Phonological substitution patterns in Yemeni Ibbi Arabic child speech: a markedness and natural phonology perspective

Abstract

This study investigates phonological substitution patterns in the early speech of children acquiring Yemeni Ibbi Arabic (YIA) as their first language. The research aims to identify the characteristics and motivations behind these substitutions, drawing on data from three monolingual Arabic-speaking children aged between 1;9 and 3;6. The analysis reveals common substitution processes, including fronting, backing, lateralization, fortition, emphasization, and de-emphasization. These patterns are examined in relation to the children’s stages of consonant acquisition, the Substitution Phonological Pattern (SPP) hypothesis, markedness theory, and natural phonology. The findings suggest that substitutions are influenced by the child’s current stage in the consonant acquisition timeline and reflect developmental challenges associated with mastering more complex sounds. The study provides empirical evidence supporting the SPP hypothesis, which highlights the children’s strategic adaptations to phonetic and articulatory challenges. Additionally, the results align with the concepts of markedness and natural phonology, demonstrating a preference for simpler, more common phonological elements over complex ones. The research contributes to the understanding of phonological acquisition in Arabic and offers insights into the interplay between universal tendencies and language-specific patterns in child language development.

Keywords:

• Phonological substitution
• markedness
• natural phonology
• phonological acquisition

REVIEWING EDITOR:

• Jeroen van de Weijer, College of International Studies, Shenzhen University, Shenzhen, China

SUBJECTS:

• Introductory Linguistics
• Applied Linguistics
• Psycholinguistics

1. Introduction

Language acquisition is a fundamental aspect of cognitive development that reveals the complex processes through which children internalize the intricate sound system of their native language. This process is not merely a passive absorption of linguistic input but a dynamic interaction between the child’s innate cognitive abilities and their linguistic environment. One particularly interesting phenomenon observed during this interaction is phonological substitution, which highlights the linguistic competence of children as they employ adaptive strategies to navigate the phonetic landscape of their language. Studying first language acquisition is crucial for linguistic theory, as understanding the development of a child’s language system can provide valuable insights into the adult form of the language.

This study examines phonological substitution errors in the early speech of children acquiring Yemeni Ibbi Arabic (YIA), an understudied Arabic dialect spoken in Yemen. The investigation is grounded in the broader discourse on phonological acquisition, where the concept of markedness plays a crucial role in understanding the patterns and tendencies observed in language acquisition.

By examining the speech of three children aged 1;9 and 3;6, this study aims to identify the common patterns of phonological substitution within the broader frameworks of markedness and natural phonology, as well as Orfson-Offei’s (2023) Substitution Phonological Pattern hypothesis (SPP). This study contributes to the growing body of research on Arabic phonology and child language acquisition by situating the investigation of YIA phonological substitution within the context of markedness and its critiques.

1.1. Markedness, natural phonology, and SPP

The following subsections provide an overview of some of the basic concepts on which this paper is focused: markedness, natural phonology, and the Substitution Phonological Pattern hypothesis (SPP)

1.1.1. Markedness

Originally introduced by Trubetzkoy (1931) and further developed by Jakobson (1932, 1939), markedness was initially applied to phonological distinctions, suggesting that certain phonological features are inherently more ‘marked’ than others, which indicates a degree of complexity or rarity (Haspelmath, 2006). This concept was extended by Jakobson (1932, 1939) to semantic distinctions, proposing that marked elements carry additional semantic specificity compared to their unmarked counterparts.

Markedness is inherently comparative (Johnson & Reimers, 2010), which means it is used to express a relationship between at least two linguistic elements or phenomena, highlighting differences in complexity, frequency, acquisition, and phonetic difficulty. This concept establishes a hierarchy or spectrum where one element ‘X’ is considered marked relative to another element ‘Y’ based on various criteria such as phonetic complexity, frequency in languages, sequence in child language acquisition, and implications in language presence and change.

Haspelmath (2006) provides an in-depth critique of markedness where he distinguishes twelve different senses of markedness, grouping them into four categories: complexity, difficulty, abnormality, and multidimensional correlation. Three of the twelve senses relate to phonological acquisition. First, the Trubetzkoyan Markedness where certain phonological features are inherently more complex or ‘marked’. For instance, voiced consonants are considered marked relative to voiceless consonants in many languages (Trubetzkoy, 1931; also see Maddieson, 1984, p. 45). Second, Phonetic Markedness where some sounds are harder to produce and thus considered marked. For example, [θ] is less common in world languages and might be considered more marked than more widely used sounds like [t] or [d] (Gordon, 2016). Third, Distributional Markedness, where linguistic features with limited distribution within a language are marked. For instance, phonemes that only appear in loanwords.

The close relationship between markedness and naturalness leads us to the theory of Natural Phonology, which provides a framework for understanding how children navigate the marked and unmarked elements of their language during the acquisition process.

1.1.2. Natural phonology

The concept of markedness is closely tied to naturalness, often equated with it, which suggests that unmarked phenomena are ‘natural’ due to their simplicity, commonality, or early acquisition in child language.

Natural phonology, as proposed by Stampe (1969, 1972), posits that phonological development in children is governed by a set of ‘natural processes’ based on phonetic criteria. These processes are inherent in all children and are gradually suppressed as they learn the specific phonological system of their language. The natural processes reflect the phonetic motivations behind phonological patterns, such as ease of articulation and perceptibility, and are thought to underlie universal phonological tendencies across languages.

