Reading and Memory Skills of Children with and without Dyslexia in Greek (L1) and English (L2) as a Second Language: Preliminary Results from a Cross-Linguistic Approach

Abstract

The focus of the present paper is twofold; the first objective is to examine how children with dyslexia (henceforward DYS children) and typically developing children (henceforward TD children) performed in Greek (first language; L1) compared to English (second language; L2) in reading, phonological awareness (PA), rapid automatized naming (RAN), working memory (WM), and short-term memory (STM) tasks. Our second goal is to investigate DYS children’s performance compared to that of TD children in the L1 and L2 domains mentioned above. Thirty-two (DYS = 16; TD = 16) school-aged children (9;7–11;9 years old; M_age = 130.41), basic users of English (level ranging from A1 to A2), carried out a battery test in L1 and L2, respectively, including reading, PA, STM, and WM tasks. More specifically, the tasks were the following: word and nonword decoding, reading accuracy and reading fluency, word and nonword reading per minute, PA, RAN, nonword repetition, as well as forward, backward, and digit span sequencing. This is a work-in-progress study, and preliminary results reveal that DYS students exhibit important reading and memory deficits in both languages. The data analysis indicated that DYS children have particular difficulties and statistically significant differences in L1 and L2 compared to TD in all tasks. In conclusion, this is the first study, at least in Greek, which assesses both reading and memory skills of DYS children in L2. The results reveal deficits in both languages, and the overall findings contribute to theories on the transfer of difficulties of linguistic skills between L1 and L2, while memory scores also underline this co-occurrence. Future implications of this study include a combination of reading and cognitive activities in the teaching methods of English teachers to improve DYS children’s overall performance in learning English as L2.

1. Introduction

Over the last decades, foreign language learning has been a central topic in many educational policies; hence, many countries have chosen to integrate English as an L2 into their national curriculum. Knowing how to communicate in a non-native environment is a remarkable asset for many individuals regarding their personal and career prospects. Without a doubt, it seems inequitable to deprive students with specific learning disorders (SLDs) of this critical skill. It is a common fact that these individuals experience several difficulties in their native language and also when acquiring a second/foreign language or any other additional language (Cimermanová 2015; Kormos 2020; Kormos and Nijakowska 2017). Research has shown that learning an L2 is a considerable challenge for children with SLDs (Ferrari and Palladino 2007; Ganschow and Sparks 1995). Likewise, students with dyslexia (DYS students) with significant deficits in their mother tongue are highly likely to exhibit these weaknesses when they learn and study another language, too (Kormos 2017; Łockiewicz and Jaskulska 2016; Maunsell 2020; Siegel 2016). This is commonly attributed to the fact that dyslexia is a persistent, lifelong language condition, which means that individuals face challenges in learning any language, regardless of whether it is their first, second, or even third language (Siegel 2016).

1.1. Dyslexia in L1 and L2

Dyslexia is considered the most prevalent learning disability (Kormos et al. 2019; Maunsell 2020) among the existing subtypes of SLDs. According to the Fifth Edition of the Diagnostic and Statistical Manual of Mental Disorders of the (DSM-5, American Psychiatric Association 2013), dyslexia meets the criteria of a neurodevelopmental disorder that manifests itself as an impairment specified in word reading accuracy, reading rate, or fluency, and/or reading comprehension. Individuals with dyslexia (DYS individuals) experience challenges in reading and spelling despite receiving adequate tuition from educators or parents, while these difficulties are not aligned with their age and IQ (Snowling et al. 2020). To date, it is still hard to distinguish one concrete definition of dyslexia, as over the years, many perennial studies have approached this disorder to delineate it best. However, it is unanimously recognized that weaknesses in phonological processing skills constitute a core deficit in populations with dyslexia (Ramus et al. 2003; Ramus and Szenkovits 2008; Snowling 2001; Vellutino et al. 2004; Ziegler and Goswami 2005) along with poor word reading abilities. In general, decoding deficits in dyslexia primarily stem from poor phonological representations of speech sounds (see Hulme and Snowling 2009, for review).

Furthermore, since dyslexia is a multifactorial disorder, various symptoms can emerge, simultaneously or not, and in different domains. In fact, there has been growing adoption of the view that deficits in dyslexia extend beyond phonological weaknesses. To be more specific, children and adults with dyslexia face difficulties in a series of tasks including PA (Cantiani et al. 2013; da Silva et al. 2020; Smail et al. 2022; Stampoltzis et al. 2020; Talli et al. 2013), reading fluency and accuracy (da Silva et al. 2020; Giazitzidou and Padeliadu 2022; Melloni and Vender 2022; Rothou and Padeliadu 2019), vocabulary (Giazitzidou and Padeliadu 2022; Maassen et al. 2022; Rothou and Padeliadu 2019), morphosyntax (Cantiani et al. 2013; Talli et al. 2013), and spelling (Giannouli and Pavlidis 2014). Additionally, memory impairments might also be clinical markers of dyslexia. In particular, DYS individuals show poor performance on cognitive skills such as RAN (Bogdanowicz et al. 2014; Carioti et al. 2022; da Silva et al. 2020; Stampoltzis et al. 2020; Jones et al. 2009), WM (da Silva et al. 2020; Gathercole and Pickering 2000; Masoura et al. 2021; Smail et al. 2022) and STM measures (Bogdanowicz et al. 2014; Cantiani et al. 2013; Lazzaro et al. 2021; Masoura et al. 2021; Zhao et al. 2015).

Although until recently, the majority of research undertaken into dyslexia included monolingual populations, there has been a gradual increase in studies investigating the aforementioned characteristics of dyslexia in L2. English as a foreign language can be very challenging to learn, sometimes beyond dyslexic abilities (Kormos and Smith 2012). On top of that, regardless of their native language, learners of English as an L2 tend to exhibit common and similar difficulties (Lundberg 2002). Notwithstanding, it is suggested that DYS students who study English as L2 can develop better literacy skills and show an advantage over monolingual dyslexics (Jarsve and Tsagari 2022; Siegel 2016). Dyslexia and second language acquisition (SLA) can co-exist in the sense that participation in foreign classes is anything but discouraged.

A wide variety of factors affect how learners with dyslexia (DYS learners) and regular learners confront the process of learning a foreign language. According to the Orthographic Depth Hypothesis, the degree of transparency in alphabetic languages renders specific tasks more difficult due to PA irregularities (Buetler et al. 2014; Geva and Siegel 2000; Katz and Feldman 1983; Katz and Frost 1992; Schmalz et al. 2015). If a language’s orthography is complex, learners may struggle with both word and nonword reading (Seymour et al. 2003). By the same token, since English is an opaque (non-transparent) language with an irregular orthography and inconsistent relationships between letters and sounds, it becomes more demanding for DYS learners, and sometimes even for TD students, to decode. On the other hand, reading and writing in English as L2 may not be considered as such complex tasks because learners have experienced similar literacy skills in their L1 and are equipped with appropriate and fundamental linguistic knowledge in general (Pinter 2006). However, becoming a proficient reader in one’s L1 can vary depending on individual differences among learners. As a result, it is highly likely for DYS students to face difficulties initially exhibited in L1 during L2 learning (Helland and Kaasa 2005).

1.2. Cross-Linguistic Influence between L1 and L2

It is important to note that cross-linguistic influences can be detected when an individual learns two languages simultaneously. Kellerman and Smith (1986) first introduced the term cross-linguistic influence, defining it as the phenomenon where a person applies their knowledge of one language to another language they are learning. This process involves both linguistic and cognitive elements, allowing the individual to draw upon their existing language skills to aid in learning a new language (Moattarian 2013). Another similar definition is language transfer, which occurs when similarities and differences between the new language and any previously (and maybe imperfectly) acquired language impact the learning process (Odlin 1989). Regarding this definition, Archibald (1998) pointed out that one of the most noticeable characteristics of an L2 learner’s speech is its similarity to their L1. Consequently, language transfer is possible when features or rules of one language (L1) affect the use of the target language, such as the carryover of vocabulary, grammar structures, and pronunciation patterns. This type of transfer is possible when L2 is learned after L1, with the learners primarily relying on their L1 in L2 learning. Besides, it has been observed that, especially when children begin learning two languages in early childhood, they exhibit some types of mixing, which may confirm that transfer between L1 and L2 does indeed occur (Odlin 1989).

