaire and noun bias in their input to infants; and Exploratory Question 4: Is there a relationship between maternal education and noun bias in their input to infants?
To examine whether maternal education level and degree of collectivism were related to noun bias in maternal input, we conducted two linear regressions. The dependent variable in the first regression was noun bias in terms of tokens in the input, and in the second regression, noun bias in terms of types in the input. Independent variables in both were mothers’ values and education.
In the first regression, results showed the three factors together explained 9% of the variance (R2 = .092, F(2, 28) = 1.416, p = .26). We found no significant relationship between mothers’ values and token-noun bias in the input (β = .002, p = .589), nor between maternal education and noun-token bias (β = .03, p = .118).
In the second regression, results showed the three factors together explained only 3% of the variance (R2 = .035, F(2, 28) = 0.514, p = .604). We found no significant relationship between mothers’ values and type-noun bias in the input (β = −0.001, p = .853), nor between maternal education and type-noun bias in the input (β = .018, p = .332).
6. Discussion
The purpose of the current study was to examine the relationship between maternal input, the strength of noun bias, and cultural values, in the receptive lexicon of Palestinian-Arabic-learning infants. Specifically, the study sought to investigate whether noun bias exists in Palestinian Arabic comprehension during the early language development stages, whether correlations exist between noun and verb usage in maternal input and noun bias in children’s receptive lexicon, and whether collectivist values and education level are related to mothers’ choice of nouns and verbs in speech with their children.
Research indicates a preference for nouns over other categories such as verbs and adjectives in the lexicon during the initial language-development stages across different languages (Bates et al., Reference Bates, Marchman, Thal, Fenson, Dale, Reznick and Hartung1994), with some studies suggesting this preference is universal (Gentner, Reference Gentner and Kuczaj1982; Golinkoff & Hirsh-Pasek, Reference Golinkoff and Hirsh-Pasek2008). However, more recent studies oppose the claim that noun bias is universal, arguing that it depends on linguistic structure and culture (Bornstein et al., Reference Bornstein, Cote, Maital, Painter, Park, Pascual and Vyt2004). Few studies have examined noun bias in Palestinian Arabic, primarily focusing on the expressive vocabulary. The Palestinian society in Israel presents an interesting test case for analysis, since it is defined on the one hand as a collectivist and conservative culture – albeit with very heterogeneous cultural and religious groups, and on the other hand, the morphological structure of Arabic should make verbs much harder to learn than nouns, as they have many more inflections and conjugations than nouns.
The first hypothesis addressed the existence of a noun bias in the receptive lexicon of infants aged 16–24 months. We hypothesized that Palestinian Arabic would show a noun bias in the early lexicon based on previous studies showing a noun bias in the expressive lexicon (Saiegh-Haddad & Henkin-Roitfarb, Reference Saiegh-Haddad, Henkin-Roitfarb, Saiegh-Haddad and Joshi2014; Rosenhouse, Reference Rosenhouse2000). In line with our hypothesis, we found significant evidence of noun bias in the receptive lexicon as measured by the CCT, with 26 out of 31 infants scoring proportionally higher on noun trials and only 4 showing a verb bias.
This finding aligns with the literature describing several reasons for noun bias in early language acquisition. Researchers have pointed out that noun concepts describing objects are more perceptually stable and concrete, making them easier to acquire compared to verbs that describe actions and require understanding complex relationships between objects (Gentner, Reference Gentner and Kuczaj1982). Verbs present greater challenges because they cannot be learned through observation alone and require understanding linguistic structures (Mandler, Reference Mandler2022). Another possible explanation for a noun bias is the linguistic structure of Arabic itself, which is morphologically complex especially regarding verbs, with numerous verb patterns, inflection, and conjugation forms that could make nouns comparatively easier to acquire.
An important question remains regarding how the strength of noun bias in Palestinian Arabic compares to other languages like English (with a simpler verb morphology) and Hebrew (with a similarly complex verb morphology). Is this bias stronger due to Arabic’s linguistic complexity, or perhaps weaker due to cultural factors beyond linguistic structure? Currently, such cross-linguistic comparative studies are lacking. To understand whether the observed noun bias stems primarily from language structure or is also influenced by parents’ verbal behaviour, we examined possible relationships between cultural values, maternal input, and noun bias in the early receptive lexicon in Arabic.
The second hypothesis thus referred to maternal input and its relationship with infants’ receptive lexicon. We expected to find correlations between maternal input in picture narrative descriptions and noun bias in children’s lexicon, based on studies examining relationships between maternal input in different situations and early lexicon composition (Chai et al., Reference Chai, Low, Wong, Onnis and Mayor2021).