In terms of language acquisition, Natural Phonology holds that children are born with access to the full range of these universal phonological processes. As they acquire their native language, they must learn to systematically constrain and limit the processes according to the specific phonological system of that language. Processes that are not suppressed continue to operate in the child’s speech, resulting in common patterns seen in child language. Successful acquisition involves inhibiting the processes to match the adult phonological grammar. This contrasts with phonological rules, which are language-specific and must be learned by the child through learning and exposure (Dressler, 1984; Fromkin et al., 2018).

The relationship between markedness and natural phonology is significant. Marked phenomena are typically those that are not favored by natural phonological processes. For instance, complex consonant clusters, which are marked, are often simplified by children, which reflect natural phonological processes. Similarly, the presence of certain sounds only in more complex linguistic environments aligns with the notion that these sounds are marked and that their acquisition and usage are constrained by natural phonological processes.

However, there are instances where child phonology appears to be more natural than adult language, such as consonant harmony (e.g. ‘dog’ becomes ‘gog’), which is common in child language and can affect the primary place of articulation (Johnson & Reimers, 2010; Levelt, 2011) but rare in adult languages (Gordon, 2016).

The key distinction between markedness and natural phonology lies in their focus and application. Markedness provides a comparative framework that helps linguists understand the relative complexity and commonality of phonological elements across languages and in language acquisition. It is more about the attributes of the phonological elements themselves. In contrast, natural phonology focuses on the processes that children use to manage these elements as they learn language. It is more concerned with the dynamic aspects of how children interact with their linguistic environment during the acquisition process.

However, these two concepts are not entirely separate or opposites; rather, they complement each other in the study of phonological acquisition. Markedness can explain why certain phonological features are universally more difficult and thus avoided or simplified by children, while natural phonology describes the specific processes through which these simplifications occur. Together, they offer a comprehensive view of how children develop their phonological systems, accounting for both the inherent properties of phonological elements and the universal strategies children employ to acquire their native language’s phonology.

While Natural Phonology offers insights into the universal processes that guide phonological acquisition, the Substitution Phonological Pattern (SPP) hypothesis, proposed by Asare and Orfson-Offei (2023), focuses specifically on the systematic patterns of phonological substitution observed in children acquiring a second language.

1.1.3. Substitution phonological Pattern hypothesis (SPP)

The SPP hypothesis (Asare & Orfson-Offei, 2023) suggests that phonological substitution patterns are not random but are systematic responses to the challenges the children face due to the phonetic and phonological disparities between their L1 and English. The study grounds its findings within the framework of Natural Phonology, suggesting that these substitution patterns are natural phonological processes that reflect the children’s phonetic motivations, such as ease of articulation and perceptibility. These processes are inherently present in all children but are gradually modified or suppressed as they learn the specific phonological rules of their L2.

In their study, Asare and Orfson-Offei (2023) focused on a group of thirty preschool children who speak various Ghanaian languages as their first language (L1), such as Akan, Ga, and Ewe, and are in the process of acquiring English. The children’s English pronunciations were recorded and analyzed to understand how they navigate the pronunciation of English sounds that are challenging for them due to differences in the phonological systems of their L1 and English. Factors influencing the occurrence of SPP in the children’s speech were identified, including the absence of specific English sounds in their L1, the developmental stage of their articulatory organs, and their age. Older children showed fewer instances of substitution, indicating a developmental trajectory in phonological acquisition where children become more proficient in articulating L2 sounds as they grow older and receive more exposure and practice.

Although Asare and Orfson-Offei’s (2023) Substitution Phonological Pattern (SPP) hypothesis primarily addresses second language acquisition in children, integrating the concept of markedness with their hypothesis extends its applicability to first language acquisition in a monolingual context.

The SPP hypothesis provides a foundation for examining phonological substitution patterns in second language acquisition. In the following section, we will review the concept of phonological substitution and its manifestation in Arabic-speaking children in the context of first language acquisition.

1.2. Phonological substitution

Compared to other aspects of language acquisition, phonological substitution has not been widely covered in word languages, including Arabic. Phonological substitution, a concept central to understanding language acquisition and errors in speech production, refers to the replacement of one phoneme for another during speech. Children with specific language impairment (SLI) often display significant difficulties in language acquisition, differing from children with typically developing language (TD) peers in their ability to produce and comprehend language. This discrepancy is attributed to a range of factors including phonological memory limitations, phonological sensitivity, and motor articulatory challenges.

In Arabic, Dyson and Amayreh (2000) identified four substitution patterns in Arabic children. The most common was de-emphasization, which involves the loss of the secondary articulation (emphatic feature) in consonants and was marked with a decline with age but remained present to a considerable extent even in the oldest children studied. Second, lateralization, which involves the substitution of/l/with [r]. This process demonstrated a clear developmental trajectory with a quick decline by the age of 4;0 to 4;4. Third, fronting, in which back consonants (e.g. /k/ and /q/) are replaced with front ones (e.g. /t/). This pattern was relatively infrequent and showed a decline with age. Fourth, stopping, which involves the replacement of fricatives and affricates with stop consonants.

Having explored the general patterns of phonological substitution, we now turn our attention to the specific developmental trajectory of Arabic consonant acquisition, as outlined in the seminal work by Amayreh and Dyson (1998).

1.3. Acquisition of Arabic consonants

In Arabic, Amayreh and Dyson (1998) showed clear developmental trends in the acquisition of consonants in three periods. The three periods of Arabic consonant acquisition are the early period (2;0 to 3;10), the intermediate period (4;0 to 6;4), and the late period after 6;4. During the early period, children acquire at least 10 standard consonants (/b/,/t/,/d/,/k/,/f/,/ħ/,/m/,/n/,/l/,/w/). The intermediate period is characterized by the acquisition of the fricatives /s/,/ʃ/,/χ/,/ʁ,/h/, and the liquid /r/. The late period includes the acquisition of emphatic consonants, /q/,/ʔ/, interdental fricatives, the affricate /dʒ/, and the remaining voiced fricatives.