Phonological and phonetic transfer from L1 is highly likely in L2 learners, either children or adults, who make significant errors while acquiring the new language due to the influence of their L1 (Archibald 1998; Zampini 1994). In Zetterholm’s (2024) study, phonological transfer from L1 to L2 was apparent when participants were asked to produce both oral and written speech, demonstrating an important correlation between these skills. Nonetheless, as mentioned above, cross-linguistic interaction in the phonological domain may be more prevalent in younger learners throughout the early stages of syllable structure development (Kehoe 2020). Undoubtedly, during the process of learning a foreign language, an intertwined linguistic system is developed as L1 and L2 interact with each other, aiding in comprehension and expression. For instance, Kehoe (2020) indicated that perceiving and producing complex structures in one language definitely facilitates the production of these structures in the other language. On the other hand, according to Major (2008), L2 learners may encounter difficulties in perceiving differences in the L2 due to the transfer of their L1 perceptual systems when exposed to L2.

In scholarly literature, the terms cross-linguistic transfer and cross-linguistic influence are occasionally used interchangeably (Jarvis and Pavlenko 2008). However, in this paper, the term transfer of difficulties is preferable for various reasons. According to hypothesis Russak and Zaretsky’s (2021), cross-linguistic influences of L1 skills are possible while acquiring English as L2 and, as Łockiewicz and Jaskulska (2019) underpinned, this procedure reflects a potential transfer of language difficulties. It is generally acknowledged that DYS readers with weaknesses in their L1 are liable to experience similar difficulties in their L2, a fact ruled by the transfer of skills from one language to the other. More specifically, data from L2 learners of English reveal a correlation between struggling with reading in English and facing similar challenges in their L1, particularly in alphabetic languages. These results show a mutual relationship between poor reading skills in L1 and L2 (Gao et al. 2019). Further, a very recent study by (Venagli and Kupisch 2024) provided consistent results, offering valuable insights from the learners’ perspective. The researchers compared the self-perception of L1 (German or Italian, which are transparent languages as Greek) and L2 (English) competence in DYS children using questionnaires and found an important interrelationship. Not surprisingly, the participants rated their reading and writing skills as significantly low in both their L1 and L2. Therefore, these findings demonstrated the interconnected nature of L1 and L2 language skills and underlined that perceived weaknesses in one language may also be transferred to the other.

Notably, difficulties in L2 arise from inadequate phonological skills in L1, which are indicative of the fundamental challenges in dyslexia. As a result, it can be hypothesized that DYS learners may primarily experience L2 phonologically-based issues similar to those in their L1. The presence of these factors may suggest that individuals are able to apply their L1 knowledge and skills in L2 phonology, morphology, orthography, and naming speed across different languages, indicating a transfer of language skills and overall cognitive processing abilities (Chung and Ho 2010; Gao et al. 2019; van der Leij and Morfidi 2006). To say the least, the transfer of specific phonological and reading-related skills across different languages suggests the existence of universal linguistic underpinnings and a universal phonological core for reading in two languages (Gao et al. 2019; Perfetti et al. 1992).

In order to approach this transferability of difficulties, the current study is structured upon a widely accepted hypothesis, the Linguistic Coding Difference Hypothesis (LCDH; Ganschow et al. 1991; Sparks and Ganschow 1993; Sparks et al. 1989). The LCDH predicted that students’ poor performance in L2 originates from fundamental difficulties in relevant areas in their L1. The authors emphasized that L1 learning and L2 learning are causally connected and that phonology or phonological coding is the most challenging aspect for learners. According to the LCDH, the phonological code underlines the interrelation between L1 and L2 acquisition capabilities. In a nutshell, it is speculated that phonology, syntax, and semantics are critical linguistic skills that impact L2 learning.

Furthermore, the Central Processing Hypothesis (CPH; Cossu et al. 1988; Geva and Siegel 2000; Stanovich 1984) highlighted that the nature of the orthography (based on the Orthographic Depth Hypothesis mentioned earlier) is not the only factor affecting reading acquisition. Learners’ reading profiles in L1 and L2 will correlate even if these two orthographic systems differ in complexity and regularity. Conversely, some underlying cognitive processes, such as memory (short-term verbal memory and serial naming) and linguistic components (phonological skills), contribute to acquiring reading skills in L1 and L2, respectively. Moreover, this shared collection of fundamental cognitive concepts can accurately forecast individual variations in reading abilities across all languages (Geva et al. 1993). Additionally, developing reading skills can be complicated by linguistic and cognitive challenges, regardless of weaknesses in the learners’ L1 or L2 (Geva and Siegel 2000). To sum up, it is postulated that cognitive skills such as phonological processing (PA and memory) and RAN, which support spelling and reading, are universal across various writing systems rather than language-specific (Moll et al. 2014). However, there is a lack of Greek data available to support or verify this finding, as well as data stemming from L2 research.

1.3. Relevant Research and Links with the Present Study

Taking everything into consideration, it is unambiguous that dyslexia differentiates the way children master linguistic and cognitive skills not only in L1 but also while learning L2 (Kormos et al. 2019). There is compelling evidence from remarkable studies exclusively in English as an L2, confirming the transfer of difficulties from the student’s native tongue to English (Slovenian: Kormos et al. 2019; Dutch: van Setten et al. 2017; German: Maurer et al. 2021; Polish: Łockiewicz and Jaskulska 2016; Chinese: Chung and Lam 2020; Li et al. 2018; Tong and McBride 2017; Spanish: Álvarez-Cañizo et al. 2023; Suárez-Coalla et al. 2020; Italian: Bonifacci et al. 2017; Fazio et al. 2021; Greek: Tsesmeli et al. 2021; Sotiropoulos and Hanley 2017; Andreou and Segklia 2017; Andreou and Baseki 2012). As a consequence, it has been observed that similarities and differences between L1 and L2 do not hinder the transitional process. This supports the notion that difficulties in both languages seem to co-occur.

The present study aims to cast further light on how DYS children learn English as L2. As far as we know, there has been limited research in the Greek language on the transfer of linguistic skills between L1 and L2 (English) in students with and without dyslexia. According to the LCDH, we assume that L1 abilities are a foundation for L2 learning, while both L1 and L2 acquisition rely on basic language learning mechanisms. We also aim to investigate whether acquiring essential reading and memory skills in L1 significantly impacts the L2 learning process and whether difficulties in L1 learning can result in poor L2 proficiency. Furthermore, there is a significant research gap regarding the memory skills of L2 learners with and without dyslexia, which the present study aims to address. By examining participants’ reading and memory skills, we speculate that based on the CPH, reading acquisition is not solely dependent on the orthographic system of a language but on cognitive processes as well. As a result, our study examines whether deficits in certain cognitive skills, i.e., STM, WM, and RAN, hinder the acquisition of basic reading skills in DYS and TD children, regardless of the language or script being learned, such as Greek or English.

In particular, our first objective is to examine DYS and TD children’s performance in Greek (L1) compared to English (L2) in a series of tasks encompassing reading, PA, RAN, WM, and STM. Secondly, we seek to investigate DYS learners’ performance compared to that of TD children in L1 and L2 in the aforementioned domains. Based on the LCDH supporting cross-linguistic transfer of difficulties and the underlying cognitive mechanisms crucial for language learning, we speculate that participants’ with dyslexia (DYS participants’) performance will be lower in both languages and across all tasks, verifying that weaknesses detected in L1 will be transferred in L2. Last, we highlight the domains that appear more challenging for DYS learners, namely the tasks with significant differences between the two groups. Τhe novelty of this study concerning dyslexia in L1 and L2 is that it includes the assessment of not only reading and PA skills, but also cognitive skills, namely RAN, STM, and WM. Such skills may appear deficient in L2 as well as in L1 in DYS populations, despite differences in the two orthographic systems. Depending on the aims of the present study, the following hypotheses were addressed:

• DYS children’s performance in English (L2) is expected to be lower than in Greek (L1) in reading, PA, and RAN, as well as in WM and STM.
• TD children are expected to perform better in Greek (L1) than in English (L2) in reading and PA and similarly in L1 and L2 in RAN, WM, and STM.
• The DYS group’s performance in L1 is expected to be lower in reading, PA, RAN, WM, and STM compared to that of the TD group.
• DYS children are expected to obtain lower scores in L2 on reading, PA, RAN, WM, and STM compared to TD children.