To examine this hypothesis, we tested three main relationships: First, the relationship between number of noun types and verb types in maternal input and noun bias in infants’ CCT performance; Second, the relationship between noun and verb tokens used by mothers in picture narratives and noun bias in the CCT; Third, the relationship between Type–Token Ratio (TTR) of verbs and nouns in picture narrative descriptions and the CCT noun bias.
A general noun bias was identified in maternal speech. No significant relationship was found between the number of nouns used by mothers in input – both in noun tokens and type – and noun bias in the CCT. In contrast, as hypothesized, significant relationships were found with verb usage, both in terms of tokens and types: more extensive verb use in maternal input correlated with lower noun bias in infants’ CCT results. Although this relationship explains only a small percentage of variance, it is found in both verb types and tokens. Finally, no significant relationship was found for TTR ratio of verbs and nouns in maternal input to noun bias on the CCT. This last finding might simply be due to the structure of Arabic; having more options of how to present a verb (different tenses, verbal-templates, gender, and plurality) might lead to a higher TTR than for nouns.
A possible explanation for our results regarding types and tokens is that verbs require more understanding of complex relationships between concepts, syntactic structures, and contexts. When parents use more verbs, they may provide richer and more complex context stimulating infants to understand relationships between different words, including nouns, in ways that might reduce early noun bias. Thus, exposure to rich linguistic structures may encourage diversification in toddlers’ lexicon. In addition, extensive exposure to verbs may direct infants’ attention to actions and occurrences in the environment rather than to objects themselves. When extensive verb use occurs, infants need to understand different types of interactions and processes, potentially developing categorization ability and understanding beyond individual objects. That is, focusing on verbs may lead to the development of a more diverse lexicon not necessarily relying mainly on nouns.
These findings connect to previous research examining the influence of maternal input on early lexicon composition across different languages. Studies examining language acquisition in Chinese, Japanese, and Korean found that parents in these cultures typically use more verbs compared to Western language speakers. This increased verb exposure correlates with richer verb lexicons in children and reduced noun bias compared to English-speaking children, who show a stronger tendency toward nouns (Tardif, Reference Tardif1996; Choi & Gopnik, Reference Choi and Gopnik1995). Our findings similarly suggest that increased exposure to verbs may affect the balance between nouns and verbs in language acquisition.
The lack of correlation between noun usage in maternal input and noun bias in infants may reflect a natural developmental process. In early linguistic development, infants may develop a noun bias due to their natural interest in tangible and familiar objects, regardless of specific maternal input patterns. Since nouns provide clear reference points for concrete concepts, additional emphasis on nouns in maternal input may not significantly alter infants already established preference for them. Object names like “ball,” “cake,” or “book” are relatively easy to learn through observation, regardless of specific maternal input patterns. In contrast, understanding verbs requires more sophisticated cognitive processing of actions and relationships between objects, making verb-rich input potentially more impactful on lexical structure development. While noun bias could be a default pattern, exposure to verbs may enrich children’s understanding of relationships between objects and actions, promoting a more balanced lexical development. That said, the lack of significant noun effects might simply be due to low statistical power, as the non-significant effect was in the expected direction, and no significant statistical support for H1 does not equal support for H0.
It is important to note that the maternal input in our study was collected through picture narrative descriptions, which explain only a small percentage of the variance in children’s noun bias. Picture descriptions provide only partial insight into the linguistic input toddlers receive throughout their daily interactions with mothers. Different contexts and interaction types between mothers and children might reveal stronger connections between maternal input patterns and lexicon composition. For example, Tardif et al., (Reference Tardif, Gelman and Xu1999) noted that the context in which communication occurs affects the types of words parents use and, consequently, the words children tend to acquire. The study emphasized that the context in which children are assessed is crucial for understanding their preference for nouns or verbs. Additionally, infants’ lexical development is influenced by numerous factors beyond maternal input, including interactions with other family members, exposure to language in their social environment, and individual tendencies.
In our study, we asked mothers to respond to questions about cultural perceptions related to child-rearing, with ratings ranging from collectivistic perceptions (viewing the child as dependent on family and society) to more individualistic perceptions (viewing the child as autonomous). We found no significant relationship between cultural values and noun bias in maternal input. This finding may result from the relative homogeneity of our sample, as most mothers came from the same geographical area (Northern Triangle), which likely reduced demographic variability among participants. Alternatively, mothers’ collectivistic values may also not directly translate to their word choice patterns with children. The influence of cultural values might manifest primarily in interaction styles, attitudes toward specific topics, and broader communication approaches rather than in preference for particular word categories. Mothers with more individualistic attitudes may still use rich and diverse language appropriate to their child’s developmental level and the specific communication context.