The study by Amayreh and Dyson (1998) provides an overview of the developmental patterns in the acquisition of Arabic consonants among Jordanian children. This research, grounded in the collection and analysis of data from 180 monolingual Arabic-speaking children aged between 2:0 and 6:4 years, addressed several questions regarding the correctness of consonant production at various ages, the mastery of phonemic sounds, the variation of consonant accuracy within sound classes by word position, and the comparison of Arabic phoneme acquisition timelines with those of English. Central to this study is the establishment of three distinct periods in the developmental trajectory of Arabic consonant acquisition: the early period (2:0 to 3:10), the intermediate period (4:0 to 6:4), and the late period (after 6:4).

During the early period, children acquire fundamental consonants including /b/,/t/,/d/,/k/,/f/,/ħ/,/m/,/n/,/l/, and /w/, forming the foundational blocks of their phonological inventory. This early acquisition underscores the rapid phonetic and phonological advancements occurring within the first years of life, with a clear emphasis on the mastery of essential consonants.

The intermediate period (4:0 to 6:4) is marked by the acquisition of more complex and less frequently used phonemes, including fricatives /s/,/ʃ/,/χ/,/ʁ/, and /h/, as well as the liquid /r/. This stage reflects a refinement in articulatory skills and an expansion of the child’s phonological repertoire, which accommodates sounds that require more precise manipulation of the speech apparatus and are often influenced by the phonetic context within which they occur.

The late period (after 6:4) introduces the acquisition of emphatic consonants, /q/,/ʔ/, interdental fricatives, the affricate /dʒ/, and remaining voiced fricatives, indicating a further sophistication in phonological development. The study’s findings suggest that these sounds, which may be influenced by linguistic factors such as phonetic environment and lexical frequency, represent the culmination of early phonological acquisition processes.

Amayreh and Dyson (1998) also examine consonant accuracy in relation to word position, showing that medial consonants are produced more accurately than those in initial and final positions. Moreover, the study contrasts the ages of phoneme acquisition in Arabic with those reported for English. While many consonants show similar timelines in both languages, specific phonemes such as /f/,/t/, and /l/ are acquired earlier in Arabic.

1.4. Phonemic system of YIA

Yemeni Ibbi Arabic (YIA) as a dialect spoken in the central part of Yemen and has several phonological features and system like the Standard Arabic (SA) and many other varieties and dialects of Arabic. The dialect may have prominent differences in terms of morpho-syntactic categories. The use of the voiceless velar /k/ at the end of the verbs for the first-person singular for feminine and masculine is the most common in comparison to the other dialects in Yemeni and other Arabic dialects as in [Ɂashtruk qalam] ‘I bought a pen’. Since the focus of the research article addresses more on the sound patterns and on one of the phonological processes, metathesis, a phonemic inventory of YIA is represented in Table 1.

Table 1. Phonemic inventory of YIA (Qasem, 2023).

	Labial	Labio-Dental	Inter-dental	Alveolar	Palato-alveolar	Velar	Uvular	Pharyngeal	Laryngeal
Plosive	b		t   d			k g	q		Ɂ
Emphatic			ṭ
Fricative		f	θ	s       z	ʃ     ʒ		χ  ʁ		h
			ð
Emphatic			ð̣	ṣ				ħ   ʕ
Nasal	m			n
Lateral				l
Tap				r
Glides	w				j

An aspect of the phonemic inventory under scrutiny involves emphatic consonants and their phonetic characteristics—whether they are pharyngealized, uvularized, or even glottalized. The academic discourse on Arabic dialects reveals a lack of consensus among researchers concerning several aspects: the precise articulation locations (primary or secondary), the specific part of the tongue involved, and the extent of tongue movement (retraction or elevation) required to produce these consonants. For instance, emphatic consonants across different Arabic dialects have been variously categorized as velarized (Nasr, 1959), pharyngealized (Al-Ani, 1970; Al-Tamimi et al., 2009; Bin-Muqbil, 2006; Hess, 1998; Kenstowicz, 1994), uvularized (Shar, 2012; Zawaydeh & de Jong, 2011), and glottalized (Watson, 2002). This disagreement may stem from several factors: the dialectal variations present within the Arabic language, individual variances among speakers of the same dialect, and the diversity of methodologies (such as X-ray, MRI, ultrasound, videofluoroscopy, etc.) employed in their study, each with its unique limitations. The findings presented in this paper provide implications on the phonetic characteristics of emphatic consonants in YIA, which is one of the key aspects investigated in this study of phonological substitution patterns in child language acquisition.

In existing research, phonological substitution has gained limited attention, despite being common in the speech of children acquiring either their first language or a second language. Consequently, this study aims to examine the substitution process in children’s speech by identifying the causes and motivations behind these errors, and to explore the context in which they occur. The primary objectives of this study are to:

1. Investigate the nature of phonological substitution as it appears in the speech of YIA acquirers.

2. Determine the factors that lead to phonological substitution in the speech of YIA acquirers.

2. Methods

2.1. Participants

The dataset in the current study was collected from three monolingual Arabic-speaking children, all of whom are children of parents who speak YIA. None of the children exhibited signs of language disorders or hearing impairments. These children were not exposed to any secondary language (L2) environments. The group consisted of two females and one male (Table 2).

Table 2. Age ranges and genders of participants.