2. Materials and Methods

2.1. Participants

The sample of the present study consisted of Greek-speaking children between 9;7 and 11;9 years old. All students attended primary school and learned English as L2. The specific age range corresponds to children learning English since the first grade and have already acquired an A1/A2 proficiency level. A language placement test, the Greek state certificate for foreign language proficiency (known as KPG), was conducted before the battery test and, according to the Common European Framework of Reference for Languages (CEFR) scale, verified this elementary level, classifying participants as basic users of English. General inclusion criteria for participation in this study included Greek native language, English as an L2, and non-verbal IQ superior or equal to 85 (assessed using Raven’s Progressive Matrices; Raven 2004). Bilingual children who learned English as a third language, children with general cognitive delay, as well as children with mental or neurological disorders were excluded from participation.

Participants were divided into a group of DYS children and a TD group. All members belonging to the DYS group had previously received the diagnosis of specific learning disability with an impairment in reading provided by the official governmental agency for diagnosing learning and speech-language difficulties in Greece (KE.D.A.S.Y.; Centers of Interdisciplinary Assessment, Counseling, and Support) after being assessed by accredited clinical professionals in the region of Macedonia. In addition, all DYS participants performed at least 2 SD (standard deviations) below average on the reading subscales of a Greek standardized test, Test A (Padeliadu and Antoniou 2008). DYS children were recruited from Speech and Language Therapy Centers, Rehabilitation Centers, Centers of Creative Activities for Children, and two public primary schools, all located in different areas in Thessaloniki. In the TD group, participants were enrolled in two separate public primary schools in Thessaloniki. These students demonstrated normal school performance without any special reading or writing difficulties in their age-appropriate classroom based on teachers’ and parents’ reports, as well as their performance on Test A in reading subscales (greater than or equal to average). Participants in both groups had been learning English since first grade and attended English lessons in foreign language centers for three years on average.

To ensure baseline comparability, we utilized descriptive statistics and found no significant differences between the two groups in terms of chronological age, non-verbal reasoning, and English level (p > 0.05). The number of participants in each group was the same (n = 16; boys, 7, girls, 9). The main characteristics of the participants are accumulated in

Table 1. Demographic characteristics of the two groups separated and together in terms of non-verbal IQ, chronological age, and English level.

Groups	DYS (n = 16)		TD (n = 16)		Participants (n = 32)
Gender	Boys = 7 Girls = 9		Boys = 7 Girls = 9		Boys = 14 Girls = 18
Variables	M	SD	M	SD	M	SD	sig.
Chronological age (months)	129.50	7.86	131.31	8.00	130.41	7.86	0.491
Raven (raw scores)	30.62	1.96	31.50	1.82	31.06	1.91	0.224
English level (A1–A2)	1.38	0.465	1.47	0.427	1.42	0.442	0.539

.

2.2. Procedure

Ethical approval for the current study was granted by the Aristotle University of Thessaloniki (AUTh) Research Ethics Committee (protocol code 188367/2022) in accordance with AUTh requirements. Before data collection, all participants received written informed parental consent, which authorized children to participate in this survey. DYS and TD groups were required to complete a battery test in L1 and L2, including reading, PA, STM, and WM tasks. The researcher administered standardized and non-standardized tests in Greek and English as well. All children underwent individual assessment in a quiet room provided by the school or the equivalent center. Each assessment was conducted orally by the same examiner, who was also responsible for scoring. The evaluation process lasted approximately 2 h (1 h for the Greek tests and 1 for the English tests) and was divided into four sessions (30 min each). The assessment tools were administered in the same order for all children in L1 and L2 respectively. However, measuring consecutive similar tasks was avoided due to being too demanding for each participant (e.g., performing two cognitive tasks in a row). As a result, the researcher allocated the tasks to ensure reliable outcomes for both groups’ reading and cognitive assessments. Last but not least, the researcher attempted to choose tasks that were as equivalent as possible in English and Greek.

2.3. Assessment Tools

2.3.1. Non-Verbal Intelligence

Raven’s Educational Colored Progressive Matrices Test (CPM; Raven 2004), along with the Greek standardization (Sideridis et al. 2013), was implemented to assess subjects’ non-verbal intelligence. The test required children to select one of the six options that best completed a matrix with a missing part. There were 36 problems with three different sets of increasing difficulty. Only participants with non-verbal IQ superior or equal to 85 were included in this study.

2.3.2. Assessment in Greek

Reading Tasks

• Word and nonword decoding

Test A (reading test; Padeliadu and Antoniou 2008), a standardized Greek diagnostic tool, was used to evaluate participants’ decoding skills in real words and nonwords. In particular, the first subdomain of decoding involves three separate subscales appropriate for assessing decoding skills in single-word reading. In the first exercise, children were asked to read aloud and decode 24 nonwords in Greek. The second task followed the same procedure with 53 real Greek words. Finally, in the third task, each participant had to distinguish between real words and pseudowords and also be able to decode the former correctly. This subscale consisted of a list with 36 words in total. For each exercise, every correct response accounted for one point, while zero points were given for the incorrect answers. Raw scores and percentiles based on the normative data (50th percentile is the mean score) were initially noticed. Percentages out of raw scores (24 nonwords and 53 words) were also calculated to ensure accurate comparison with the English word/nonword decoding task. Without this step, comparing Greek percentiles with English raw scores would not be possible, so we converted both into percentages. Finally, we compared the scores of the word decoding subscale with two English tasks, once with regular words and once with irregular words (since Greek regular words gradually become more difficult in this task) to match the English tasks (see Section 2.3.2. word and nonword decoding) which included regular and irregular words separately.

Reading Fluency

Reading fluency was assessed with the text fluency subscale of Test A (reading test; Padeliadu and Antoniou 2008). According to the instructions, children were required to read a text aloud (total number of words: 279) as rapidly and accurately as possible, only for one minute. Correct responses included the words a child could read in 60 s minus the words read incorrectly (i.e., errors). Both raw scores and percentiles were enumerated to fulfill the requirements of the present study. In particular, raw scores were used to accurately compare the English reading fluency task.

Reading Accuracy

Giro giro oloi (Talli 2010), a Greek adaptation of the French Alouette test (Lefavrais 1967), constitutes a meaningless text with primarily rare and unknown words that participants had to read aloud (total number of words: 271) in 180 s without paying attention to its contextual information. This test evaluates students’ reading level by taking into account both accuracy and fluency. A composite score was calculated by adding the total time in seconds, the total number of errors, and the words that remained non-read.

Word and Nonword Reading per Minute

During the word reading per minute (Talli 2010) assessment, participants were instructed to read a list of high-frequency Greek words as fast and accurately as they could in 1 min. The list consisted of 50 words with one and two syllables. Correct phonological decoding and stress within 60 s formed the total word reading score. Similarly, the assessor used an identical approach to evaluate the nonword reading assessment and determine the score (Talli 2010). Pseudoword selection involved a list of 50 one- and two-syllable words that complied with the phonotactic structure of the Greek language.

Phonological Awareness (PA) Tasks

To measure participants’ PA, the assessor implemented a Greek non-standardized PA test based on the French battery EVALEC (Sprenger-Charolles et al. 2005), encompassing three separate tasks: one syllabic and two phonemic (Talli 2010). In particular, the syllabic test included 10 tri-syllabic CV pseudowords, while the other two tests consisted of 24 tri-phonemic pseudowords, 12 with a CVC structure and 12 with a CCV structure. The children were instructed to delete the first element of an item, either a syllable or a phoneme, depending on the task given. Accuracy scores and percentages were calculated, whereas no time restriction was mentioned.