The Palestinian Arab society in Israel has undergone significant cultural changes over time. Ben–Arieh et al. (Reference Ben–Arieh, Khoury–Kassabri and Haj–Yahia2006) found intergenerational shifts in child-rearing perspectives: while older generations tended toward authoritarian parenting styles, younger generations more often adopt authoritative approaches, setting clear boundaries while providing emotional support and guidance. This generational change suggests that examining the relationship between culture and maternal input requires deeper investigation of various cultural dimensions.
It is worth noting that although our cultural values questionnaire was based on previous research examining culture and maternal input, it did not undergo validity and reliability testing. The lack of established reliability and validity might have affected our results, as the questionnaire may not have consistently distinguished between different degrees of cultural orientation, limiting our ability to identify relationships between cultural values and lexical composition in maternal speech. The differentiation between individualist and collectivistic values as two points on a single line might also be simplistic. Collectivistic and individualist values can co-exist in the same individual (Triandis, Reference Triandis1994).
Regarding our null result for any effect of maternal education on input, while studies indicate that factors beyond cultural aspects, such as socioeconomic status, may influence lexical composition (Hoff, Reference Hoff2006), education’s influence on linguistic patterns may be less direct, and differences in linguistic input are not always clearly expressed in preferences for specific word categories like nouns or verbs. Studies show that maternal language patterns stem not only from demographic characteristics but also from parents’ child-rearing philosophies and interaction styles. Parents who respond sensitively and consistently to their infants tend to encourage children to learn words and understand their meanings across different contexts (Tamis-LeMonda, et al., Reference Tamis-LeMonda, Kuchirko and Song2014). We therefore did not have any directional hypothesis about maternal education, but rather treated this analysis as exploratory.
Our study has several limitations that should be considered when interpreting the results. First, the sample size was relatively small and homogeneous in terms of geography. The limited sample size may have restricted our ability to identify significant relationships between variables. Additionally, the homogeneity of our sample – with most participants coming from the Northern Triangle area – likely reduced variability in cultural characteristics. Future research would benefit from more representative sampling that includes populations from diverse geographic locations, potentially revealing cultural or demographic differences that influence maternal input and noun bias strength. Comparative research examining noun bias in Arabic versus other languages could further illuminate the relationship between culture and the strength of noun preference.
A second limitation involves our use of picture narratives to assess input. This methodology may not fully capture the richness and variety of parent–child linguistic interactions in daily life. Picture narratives may miss important aspects of more natural interaction contexts between mothers and children. Future studies could incorporate additional tools for evaluating linguistic input, such as observations of naturalistic parent–child interactions or recordings of everyday conversations, to assess maternal language in more authentic and diverse contexts.
To summarize, our findings strongly support the existence of noun bias in the receptive lexicon of Palestinian Arabic-speaking infants, likely stemming from both morphological characteristics of the language and cognitive aspects of language learning. Interestingly, while the number of nouns in maternal input did not significantly relate to children’s noun preference strength, the amount of verbs in maternal input showed a significant negative relationship with noun bias – greater verb usage corresponded with diminished noun bias in infants’ receptive vocabulary. We did not find significant relationships between mothers’ education level or cultural factors and their linguistic input patterns.
These results highlight the importance of considering language structure, cultural context, and maternal input together when studying early lexical development. The findings suggest that while noun bias may be a default pattern in early language acquisition, exposure to verb-rich input may help balance children’s lexical development. However, given the limitations of our research tools and sample, these findings should be considered preliminary groundwork for future more extensive and diverse investigations of how language structure, cultural context, and maternal input interact in shaping children’s early word learning.
Data availability statement
The materials, processed data and R analysis scripts that support the findings of this study are openly available in the Open Science Framework (OSF) repository at https://osf.io/vwxac/?view_only=4328f162828b4f05a177eaccbcb955f2.
Funding statement
The study was funded by startup funds from the School of Psychological Sciences at the University of Haifa to the second author.
Competing interests
The authors declare none.
Declaration of generative AI and AI-assisted technologies in the writing process
During the preparation of this work the authors used ClaudeAI in order to create the first draft, by translating the first author’s thesis to English, and to proofread the final version. We also used ClaudeAI to help with the R scripts to create the graphs. After using this tool/service, the authors reviewed and edited the content as needed and take full responsibility for the content of the publication.
Open access