Child	Period of data collection	Gender
Layan	1;9–3;6	Female
Shada	1;10–3;2	Female
Ali	2–3;6	Male

2.2. Data collection

Before beginning data collection, the researchers obtained the required permissions from the research ethics committee in the department to which they are affiliated. The parents of the children involved were thoroughly informed about the procedures, the methods of data anonymization, and the use of the data in this study. To protect their privacy, all children were assigned pseudonyms.

Data were collected longitudinally as part of an ongoing corpus that aims to document the syntactic acquisition of YIA. The data collection process was designed to capture children’s speech in a variety of naturalistic contexts to provide a comprehensive understanding of their language development. The researchers employed multiple data collection tools, including voice recorders and diary entries, to ensure a rich and diverse dataset.

Voice recordings were made in various settings, such as children’s homes, playgrounds, and nurseries, to capture their interactions with peers, parents, siblings, and other family members. These recordings were made at regular intervals over the course of several months, allowing for the observation of phonological substitution patterns across different stages of development. The researchers also made note of the specific contextual details for each recording, including the interlocutors, the nature of the interaction (e.g. playtime, mealtime, storytelling), and any relevant linguistic features (e.g. exposure to different dialects, use of baby talk or motherese).

In addition to voice recordings, the researchers maintained detailed diary entries to document children’s language use in contexts where audio recording was not feasible or appropriate. These diary entries included observations of children’s speech during daily routines, such as bathtime or bedtime, as well as their interactions with extended family members or caregivers. The researchers also noted any instances of code-switching or exposure to other languages in these diary entries.

To ensure the naturalness and authenticity of the collected speech samples, the researchers primarily relied on unobtrusive observation and recording methods. They aimed to minimize their interference in children’s natural interactions and communication patterns. However, in rare instances where specific linguistic structures or phonological patterns of interest were not naturally occurring in a specific setting, the researchers engaged with the children to elicit these forms through play-based activities or targeted prompts. It is worth noting that the use of elicited repetition is a common and accepted practice in child language research, particularly when studying specific linguistic structures or patterns that may not occur frequently in spontaneous speech (Devescovi & Caselli, 2007; Eisenberg et al., 2001; Klem et al., 2015; Marinis & Armon-Lotem, 2015). The data analyzed in this paper constitutes the instances of phonological substitution in the corpus.

At the start of the data collection process, the children’s ages ranged from 1 year and 9 months to 2 years, which positions them within the early development phase as described by Amayreh and Dyson (1998). By the end of the data collection period, their ages had progressed to between 3 years and 2 months and 3 years and 6 months. To analyze the children’s speech development, their utterances were compared with standard adult utterances of Yemeni Ibbi Arabic.

3. Results and discussion

In this section, we describe the characteristics of phonological substitution errors observed in the dataset. Overall, we identified 36 instances of phonological substitution, distributed as 12 instances for Ali, 10 instances for Layan, and 14 instances for Shada. To visualize the distribution of these substitution types across the three children, Figure 1 presents a plot of the frequency of each substitution type for Ali, Layan, and Shada.

Figure 1. Frequency of substitution types across children.

While the observed substitution patterns were consistent across the three participants, individual differences in the frequency and types of substitutions were also noted. For instance, Ali and Shada exhibited a slightly higher frequency of fronting compared to Layan, with each having 5 instances of fronting while Layan had 4. Lateralization was equally distributed among the three children, with each having 3 instances. Backing was also relatively evenly distributed, with Ali and Shada having 3 instances each and Layan having 2. However, some substitution patterns were only observed in certain participants. Layan was the only child who exhibited fortition and emphasization, with 1 instance of each. De-emphasization was only observed in Shada’s speech, with 1 instance. These individual differences may be attributed to factors such as age, language exposure, and potential developmental variability. Further research with a larger sample size could help elucidate the extent and implications of these individual differences in substitution patterns among YIA-speaking children.

Table 3 displays the total number of instances of phonological substitution identified within the data along with some examples.

Table 3. Sample of phonological substitution patterns.

Process	Number of occurrences	Adult speech	Child speech
Fronting	14	/ku:rah/	/tu:rah/
Lateralization	9	/taru:h/	/talu:ħ/
Backing	8	/χargana/	/ħardna/
Fortition	1	/θaʕlab/	/taɁlab/
Emphasization	1	/alqaiʕah/	/alṭaliʕah/
Debuccalization	2	/ħilwah/	/hilwah/
De-emphasization	1	/ʃurṭah/	/ʃultah/

The table above describes various phonological substitution patterns where specific target sounds are systematically replaced with alternative sounds. The first pattern seems to involve simplification of the articulatory effort. For instance, the substitution of /θ/ and /ṭ/ with /t/ suggests moving from more specific and possibly more effortful articulations to a simpler, more universally found articulation. Similarly, replacing /k/ with a /t/ indicates a change to a sound that requires less precise positioning of the tongue.

Moreover, there is a tendency towards alveolar sounds where various sounds are substituted with alveolar sounds (especially /t/), which might reflect a universal tendency towards centralization in the mouth, where alveolar sounds are produced. This could be due to the alveolar sounds being less marked universally as they are cross-linguistically the most frequently occurring place of articulation (Maddieson, 1984). The shift from specific articulatory positions that are less common and more challenging for the child, towards those that are more central and commonly found across languages, indicates a natural tendency towards simplification in line with the principles of markedness and natural phonology.