Memory (and Cognitive) Tasks

• Short-term memory (STM)

(i)

Forward digit span

The Greek adaptation (Stogiannidou et al. 2017) of the Wechsler Intelligence Scale for Children—Fifth Edition (WISC-V; Wechsler 2014) was administrated in the present study. In particular, we used the forward digit span test to measure verbal STM (VSTM). This task required participants to listen carefully to verbal numbers and immediately repeat them in the same order as presented by the examiner. Number sequences begin with two digits and reach up to 10. There are two trials for each sequence; a correct trial is scored one point, whereas an incorrect one equals zero points. The test terminates when the examinee fails in both trials of a specific sequence. After calculating raw scores, they were converted into standard scores using Greek norms, considering all three test subscales.

(ii)

Phonological short-term memory (PSTM)

A Greek nonword repetition test with available norms for children aged 7–13 years old (Talli et al. 2023), which is an adaptation of the French battery EVALEC (Sprenger-Charolles et al. 2005), was used to assess phonological STM (PSTM). Children were required to repeat 24 pseudowords of increasing length as accurately as possible. The task encompassed three- to six-syllable items with both consonant–vowel (CV) and consonant–vowel–consonant (CVC) syllables. Raw scores (number of correctly repeated syllables) and percentiles were taken into account in this study, while we also enumerated participants’ spans depending on the number of syllables (max six).

(iii)

Verbal short-term memory (VSTM)

A Greek adaptation (Chrysochoou 2006) of a task included in the Working Memory Test Battery for Children (WMTB-C; Pickering and Gathercole 2001) was selected in order to measure subjects’ VSTM. The specific task comprised serial recall of spoken verbal stimuli. Children had to repeat lists of real two-syllable words presented in several blocks of trials (six trials per block). For instance, in the first block, participants had to repeat one-word trials; in the second block, two-word trials, and so forth. Administration was discontinued after three failed trials in a particular block. In the current study, we considered both raw scores and percentages.

Working Memory (WM)

(i)

Backward digit span

To assess the subjects’ WM, we administered the backward digit span test of WISC-V GR (Wechsler 2014; Greek version by Stogiannidou et al. 2017). In this task, we evaluated participants’ ability to manipulate information in WM. More specifically, children were prompted to recall and repeat a series of numbers spoken aloud by the researcher following the reverse order. As in the forward condition, the initial number of the digits in both sequences was two, and then increasingly, they reached up to 10. Participants have to succeed in at least one of the trials of each sequence for the test not to stop. Raw totals for each subscale were summed and then converted into standard scores and percentiles based on Greek norms.

(ii)

Digit span sequencing

Digit span sequencing is the last subscale of the WISC-V GR (Wechsler 2014; Greek version by Stogiannidou et al. 2017) evaluating WM and must be administered after the forward and backward spans. The specific task asked children to recall a series of numbers presented verbally by the researcher and repeat it following the ascending order. The number of digits in each sequence was the same as the previous subscales, and the discontinue rule followed the same pattern. According to the Greek norms, raw totals for forward, backward, and sequencing digit spans were added and turned into standard scores and percentiles.

Rapid Automatized Naming (RAN)

Children were administered the Picture Naming Speed task from the Phonological Assessment Battery test (PhAB; Frederickson et al. 1997). After receiving a Picture Naming Card, participants were required to repeatedly name in Greek as quickly as possible a random sequence or five common objects (a table, a door, a box, a ball, and a hat) presented on the card. There were five rows with 10 pictures in each, and children were driven by the visual stimuli (object) they received. Scores included the time in seconds children needed to complete the test.

2.3.3. Assessment in English

With the exception of the reading fluency and accuracy measure, the English assessment tools used in this study were standardized tests. However, as the normative data of each measurement pertain to English native speakers, it was impossible to apply these norms to students whose native language is Greek and who learn English as L2. As a solution, we utilized percentages across all tasks to prevent relevant issues and maintain cross-linguistic differences within reasonable limits.

Reading Tasks

• Word and nonword decoding

The Castles and Coltheart Test 2 (CC2; Castles et al. 2009) was used to assess decoding skills through single-word reading in English. The entire test included 40 regular words, 40 irregular words, and 40 nonwords, presented to the examinee one at a time, in mixed order and with gradually increasing difficulty. Children’s score on the test was the sum of their correct responses (including all three categories) until administration ceased (after five consecutive items were misread). Furthermore, the categories were graded separately to determine differences in students’ competency between word and nonword reading. After completing each task, we calculated both raw scores and percentages. Finally, the scores of the three subscales were added together to obtain the total competence score for the test.

Reading Fluency and Accuracy

The researcher designed a non-standardized test as a part of her Ph.D. thesis to evaluate participants’ reading accuracy and fluency. The text contained 288 words with varying frequencies—high, high/medium, and medium/low. The frequency of the words was controlled using two basic books that are suitable for English (second) language learners—“The ESL/ELL Teacher’s Book of Lists” (Kress 2008) and “The Reading Teacher’s Book of Lists” (Fry et al. 2000). Before the text was initially used, it was proofread by a native English teacher to ensure accuracy in terms of spelling, grammar, syntax, and punctuation. Scoring was calculated in two different periods: first, by adding the number of words children were able to read correctly within 60 s; secondly, by adding the total time in seconds (maximum 180 s) children needed to complete the whole test, the total number of errors they committed and the number of the words remained non-read (based on the scoring pattern indicated by the Greek reading test Giro giro oloi).

Word and Nonword Reading per Minute

The Test of Word Reading Efficiency—Second Edition (TOWRE-2; Torgesen et al. 2012) was implemented to measure participants’ ability to pronounce regular words (sight word efficiency) and phonemically regular nonwords (phonemic decoding efficiency). Instead of 45 s, given in the original test instructions, we decided to provide participants with 60 s to achieve complete correspondence with the Greek task. Accuracy scores for each task included only the number of correctly decoded words and nonwords within the specified time limit (1 min).

Phonological Awareness (PA) Tasks

This study used two PA tasks of the Phonological Assessment Battery test (PhAB; Frederickson et al. 1997). First, the alliteration test evaluated children’s ability to isolate the initial sounds in single-syllable words. In particular, the examinee listened to three words (=1 test item) and was required to say aloud only two of them that started with the same sound. In the specific task, there were 10 test items separated into two parts of five each. Part 1 included easier words with single consonants, and for the test to continue, children had to respond correctly to at least three test items; otherwise, administration stopped. Accordingly, more difficult words with consonant blends were presented in Part 2. The total number of correct responses in Part 1 and Part 2 equaled the total score on the alliteration test. Secondly, in the rhyme test, participants were required to identify rhyming in single-syllable words. More specifically, they listened to three words (=1 test item) and were required to say aloud only two of them that rhymed. This subscale consisted of 21 test items with three words, each divided into two parts with increasing difficulty. There were 12 test items in Part 1. An essential precondition for the administration to continue was having at least nine correct responses in Part 1. Once the child finished the first part, there were nine more test items in Part 2. We added Part 1 and Part 2 out of 21 examples to calculate the total scoring of the rhyme test. Therefore, for the needs of the present study, we added the raw scores of both subscales (alliteration + rhyme) and converted them into percentages.

Memory (and Cognitive) Tasks

• Short-term memory (STM)

(i)

Forward digit span

To evaluate participants’ STM, we used the forward digit span subscale of the Wechsler Intelligence Scale for Children—Fifth Edition (WISC-V; Wechsler 2014). This task required participants to listen carefully and repeat numbers in exact order. The only difference between the Greek and English assessments was that digits were presented verbally in English. The administration process, number sequences, and discontinuation rules were the same as those used in the L1 task. Only raw scores were taken into account for the English assessment.

(ii)

Phonological short-term memory (PSTM)

The Children’s Test of Nonword Repetition (CNRep; Gathercole and Baddeley 1996) was used to measure children’s PSTM. Participants listened to a single unfamiliar phonological item and were instructed to repeat it immediately as accurately as possible. The repetition accuracy was scored for 40 such nonwords in mixed order with syllables varying from two to five (10 words for each). Raw scores were determined by counting the number of correctly repeated words, which were then converted into percentages. Spans were also calculated based on the number of syllables (max five).

(iii)

Verbal short-term memory (VSTM)

The Rey Auditory Verbal Learning Test (RAVLT; Rey 1964; Schmidt 1996; Spreen and Strauss 1998) was selected to evaluate VSTM. Participants were given a list (List A) of 15 unrelated, one- and two-syllable words repeated over five different trials. Each time they listened to the whole list, students were asked to recall as many words as possible following their own order. We administered only the immediate recall subscale of the particular test. Scoring involved correct word recalls of all five trials without intrusions (words that did not belong to the list) turned into percentages.

Working Memory (WM)

(i)

Backward digit span

The backward digit span subtest of the WISC-V (Wechsler 2014) was implemented to measure WM. As in Greek, the researcher spoke a series of numbers aloud, but this time using English digits, and asked children to repeat them in reverse order. The administration process, number sequences, and discontinuation rules were identical to those in the L1 assessment. For this L2 tool, only the raw scores were retained.

(ii)

Digit span sequencing

According to the digit span sequencing (WISC-V; Wechsler 2014) used to evaluate WM, children were prompted to recall and repeat verbally presented numbers in English in ascending order. The scores included in the analysis were only raw scores. No differences concerning the administration procedure, number of sequences, and ceasing rules were mentioned between L1 and L2 assessment.

Rapid Automatized Naming (RAN)

To measure participants’ rapid naming skills, we administered the original Picture Naming Speed task from the Phonological Assessment Battery test (PhAB; Frederickson et al. 1997). The participants were asked to name the objects in English after being presented with the same visual stimuli in Greek. The test score was based on the time taken by the children to complete it.

2.4. Statistical Analysis

IBM SPSS Statistics for Windows, version 29, was used for data analysis. Since we had a small sample size, we initially conducted the Shapiro–Wilk and Kolmogorov–Smirnov tests to select an appropriate statistical method. The results rejected the null hypothesis and revealed that the distribution of the majority of the tasks departed significantly from normality. Based on this outcome, only non-parametric tests were used in this study. The subsequent statistical procedure included the Wilcoxon Signed Ranks test, a non-parametric statistical method used to compare two related samples (within-group comparison). The Wilcoxon test provided results related to two research questions in the current study. First, we compared DYS children’s performance in Greek and English, and secondly, we followed the same process to compare how TD participants’ performance was in L1 and L2 tasks correspondingly. Moving to our third and fourth research question, we utilized computing descriptive statistics and the Mann–Whitney U test. Descriptive statistics were used to calculate and define the mean scores and standard deviations of both groups in all variables in L1 and L2 separately. The Mann–Whitney U test was used as a nonparametric test to compare the two independent samples (between-groups comparison) and evaluate whether they exhibited significant differences. In particular, this test was utilized two times. In the first case, we compared DYS and TD performance in all tasks in L1, and then, likewise, we made the same comparison regarding L2 measurements.

3. Results

3.1. DYS Performance in Greek as Compared to English in Reading and Memory Tasks

For the analysis of DYS participants’ performance in L1 and L2, the Wilcoxon signed-rank test was used for all variables, including decoding (word, irregular word, and nonword), reading accuracy and fluency, word and nonword reading per minute, RAN, PA, and STM (forward digit span, PSTM, VSTM) and WM (backward digit span and digit span sequencing). Results are presented in

Table 2. DYS (n = 16) within-group analysis comparing Greek and English reading and memory tasks.

Variables	Greek	English	Z	Sig.
Decoding WR	72.18	54.68	−3.516	<0.001 ***
Decoding IWR	72.18	25.15	−3.517	<0.001 ***
Decoding NWR	53.91	43.12	−2.871	0.004 **
Decoding total	75.32	40.99	−3.517	<0.001 ***
Reading fluency	73.19	40.56	−3.522	<0.001 ***
Reading accuracy	324.88	362.81	−3.361	<0.001 ***
WR per minute	56.00	41.13	−3.440	<0.001 ***
NWR per minute	31.56	25.87	−3.126	<0.001 ***
RAN	66.06	65.69	−0.114	0.909
PA	58.68	43.14	−3.519	<0.001 ***
STM forward digit span	7.56	6.50	−2.846	0.004 **
WM backward digit span	5.81	5.31	−1.999	0.046 *
WM digit span sequencing	5.44	4.31	−2.342	0.019 *
PSTM	95.50	64.68	−3.517	<0.001 ***
PSTM Span	3.81	3.31	−2.828	0.005 **
VSTM	39.90	40.25	−0.155	0.877

* p < 0.05, ** p < 0.01, *** p < 0.001.

for Greek and English, respectively.

A Wilcoxon signed-rank test indicated significant differences between the mean values of L1 and L2 performance in most tasks. More specifically, the Wilcoxon test revealed that scores were significantly different in word decoding (z = −3.516, p ≤ 0.001), irregular word decoding (z = −3.517, p ≤ 0.001), and nonword decoding (z = −2.871, p = 0.004) together with the total decoding score (z = −3.517, p ≤ 0.001). Further, reading accuracy (z = − 3.361, p ≤ 0.001) and fluency (z = −3.522, p ≤ 0.001), word reading per minute (z = −3.440, p ≤ 0.001) and nonword reading per minute (z = −3.126, p ≤ 0.001) were distinguishable. Significant differences were also reported regarding PA (z = −3.519, p ≤ 0.001). In the memory tasks, the Wilcoxon test showed that performance significantly differed in STM variables such as the forward digit span (z = −2.846, p = 0.004) and PSTM (z −3.517, p ≤ 0.001) with its corresponding span (z = −2.828, p = 0.005), as well as in WM tasks namely the backward digit span (z = −1.999, p = 0.046) and the digit span sequencing (z = −2.342, p = 0.019). On the other hand, DYS participants obtained similar scores on RAN (z = −0.114, p = 0.909) and VSTM (z = −0.155, p = 0.877) in Greek and English, respectively, without any differences.

3.2. TD Performance in Greek as Compared to English in Reading and Memory Tasks

The Wilcoxon signed-rank test was used to analyze how TD participants performed in the L1 battery test compared to the equivalent L2 test. The Wilcoxon test was applied to all variables, encompassing word, irregular word, and nonword decoding, reading accuracy and fluency, word and nonword reading per minute, RAN, PA, STM (forward digit span, PSTM, VSTM), and WM (backward digit span and digit span sequencing). The results for the Greek and English variables are presented in

Table 3. TD (n = 16) within-group analysis comparing Greek and English reading and memory tasks.

Variables	Greek	English	Z	Sig.
Decoding WR	95.73	85.62	−3.465	<0.001 ***
Decoding IWR	95.73	59.06	−3.518	<0.001 ***
Decoding NWR	81.74	72.96	−2.984	0.003 **
Decoding total	95.91	72.56	−3.519	<0.001 ***
Reading fluency	129.88	106.44	−3.259	0.001 **
Reading accuracy	179.06	176.38	−0.233	0.816
WR per minute	113.25	78.31	−3.517	<0.001 ***
NWR per minute	66.75	47.31	−3.517	<0.001 ***
RAN	41.31	40.00	−1.084	0.279
PA	95.26	81.24	−3.520	<0.001 ***
STM forward digit span	10.13	9.44	−1.637	0.102
WM backward digit span	7.88	8.38	−1.513	0.130
WM digit span sequencing	8.31	6.93	−3.376	<0.001 ***
PSTM	104.75	87.65	−3.517	<0.001 ***
PSTM span	5.62	4.50	−3.626	<0.001 ***
VSTM	57.73	62.51	−1.989	0.047 *

* p < 0.05, ** p < 0.01, *** p < 0.001.

.