3.1. Fronting

Fronting is a phonological process that involves the substitution or replacement of consonants articulated at the back of the oral cavity, especially /k/ and /q/, with those produced closer to the front, often at the alveolar ridge. This process is particularly common in child phonology. Lowe, Knutson, and Monson (1985) conducted a study analyzing speech from 1,048 children aged between 2 years and 5 months to 4 years and 5 months. They found that 6% of these children exhibited fronting, with velar fronting being more prevalent than palatal fronting. The study also revealed a decrease in fronting occurrences as children grew older, dropping from 23.3% in the youngest group to 3.5% in the oldest. Moreover, fronting can affect sounds already considered front, making them even more anterior. Johnson and Reimers (2010, pp. 14–15) observed instances where some children pronounced the palatal /ʃ/ as /s/ (e.g. ʃɪp → sɪp) and the alveolar /s/ as /θ/ (e.g. sʌn → θʌn).

In Arabic (Dyson & Amayreh, 2000), this process occurred for velar and uvular stops /k/ and /q/ (but not the uvular fricatives), but these instances were relatively infrequent and showed a decline with age, suggesting that the articulatory placement for back sounds becomes less challenging as children’s speech develops. Cross-linguistically, uvular consonants are marked (Maddieson, 2013), while velars are unmarked in world languages, but are usually acquired relatively late by children (Crowe & McLeod, 2020; Priester et al., 2011).

In the analyzed data, patterns similar to those reported by Dyson and Amayreh (2000) are observed. As demonstrated in Table 4, both the voiceless uvular plosive /q/ and the voiceless velar plosive /k/ are replaced with the voiceless alveolar plosive /t/, exemplified in the substitutions /tari:b/ for /qari:b/ ‘near’ and [tu:rah] for /ku:rah/ ‘football’. Interestingly, in our data, the process of fronting appears to be influenced by the voicing characteristic of the original consonants. For instance, in the adult pronunciation of /gami:l/, the voiced velar stop /g/ is substituted with the voiced alveolar plosive [d]. This indicates that the child, Layan, aged 3;1, has mastered the voicing distinction between the two consonants, though she has not yet developed the ability to articulate sounds at the velum. In addition to /q/ and /k/, the palato-alveolar fricative /ʒ/, which is one of the consonants that are acquired late in Arabic (Amayreh & Dyson, 1998), is also fronted to /d/ by the child Shada.

Table 4. Fronting examples.

Child	Child Form	Adult Form	Gloss	Age
Shada	[tidah]	/kidah/	like this	1;10
Layan	[tursi]	/kursi/	chair	1;9
Layan	[samat]	/samak/	fish	1;9
Ali	[tu:rah]	/ku:rah/	football	2
Shada	[tajtah]	/kajkah/	cake	2;3
Shada	[tari:b]	/qari:b/	near	2;5
Shada	[la:daħ]	/ra:dʒaħ/	Rajeh ‘name’	2;6
Ali	[alħadeedah]	/alħadeeqah/	the garden	2;8
Layan	[dami:l]	/gami:l/	beautiful	3;1

The observed fronting patterns in our study align with cross-linguistic tendencies in child phonology. In English, Ingram (1976) reported that velar fronting was common in children aged 2–4 years, while Lowe et al. (1985) found that 6% of children aged 2;5–4;5 exhibited velar fronting. Similar patterns have been observed in other languages, such as Dutch (Altvater-Mackensen & Fikkert, 2010) and German (Schaefer & Fox-Boyer, 2017). The prevalence of fronting in our Arabic data suggests that this process is a universal simplification strategy employed by children across languages. The tendency towards alveolar sounds in child phonology has been observed cross-linguistically and is supported by empirical evidence. In a study of English-acquiring children, Stoel-Gammon (1985) found that alveolar consonants were among the most frequently produced sounds in early word productions. Similar patterns have been reported in other languages, such as Japanese (Tsurutani, 2007). This preference for alveolar sounds can be explained by the concept of ‘articulatory ease’ in natural phonology (Stampe, 1969). Alveolar consonants require less complex articulatory gestures compared to sounds produced at other places of articulation, making them easier for children to produce (Locke, 1983).

3.2. Backing

While fronting involves the substitution of back consonants with front ones, the opposite process, known as backing, was also observed in the data, albeit less frequently. Interestingly, since uvular stops were fronted, the two uvular fricatives in YIA were backend. Hodson and Paden (1981) identified ‘backing’ as a process among children whose speech was deemed essentially unintelligible. This process was observed in certain cases within the corpus. Moreover, Dyson and Amayreh (2000) did not detect any instances of backing in their study.

In our data, we observed examples of backing in three children. Layan consistently replaced the voiced uvular fricative /χ/ with the voiceless pharyngeal fricative /ħ/, as in the word [wasiħ] instead of /wasiχ/. Further examples of backing are shown in Table 5 for Shada and Ali. Contrary to the pattern observed with fronting, the resulting sounds in children’s speech may differ in their voicing characteristics compared to the adult forms, as can be seen in words like [muʕalliq] and [ṣaʕi:lih].

Table 5. Backing examples.

Child	Child Form	Adult Form	Gloss	Age
Layan	[wasiħ]	/wasiχ/	Dirty	2;4
Shada	[muʕalliq]	/muʁalliq/	Closed	3;2
Ali	[ṣaʕi:lih]	/sˁaʁi:rih/	Short	3;6

Backing patterns in child speech have been less extensively studied compared to fronting. In their study of English-speaking children, Hodson and Paden (1981) found that backing was associated with unintelligible speech. The backing of uvular fricatives in our YIA data may reflect language-specific challenges in acquiring these marked sounds, which are rare cross-linguistically (Maddieson, 2013). Further research is needed to investigate the prevalence and nature of backing processes in Arabic and other languages.

3.3. Fortition

Additionally, the data show that fricatives are sometimes substituted for plosives. Example of this are Ali’s production of /θaʕlab/ ‘fox’ as [taɁlab] and Shada’s production of /kiðah/ ‘like this’ as [kidah]. It should be noted that the voicing contrast between the adult interdental fricatives was maintained in the children’s new productions.

This observation could indicate a preference for complete closure during articulation, accompanied by a sensitivity to the voicing feature of consonants. This aligns with the observations of Dyson and Amayreh (2000, p. 98), who found that children often substituted interdental fricatives with stops that shared the same voicing characteristics. A similar pattern was identified in Dutch children learning Dutch, as noted by Altvater-Mackensen and Fikkert (2010), where the vast majority of errors involved substituting fricatives with stops. This phenomenon could be attributed to two things. First, the general linguistic principle that stop consonants are considered less complex than fricatives, a concept supported by works of Jakobson (1968), Maddieson (1984), and Kenstowicz (1994). Two, in world languages, /ð/ and /θ/ are usually among the consonants acquired late (Priester et al., 2011; Crowe & McLeod, 2020).

3.4. De-emphasization

Although not common in the data, certain sounds are subject de-emphasis. For instance, the phoneme /ṭ/ is replaced by/t/. For example, in one case, Shada replaced the emphatic alveolar stop /ṭ/ in the word /ʃurṭah/ ‘police’ with the non-emphatic [t], resulting in [ʃultah].

3.5. Emphasization

For emphasization, or added emphasis, the child Layan (at age 3;4) did the opposite, where she produced the adult form /alqajʕah/ ‘Day of Judgment’ as [alṭajʕah]. In this word, the child first fronted the uvular stop /q/ to /t/, then added emphasis to the fronted consonant. Emphasis in Arabic typically involves the articulation of certain coronal consonants with a retracted or constricted tongue root or back, which imparts a pharyngealized or uvularized characteristic to these sounds, thereby assigning them dual articulation points (Ladefoged, 2001). Interestingly, Qasem (2023) describes emphatic consonants in YIA as pharyngealized. Nonetheless, the substitution patterns observed among the children suggest that emphatics in this dialect might more precisely be described as uvularized (e.g. /tʶ/) instead of pharyngealized. Such substitutions of emphatic consonants were infrequent in the data.

In the data at hand, substitutions involving emphatic consonants were rare. Considering that coarticulation requires simultaneous articulatory gestures, it is understandable that children might struggle with these sounds. One way to explain this is that emphatic consonants in this dialect are in fact uvularized rather than pharyngealized, and that the child at this stage is still unable to produce or maintain a steady occlusion in the uvula to produce the uvular /q/. Therefore, the child fronted the primary occlusion to the alveolar ridge, and the secondary constriction to the uvula, as an attempt to produce the uvular quality of the adult sound /q/. At this developmental stage, the child might still be mastering how to produce or maintain a consistent closure at the uvula to articulate the uvular /q/. Consequently, Layan shifted the primary occlusion to the alveolar ridge, with the secondary constriction at the uvula, attempting to mimic the uvular place of articulation of the adult’s uvular stop /q/.

The difficulty in acquiring emphatic consonants can be understood within the framework of markedness theory (Jakobson, 1941). Emphatic consonants are considered marked due to their complex articulatory requirements, which involve a secondary articulation in addition to the primary place of articulation (Ladefoged & Maddieson, 1996). Marked structures are typically acquired later and are more prone to simplification processes in child language (Gnanadesikan, 2004). The substitution of emphatic consonants with their non-emphatic counterparts in our data aligns with the predictions of markedness theory and is supported by similar findings in other Arabic dialects (Ammar & Morsi, 2006; Dyson & Amayreh, 2000).

3.6. Lateralization

The lateralization process is observed in children’s speech when the voiced alveolar lateral /l/ is used instead of the voiced alveolar trill /r/ (Table 6). This phenomenon, which is widespread in children acquiring various languages, represents a common stage in phonological development (Smith, 1973). An example of this can be seen in the way children pronounce many words, such as substituting [labbi] for the adult form /rabbi/ ‘my God’. The recurrence of lateralization in multiple studies suggests that the /l/ sound is acquired more readily by children than the /r/ sound. One possible explanation for this could be the complex tongue movements required for articulating rhotics, involving simultaneous retraction and advancement of the tongue, a skill that children may not yet have fully developed (Howson & Redford, 2019).

Table 6. Lateralization examples.

Child	Child Form	Adult Form	Gloss	Age
Ali	[talu:ħ]	/taru:h/	Do you go?	2;3
Shada	/ra:dʒaħ/	Rajeh ‘name’	/ra:dʒaħ/ Rajeh ‘name’	2;6
Layan	[lu:ħ]	/ru:ħ/	Go!	3;1

This finding is consistent with studies on English (Smit, 1993) and Dutch (Beers, 1995), which report that /l/ is typically acquired before /r/. The difficulty with rhotics may stem from the complex tongue configurations required for their articulation (McGowan et al., 2004).

3.7. Debuccalization

The last phonological substitution process is debuccalization. Debuccalization is a phonological process in which a consonant sound loses its place of articulation towards the throat, typically transforming into a sound produced closer to the glottis, like the /h/ sound, or even becoming a glottal stop [ʔ]. This transformation often occurs to sibilants and other consonants in certain linguistic environments. In the dataset, debuccalization was observed twice: once with the child Shada at 2;1, who articulated the word /ħilwah/ ‘sweet’ as [hilwah], and again with the child Ali at 2; 8, who converted the adult pronunciation /θaʕlab/ ‘fox’ to /taɁlab/.