As a Wilcoxon signed-rank test indicated, TD children also demonstrated significant differences concerning their performance in L1 and L2 in specific tasks. In particular, these tasks correspond to the total decoding assessment tool (z = −3.519, p ≤ 0.001) along with the subscales of word decoding (z = −3.465, p ≤ 0.001), irregular word decoding (z = −3.518, p ≤ 0.001) and nonword decoding (z = −2.984, p = 0.003). Moreover, reading fluency (z = −3.259, p = 0.001), word reading per minute (z = −3.517, p ≤ 0.001), and nonword reading per minute (z = −3.517, p ≤ 0.001) exhibited significant differences. PA (z = −3.520, p ≤ 0.001) scores in L1 differed significantly from those obtained in L2. Further, STM performance was different in PSTM (z = −3.517, p ≤ 0.001), PSTM span (z = −3.626, p ≤ 0.001), and VSTM (z = −1.989, p = 0.047), while for WM, significant differences were noticed only in the digit span sequencing task (z = −3.376, p ≤ 0.001). However, the Wilcoxon test did not indicate any differences in reading accuracy (z = − 0.233, p = 0.816), and the cognitive variables of RAN (z = −1.084, p = 0.279), STM as assessed by the forward digit span (z = −1.637, p = 0.102) and WM based on the backward digit span measurement (z = −1.513, p = 0.130).

3.3. DYS Children’s Competence in Comparison to That of TD Children in L1 Reading and Memory Tasks

To achieve our third objective, we used descriptive statistics and the Mann–Whitney U test. Descriptive statistics were calculated for decoding, reading, RAN, and memory, displaying mean values and standard deviations (SD) for both DYS and TD. To determine if there were significant differences in the L1 battery test assessments between children with and without dyslexia, a Mann–Whitney U test was performed on matched pairs. The results for the Greek reading and memory tasks are presented in

Table 4. Mean performance (and SDs) of the DYS (n = 16) and TD (n = 16) group on all L1 (Greek) experimental tasks and Mann–Whitney results for each L1 reading and memory task.

Variables	Group	Mean	SD	Mann–Whitney U	Z	Sig. (2-Tailed)
L1 decoding (percentile)	DYS TD	12.50 60.63	6.58 11.81	0.000	−4.934	<0.001 ***
L1 decoding total percentage	DYS TD	75.32 92.91	3.91 2.19	0.000	−4.834	<0.001 ***
L1 word decoding	DYS TD	72.18 95.73	7.03 2.80	0.000	−4.842	<0.001 ***
L1 nonword decoding	DYS TD	53.91 81.74	9.17 6.73	1.500	−4.783	<0.001 ***
L1 reading fluency	DYS TD	73.19 129.88	13.45 21.07	3.500	−4.694	<0.001 ***
L1 fluency percentile	DYS TD	11.88 66.56	8.34 14.68	0.000	−4.887	<0.001 ***
L1 reading accuracy	DYS TD	324.88 179.06	40.04 28.12	0.000	−4.825	<0.001 ***
L1 word reading per min	DYS TD	56.00 113.25	11.73 21.13	0.000	−4.826	<0.001 ***
L1 nonword reading per min	DYS TD	31.56 66.75	6.62 14.53	0.000	−4.826	<0.001 ***
L1 RAN	DYS TD	66.06 41.31	12.26 4.22	0.000	−4.828	<0.001 ***
L1 PA	DYS TD	58.68 95.26	6.99 3.53	0.000	−4.874	<0.001 ***
L1 STM forward digit span	DYS TD	7.56 10.13	0.964 1.25	8.500	−4.614	<0.001 ***
L1 WM backward digit span	DYS TD	5.81 7.88	0.750 0.885	9.000	−4.599	<0.001 ***
L1 WM digit span sequencing	DYS TD	5.44 8.31	1.41 1.07	7.500	−4.620	<0.001 ***
L1 PSTM nonword repetition	DYS TD	95.50 104.75	2.85 1.61	0.000	−4.848	<0.001 ***
L1 PSTM span	DYS TD	3.81 5.62	0.543 0.500	3.000	−4.919	<0.001 ***
L1 PSTM percentile	DYS TD	17.50 68.44	12.91 14.22	0.000	−4.879	<0.001 ***
L1 VSTM word list	DYS TD	39.90 57.73	5.50 5.72	0.000	−4.854	<0.001 ***

*** p < 0.001.

.

Descriptive statistics were calculated for all variables. The means and SDs showed significantly different results for the two groups. A Mann–Whitney U test was conducted to compare scores between the DYS and the TD group in L1 measurements. The test revealed significant differences in the performance of the two groups. More specifically, the two groups differed significantly in decoding skills (n = 32, Mann-Whitney U = 0, p ≤ 0.001), namely in word (U = 0, p ≤ 0.001) and nonword decoding (U = 1.5, p ≤ 0.001), in reading accuracy (U = 0, p ≤ 0.001) and fluency (U = 3.5, p ≤ 0.001), as well as in word (U = 0, p ≤ 0.001) and nonword reading per minute (U = 0, p ≤ 0.001). Furthermore, the difference between DYS and TD participants reached significance in PA tasks (U = 0, p ≤ 0.001) and in RAN (U = 0, p ≤ 0.001). As it concerns memory tasks, the scores of the two groups were significantly different on all STM measures; thus, the forward digit span (U = 0, p ≤ 0.001), PSTM (U = 0, p ≤ 0.001) with its span evaluation (U = 3, p ≤ 0.001) and VSTM (U = 0, p ≤ 0.001). Likewise, in the WM tasks, the DYS group performed significantly worse than the TD in the backward digit span (U = 9, p ≤ 0.001) and the digit span sequencing (U = 7.5, p ≤ 0.001), respectively. As can be seen, DYS learners obtained significantly lower scores compared to the TD children on all Greek measures. Statistically, a U 0 value indicates complete separation between the two groups; that is, the data in the DYS group were strictly less than all those in the TD group.

3.4. DYS Children’s Performance as Compared to the Competence of TD Children in L2 (English) Reading and Memory Tasks

Descriptive statistics were calculated for the L2 experimental tasks, including decoding, reading, RAN, PA, and memory, to compare the performance of the two groups. Mean values and standard deviations (SDs) for both DYS and TD are accumulated in

Table 5. Mean performance (and SDs) of the DYS (n = 16) and TD (n = 16) group on all L2 (English) experimental tasks and Mann–Whitney results for each L2 reading and memory task.

Variables	Group	Mean	SD	Mann–Whitney U	Z	Sign. (2-Tailed)
L2 decoding total percentage	DYS TD	40.99 72.56	5.97 6.18	0.000	−4.851	<0.001 ***
L2 word decoding	DYS TD	54.68 85.62	7.84 5.51	0.000	−4.838	<0.001 ***
L2 irregular word decoding	DYS TD	25.15 59.06	6.85 8.65	0.000	−4.838	<0.001***
L2 nonword decoding	DYD TD	43.12 72.96	7.33 6.59	0.000	−4.830	<0.001***
L2 reading fluency	DYS TD	40.56 106.34	12.98 23.23	0.000	−4.826	<0.001 ***
L2 reading accuracy	DYS TD	362.81 176.38	41.28 42.37	0.000	−4.825	<0.001 ***
L2 word reading per min	DYS TD	41.43 78.31	7.72 8.94	0.000	−4.827	<0.001 ***
L2 nonword reading per min	DYS TD	25.87 47.31	3.73 4.89	0.000	−4.831	<0.001 ***
L2 RAN	DYS TD	65.69 40.00	8.28 5.21	0.000	−4.829	<0.001 ***
L2 PA	DYS TD	43.14 81.24	4.84 3.39	0.000	−4.865	<0.001 ***
L2 STM forward digit span	DYS TD	6.50 9.44	1.03 1.59	6.000	−4.672	<0.001 ***
L2 WM backward digit span	DYS TD	5.31 8.38	0.793 1.25	0.000	−4.891	<0.001 ***
L2 WM digit span sequencing	DYS TD	4.31 6.93	0.946 1.28	13.500	−4.422	<0.001 ***
L2 PSTM nonword repetition	DYS TD	64.68 87.65	5.39 4.78	0.000	−4.850	<0.001 ***
L2 PSTM span	DYS TD	3.31 4.50	0.478 0.632	25.500	−4.106	<0.001 ***
L2 VSTM word list	DYS TD	40.25 62.51	6.41 6.12	0.000	−4.829	<0.001 ***

*** p < 0.001.

. Non-parametric tests, Mann–Whitney U, were conducted to compare how the two groups performed on all variables and indicate potential differences between their scores. The results of the Mann–Whitney U test have also been compiled and presented in Table 5.