The substitution patterns observed in the Yemeni Ibbi Arabic-acquiring children’s speech can be further analyzed through the lens of markedness theory. According to this theory, certain phonological features are considered more marked than others, indicating a higher degree of complexity or rarity (Jakobson, 1941; Trubetzkoy, 1931). Less marked features are typically acquired earlier and are more common in the world’s languages (Jakobson, 1968).

The fronting of back consonants like /k/ and /q/ to front ones like /t/ aligns with the idea that front consonants are less marked than back consonants. Alveolar sounds, such as /t/, are among the most frequent consonants across languages and are usually acquired early by children (Locke, 1983; Maddieson, 1984). This suggests that the children in this study are substituting more marked back consonants with less marked front ones as they navigate the acquisition process.

Similarly, the lateralization of /r/ to /l/ and the fortition of fricatives to stops can be seen as a preference for less marked sounds. Lateral approximants, like /l/, are more common and typically acquired before rhotics, such as /r/ (Amayreh & Dyson, 1998). Stops are also considered less marked than fricatives, as they are more frequent in the world’s languages and are often mastered earlier by children (Jakobson, 1968).

The de-emphasis of emphatic consonants and the emphasis of non-emphatic ones highlight the marked nature of emphasis in Arabic. Emphatic consonants, which involve a secondary articulation, are less common across languages and are acquired later by Arabic-speaking children (Dyson & Amayreh, 2000). The children’s difficulties with emphatic sounds and their tendency to replace them with non-emphatic counterparts reflect the marked status of emphasis.

These findings demonstrate that the substitution patterns in Yemeni Ibbi Arabic-acquiring children’s speech are not random but are influenced by the principles of markedness. The children tend to replace more marked sounds with less marked ones, following a path of least resistance in their phonological development. This aligns with the cross-linguistic tendencies observed in child language acquisition and supports the role of markedness in shaping the acquisition process (Gnanadesikan, 2004; Jakobson, 1941).

4. Conclusion and implications

The study examined the phonological substitution in the early speech of children speaking Arabic as their First Language (L1). This research aimed at exploring the characteristics of the substitution phenomenon and the factors that motivated the substitution process to occur in child speech. This study investigated these phonological substitutions in relation to the SPP hypothesis (Asare & Orfson-Offei, 2023) as well as markedness and natural phonology.

The findings of the study showed that there are common phonological substitution patterns and processes including fronting, backing, lateralization, fortition, emphasization, and de-emphasization. Fronting, where back consonants are replaced with front ones, might be more common during the early acquisition stage when children have mastered front consonants like /b/, /t/, /d/, but are still developing the ability to articulate back consonants like /k/, /q/. The study shows examples where /t/ replaces /q/ and /d/ replaces /g/, which aligns with fronting occurring because these back consonants are not yet fully acquired. In contrast, backing, a less common process where front sounds are replaced with back ones, might be related to individual differences in the acquisition timeline or specific challenges certain children face in mastering front consonants. As children progress through the intermediate and late periods, they learn to produce more complex sounds like fricatives and rhotics. Lateralization and fortition might occur during this stage, with the replacement of /r/ with /l/ (lateralization) and fricatives with stops (fortition) being indicative of the challenges children face with these sounds, which are typically acquired later.

The types of phonological substitutions (e.g. fronting, backing, lateralization, fortition) appear to be influenced by the child’s current stage in the consonant acquisition timeline. As children progress from acquiring simple consonants to more complex sounds, their phonological errors reflect the developmental challenges associated with these stages. This connection suggests that phonological substitutions are not random but are influenced by the physiological and cognitive aspects of language development, which are, in turn, reflected in the structured acquisition of consonants as outlined by Amayreh and Dyson (1998) for Arabic-speaking children.

The study offers empirical evidence supporting the SPP hypothesis by Asare and Orfson-Offei (2023), which posits that children naturally substitute complex speech sounds with simpler alternatives as a strategic response to phonetic and articulatory challenges. This finding, additionally, highlights the adaptability of the phonological systems of the children during the early stages of language development, which highlights how they navigate speech production challenges. While the SPP hypothesis provides a useful framework for understanding phonological substitution patterns, it is important to consider that Asare and Orfson-Offei’s (2023) study focused on second language acquisition in Ghanaian children learning English. Our study extends the application of the SPP hypothesis to first language acquisition in a monolingual context, demonstrating its potential relevance beyond second language acquisition. However, further research is needed to determine the extent to which the SPP hypothesis can be generalized across different languages and acquisition contexts.

The study aligns with the concepts of markedness and natural phonology in phonological development by showing that simpler, more common phonological elements are favored by children over more complex, less common ones. This preference is reflected in the types of phonological errors and substitutions made by children, indicating an innate inclination towards simplicity and commonality in speech sounds. The findings from studies on Arabic phonological acquisition, especially Dyson and Amayreh (2000), show common patterns such as simplification processes (e.g. fronting, stopping) that are observed universally in children’s speech development. However, the presence of emphatic sounds, interdental fricatives, and uvular sounds in Arabic introduces language-specific substitution patterns not typically found in languages without these sounds. For instance, the substitution of dental or alveolar sounds for emphatic sounds in YIA might not have direct parallels in languages lacking emphatic consonants. Moreover, it is important to note that while Dyson and Amayreh’s (2000) findings provide valuable insights into Arabic phonological acquisition, their study focused on a different Arabic dialect (Jordanian Arabic) and may not fully capture the intricacies of Yemeni Ibbi Arabic. Our study builds upon their work by examining the specific patterns of phonological substitution in YIA, revealing both similarities and differences in the acquisition process across Arabic dialects.