Results indicated significant differences in all English tasks between the DYS and the TD group, defining DYS participants’ performance as poorer and considerably lower than that of the TD group. By scrutinizing these results, the two groups differed significantly in decoding skills (word, irregular, and nonword decoding, U = 0, p ≤ 0.001), regarding reading accuracy (U = 0, p ≤ 0.001) and fluency (U = 0, p ≤ 0.001), word (U = 0, p ≤ 0.001) and nonword reading per minute (U = 0, p ≤ 0.001) and in terms of PA tasks (U = 0, p ≤ 0.001). In addition, the difference between the two groups was statistically significant concerning RAN (U = 0, p ≤ 0.001). Finally, DYS children obtained considerably lower scores on the WISC-V subscales, thus the forward digit span (U = 6, p ≤ 0.001), the backward digit span (U = 0, p ≤ 0.001), and the digit span sequencing (U = 13.5, p ≤ 0.001). The STM measurements of PSTM (U = 0, p ≤ 0.001) together with its span (U = 25.5, p ≤ 0.001) and VSTM (U = 0, p ≤ 0.001) appeared also impaired in the DYS group.

4. Discussion

4.1. DYS Performance in Greek (L1) and English (L2)

Our initial research hypothesis posited that the DYS group would exhibit lower performance in L1 than in L2 in reading, PA, RAN, WM, and STM. We tested this hypothesis by conducting a series of tasks in both Greek and English and upon closer examination of the data from each task, we found compelling evidence in support of this hypothesis, which was partially confirmed. More specifically, the results revealed significant differences between the two languages in word, irregular word, and nonword decoding, reading accuracy and fluency, word and nonword reading per minute, and PA tasks. Distinct scores were also observed in STM, WM, and PSTM. However, our hypothesis was not confirmed in the case of RAN and VSTM tasks, as DYS children’s performance on these tasks in Greek and English did not differ. The study provides essential insights into these findings.

To begin with, the differences in the structure of the two linguistic systems led to statistically significant variations in PA and reading tasks. It is well established that there is a strong linkage between reading skills and the orthography of each language. Although Greek and English are both alphabetic languages, they are characterized by different orthographic transparency. Greek is considered one of the most transparent languages (Seymour et al. 2003), whereas English is generally regarded as having one of the least consistent orthographies (Share 2008; Seymour et al. 2003; Frost 2012). Poor performance on decoding, reading, and PA tasks in English as L2 is in line with the Orthographic Depth Hypothesis, as the irregularities in PA render specific tasks more difficult for the learner (Buetler et al. 2014; Geva and Siegel 2000; Katz and Feldman 1983; Katz and Frost 1992; Schmalz et al. 2015). Consequently, this orthographic distance between L1 and L2 may have hampered our participants’ development of L2 reading skills (Koda 2008). Similarly, low scores on PSTM in L2 may be attributed to the interrelation between PSTM and poor PA skills in our DYS participants. Simply speaking, difficulties in properly developing PA skills can lead to challenges with PSTM, making it difficult for DYS children to create clear and distinct phonological representations (Adel and Saleh 2021).

By the same token, since English is an opaque (non-transparent) language with an irregular orthography and inconsistent relationships between letters and sounds, it becomes more demanding for DYS learners to decode. Such an outcome corroborates Borleffs et al.’s (2019) review, which confirmed that dyslexia is affected by the level of orthographic transparency, as well as Diamanti et al.’s (2018) study, which suggested that the performance of DYS readers in Greek and English separately depended on the orthographic consistency of the language being studied.

Interestingly, against our hypothesis, in the DYS group, indistinguishable results between the two languages emerged for the cognitive measures of VSTM and RAN, with the latter having a slight advantage for the L2 task. However, the RAN score in L2 was influenced by the length of the corresponding object words. To be more specific, all the English words in the assessment tool used were one- to two-syllable words, meaning that they could be named relatively fast. Conversely, Greek words had two to three syllables, so children needed more time to name them than the English words, which were shorter. This finding aligns with Diamanti et al. (2018), who, by examining native Greek and native English DYS children separately, found that competence in RAN tasks in Greek was statistically commensurate with those in English. According to Diamanti et al. (2018), such an outcome underpins a typical profile of competence stemming from the cross-linguistic implementation of pertinent tasks, which seem unhindered by language-specific orthographies, shunting relevant differences and similarities.

On the other hand, the DYS group’s RAN results do not tie well with the hypothesis postulated by Reina et al. (2023), who envisaged that DYS participants’ lexical choice for objects in L2 might have a negative impact on their naming speed. This occurred because our participants appeared familiar with the English words used in the specific RAN task. Additionally, to our knowledge, most of the cross-linguistic studies in the literature focus on the linguistic parts of each language by testing competence in both L1 and L2. In contrast, cognitive tasks are primarily administered only in L1. However, this study is one of the first to assess cognitive tasks in L2 in dyslexia. Our results from the digit span task revealed significant differences across the two languages, consistent with the hypothesis that having two simultaneous inputs, one for each language, makes L2 cognitive tasks more challenging for DYS learners. In brief, we observed that DYS children found the L2 input more cognitively demanding as they spent extra time searching for the requested word (digit) and organizing it in the correct order. Further, low L2 memory results for the DYS group may derive from the verbal stimuli provided in the STM and WM tasks, which place a more significant cognitive burden than the visual stimuli given in the RAN task. Nevertheless, our findings do not seem to align with those of Diamanti et al. (2018) because they did not find any differences in dyslexic participants’ performance in the forward and backward digit spans in English and Greek. The justification here is that we implemented memory tasks in our participants’ L1 (Greek) and L2 (English). Notwithstanding, the authors mentioned above examined the two languages independently by evaluating only native speakers and not L1 and L2 as in our case.

On that note, no differences were detected in DYS participants’ scores in L1 and L2 concerning VSTM. Interestingly, DYS participants scored slightly higher in the English task due to the research design employed in the present study. Although we tried to choose tasks that were comparable in English and Greek, the assessment tools used to measure VSTM were different. The Greek task (Chrysochoou 2006) may be an adaptation of the English Working Memory Test Battery for Children (WMTB-C; Pickering and Gathercole 2001), but since the latter is no longer in use, we failed to administer the same task in L2. Perhaps this correspondence would yield more reliable results, but, unfortunately, a lack of access to this tool led us to replace it with a non-identical one. Therefore, the reason for the higher English VSTM score is that children listened to the given word list multiple times—a characteristic not included in the Greek task, which obviously ameliorated their results.

4.2. TD Performance in Greek (L1) and English (L2)

The second objective of the current study was to compare TD learners’ profiles in terms of reading and memory skills in both languages. Based on our hypothesis, TD children were expected to perform better in Greek (L1) and English (L2) in reading and PA, but they actually obtained similar results on RAN, WM, and STM in both languages. The results demonstrated some important differentiations, which were also noticed in the performance of DYS children. More specifically, significant differences in L1 and L2 were reported in reading, PA, and PSTM, while the rest of the memory measures were undifferentiated. On the surface, this means that our hypothesis is confirmed. In line with the Orthographic Depth Hypothesis, TD’s decoding skills in L2 were less developed due to the lack of orthographic transparency in English. Seeing that our participants were English beginner learners, the inconsistent English orthography may have negatively impacted their reading skills development (Spencer 2000).

However, to our surprise, contrary to our prediction, the word-in-context reading (accuracy) score of the TD group was slightly higher in L2. The main reason for this precedence is the nature of the tasks provided in the current study. In brief, it was found that the Greek task posed a challenge for regular readers due to the high number of low-frequency and unfamiliar words. For instance, participants even asked whether some words were real or not. Conversely, compared to the Greek task, TD children performed well on the L2 reading tool created by the researcher, which featured only a few unusual words. Extensive and continuous exposure to English from a young age may also positively impact this result. Nonetheless, higher scores were obtained for the rest of the reading tasks in Greek. This can be better explained by our participants’ low proficiency in English. Beginner L2 learners of English need more time to catch up to their level in L1, which, as a native language, has a head start. This finding is aligned with Aro and Wimmer (2003) and Caravolas et al. (2013), who found that the development of English reading skills even in native speakers is considerably slower than in other European languages with more consistent orthographies.