The substitution patterns observed in the speech of Yemeni Ibbi Arabic-acquiring children align with the predictions of the markedness hypothesis. According to this hypothesis, unmarked elements are favored over marked ones in language acquisition and use (Jakobson, 1941). Unmarked phonological features are considered simpler, more common across languages, and acquired earlier by children (Haspelmath, 2006). Our findings contribute to the growing body of evidence supporting the role of markedness in shaping phonological acquisition across languages. However, it is crucial to consider that markedness alone may not fully explain all aspects of phonological development. Factors such as language-specific phonological structures, frequency effects, and individual differences in acquisition trajectories may also play a significant role in shaping children’s phonological patterns. Future research should aim to integrate insights from markedness theory with other theoretical perspectives to develop a more comprehensive understanding of phonological acquisition.

The influence of markedness on phonological acquisition is evident in the specific substitution patterns observed in our data. The fronting of back consonants like /k/ and /q/ to front ones like /t/ reflects a preference for unmarked, simpler articulations. Alveolar sounds like /t/ are cross-linguistically more frequent and acquired earlier than velar and uvular sounds (Amayreh & Dyson, 1998; Maddieson, 1984). Similarly, the lateralization of /r/ to /l/ and the fortition of fricatives to stops indicate a tendency towards less marked, more common sounds. Lateral approximants and stop consonants are typically less marked than rhotics and fricatives, respectively (Ladefoged & Maddieson, 1996).

Furthermore, the acquisition of emphatic consonants in Arabic provides additional support for the markedness hypothesis. The de-emphasis of emphatic consonants and the emphasis of non-emphatic ones suggest that children struggle with the marked feature of emphasis, which involves a complex secondary articulation. Emphatic consonants are considered marked in Arabic and are usually acquired late by children (Dyson & Amayreh, 2000). The simplification and acquisition challenges associated with emphatic sounds further support the markedness hypothesis.

The cross-linguistic prevalence of the observed substitution patterns in our study contributes to the growing body of research supporting the role of markedness and natural phonology in shaping child phonological development across languages. The observed substitution patterns, such as fronting, stopping, and lateralization, have been reported in numerous studies on English (Ingram, 1976; Smit, 1993), Dutch (Altvater-Mackensen & Fikkert, 2010), German (Schaefer & Fox-Boyer, 2017), and other languages. The prevalence of these patterns across languages suggests that they are universal simplification strategies driven by the principles of markedness and natural phonology.

The findings of our study also have important implications for theories of phonological development and language acquisition. The observed patterns of phonological substitution, such as fronting, backing, and lateralization, align with the predictions of markedness theory (Jakobson, 1941) and natural phonology (Stampe, 1969). The preference for unmarked, less complex structures in child speech supports the idea that phonological development is guided by universal principles of simplification and ease of articulation (Gnanadesikan, 2004). However, our data also reveal language-specific influences, particularly in the acquisition of Arabic emphatic and uvular consonants, suggesting that theoretical models need to account for both universal and language-specific factors in phonological development (Vihman, 2014). Furthermore, the gradual reduction of substitution processes as children age aligns with the concept of progressive faithfulness in Optimality Theory (McCarthy & Prince, 1995), which posits that children gradually acquire more marked structures as they develop. Future research should aim to integrate these findings with existing theoretical frameworks and explore how they can inform the refinement and expansion of models of phonological acquisition.

Despite the implications provided by our study, it is important to acknowledge its potential methodological limitations. The small sample size of three children may limit the generalizability of our findings to the broader population of YIA-acquiring children. Future research should aim to replicate these results with a larger, more diverse sample to ensure the robustness of the observed patterns. Additionally, the limited contexts of data collection, primarily focusing on naturalistic settings, may not capture the full range of phonological variability in child speech. Furthermore, while researcher intervention was minimal and primarily aimed at eliciting specific linguistic structures, it is possible that this intervention influenced the children’s productions. Future studies could benefit from employing more unobtrusive data collection methods to minimize researcher influence (Eisenbeiss, 2010).

To further understand the phonological acquisition in YIA and other languages, future research will investigate the point at which each substitution process disappears in each of the children. Additionally, future research will also look at other aspects of child phonology that were not examined in this research such as vowel substitution, prosodic features (e.g. intonation, stress, rhythm), and morphophonological development by examining how children learn the phonological rules that govern word formation and alteration. Investigating these additional aspects of child phonology would contribute to a more comprehensive understanding of language acquisition processes and address existing gaps in the literature. Exploring vowel substitution patterns could reveal parallels or differences between consonant and vowel acquisition, which would tell us more about overall phonological development in children. Examining prosodic features, such as intonation, stress, and rhythm, would provide information about the suprasegmental aspects of language acquisition and how they interact with segmental phonology. Studying morphophonological development could uncover how children learn and apply phonological rules in word formation and alteration, bridging the gap between phonology and morphology in language acquisition research. By addressing these aspects, future research would paint a more complete picture of the complex nature of child language development, ultimately contributing to the refinement and expansion of theoretical models in the field.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

The authors are thankful to the Deanship of Graduate Studies and Scientific Research at University of Bisha for funding this research through the promising program under grant number (UB-Promising-47-1445).

Notes on contributors

Abdullah Alfaifi

Abdullah Alfaifi is an Assistant Professor of Linguistics at the Department of English Language and Literature, University of Bisha, Saudi Arabia. His research interests include Arabic phonetics and phonology, Arabic dialectology and documentation, and language acquisition.

Fawaz Qasem

Fawaz Qasem is an Associate Professor at Department of English language and literature, University of Bisha, Saudi Arabia. He received his Ph.D. in Linguistics and Phonetics. His research interests include language acquisition, language disorder, language and technology and AI. He works as consultant, editor, and reviewer of several scholarly journals.