Finally, yet importantly and partially consistent with our hypothesis, similar performance of the TD group in L1 and L2 on the forward and backward digit span and the RAN measures lends support to the idea that the exact cognitive mechanisms are activated while executing these tasks in either L1 or L2. More specifically, research has shown that the cognitive skills, such as RAN, that underlie reading development, at least in alphabetic languages, are not language dependent but universal across different languages and orthographies (Caravolas et al. 2013; Moll et al. 2014; Vaessen et al. 2010). Our study fills a gap in the literature by providing data in Greek and suggesting that L2 learning may yield similar cognitive skill results. Additionally, the TD group achieved a slightly higher score on the L2 backward digit span simply because the L1 task was administered first. As a result, upon encountering the task for the second time during the L2 assessment, participants became more familiar with it. However, after attempting the digit span sequencing task, all participants agreed it was the most challenging, particularly in English. This was due to the added cognitive load of manipulating digits and performing arithmetic calculations. Consequently, significant differences emerged between L1 and L2 in this WM measurement. Likewise, lower performance in the PSTM task can be due to the fact that our TD participants have not yet mastered the English phonological rules, which are considered crucial for the PSTM assessment implemented in this study.

4.3. DYS and TD Performance in Greek (L1)

According to the third hypothesis of the present study, the DYS group’s performance in L1 was expected to be weaker in reading, PA, RAN, WM, and STM than that of the TD group. The results provide evidence in support of our hypothesis that DYS children exhibit significant difficulties compared to the TD group in all the examined tasks. As expected, we found that significant differences arose in each reading and memory task. To begin with, DYS children obtained considerably lower scores on the decoding and reading measurements in Greek, displaying essential deficits. This finding is consistent with relevant studies in Greek testing DYS children compared to age-matched children with good reading abilities (Giazitzidou and Padeliadu 2022; Rothou and Padeliadu 2019; Talli 2010; Talli et al. 2016). It is also worth emphasizing that the transparent orthography of the Greek language did not seem to favor DYS participants’ word-level L1 skills. This outcome converges with previous research conducted on Slovenian DYS children, who continued to experience reading difficulties despite receiving literacy instruction in primary school (Slovenian is also a transparent language; Kormos et al. 2019). Similar results were yielded for the PA tasks verifying the Phonological Deficit Hypothesis (Ramus et al. 2003) and in line with previous research in young Greek DYS children (Giazitzidou and Padeliadu 2022; Rothou and Padeliadu 2019; Stampoltzis et al. 2020; Talli et al. 2013, 2016).

In the same vein, the findings of this study provide evidence of memory impairments among DYS children. More specifically, the DYS group’s performance on WM and STM, as assessed by the digit span tasks, was noticeably poorer than that of their matched-TD learners. This discovery aligns with Masoura et al.’s (2021) research, which observed that Greek individuals with dyslexia had lower scores on the forward and backward digit span tasks, as well as on the nonword recall list. We also noted significant differences in these tasks between the dyslexic and non-dyslexic groups. DYS children showed substantial deficits in WM, particularly regarding backward digit span and digit span sequencing scores (Smail et al. 2022). Besides, studies have shown that WM is linked to PA and significantly predicts difficulties in word reading and learning (Alloway et al. 2005; Knoop-van Campen et al. 2018; Gray et al. 2022).

Furthermore, compared to the TD group, our DYS participants experienced significant difficulties with the object rapid naming task. Research in Greek and other languages has yielded similar results (Stampoltzis et al. 2020; Talli et al. 2016; da Silva et al. 2020; Reina et al. 2023). On top of that, Stampoltzis et al. (2020) found that object naming was positively correlated with reading, implying that reading deficits may also be attributed to RAN restrictions. Extensive literature research has demonstrated significant correlations between rapid naming and reading ability. Finally, our findings align with the Double Deficit Hypothesis (Wolf and Bowers 1999), which supports that phonological processing and rapid naming reflect two prominent deficits underlying dyslexia.

4.4. DYS and TD Performance in English (L2)

Our last research hypothesis predicted that DYS children were expected to achieve lower scores on L2 reading, PA, RAN, WM, and STM than TD children. To support this statement, we compared the performance of DYS and TD participants on a battery of L2 tasks, including reading and memory measures. As anticipated, DYS individuals obtained significantly lower scores on all tasks. In particular, regarding PA, decoding, and reading skills, DYS participants exhibited lower performance in L2 due to the lack of transparency in English. This issue also affected TD children but to a lesser extent. The results extrapolated from our findings are in accordance with a considerable number of studies indicating that DYS individuals show lower performance in both L1 and L2 (English) tasks compared to their typical peers (Greek: Andreou and Baseki 2012; Andreou and Segklia 2017; Sotiropoulos and Hanley 2017; Tsesmeli et al. 2021; Slovenian: Kormos et al. 2019; Dutch: van Setten et al. 2017; German: Maurer et al. 2021; Polish: Łockiewicz and Jaskulska 2016; Chinese: Chung and Lam 2020; Li et al. 2018; Tong and McBride 2017; Spanish: Álvarez-Cañizo et al. 2023; Suárez-Coalla et al. 2020; Italian: Bonifacci et al. 2017; Fazio et al. 2021). It is not unreasonable that the aforementioned studies primarily included tasks related to English reading, decoding, and PA. DYS children exhibit significant deficiencies in these skills due to the complexity of English spelling rules (van Setten et al. 2017). It is worth emphasizing that children’s inability to master English grapheme–phoneme representations in this study may be due to their young age, unlike other relevant studies that mainly involve adolescents (Suárez-Coalla et al. 2020).

On the other hand, it is exciting to observe the performance of DYS participants on memory tasks in their L2. To the best of our knowledge, this is the first study, at least in Greek, that examines WM and STM skills of DYS children in L2. The results revealed that DYS children struggle with memory deficits in both languages. Unfortunately, there is limited relevant evidence in the literature since most cross-linguistic studies opt to assess memory tasks in the native language. For instance, Fazio et al. (2021), after assessing DYS children’s WM in L1 (Italian), revealed that WM and phonological skills influence reading accuracy in L2 (English). Consequently, both memory and phonological skills may strongly contribute to poor English reading. Further, our DYS participants’ lower performance on the L2 RAN task conforms with previous research examining RAN in both L1 and L2 (English; Chung and Lam 2020; Li et al. 2018; Reina et al. 2023). According to Chung and Lam (2020), RAN deficits can lead to weaknesses in word reading and spelling in both L1 and L2 in individuals with dyslexia.

Finally, we came to the conclusion that our findings shed light on theories supporting L1 and L2 transfer of difficulties, given that weaknesses detected in DYS children co-exist in both languages. In particular, our outcomes are in line with the LCDH (Ganschow et al. 1991; Sparks and Ganschow 1993; Sparks et al. 1989) and demonstrate that DYS children’s poor reading skills and low competence in L2 derive from relevant deficits previously exhibited in L1. Based on Sotiropoulos and Hanley’s (2017) predictions, it is suggested that individuals with dyslexia employ the same reading routes in Greek and English, which depend on similar foundation skills for reading and spelling. On top of that, as Durgunoglu et al. (1993) highlighted, there is no specific way of developing PA in a language. Still, there are similar processing types underlying two different languages. Moreover, according to the CPH (Cossu et al. 1988; Geva and Siegel 2000; Stanovich 1984), we found that the nature of the orthography is not the only factor affecting reading in L1 and L2. Despite the complexity and regularity differences between Greek and English orthographic systems, our DYS learners’ reading profiles in both L1 and L2 are still impaired. Consequently, as the CPH puts forward, memory, rapid naming, and phonological skills may significantly influence the acquisition of reading skills in L1 and L2, respectively. Previous research conducted by Maurer et al. (2021) found evidence of cross-language associations and transfer between L1 and L2 in cognitive skills such as RAN and PA. So, it can be postulated that difficulties in reading and PA skills may be due to the cognitive mechanisms involved in learning English as an L2 (Sparks et al. 2012).

As a consequence, it is critical to highlight the contribution of this study to the already existing literature concerning dyslexia in L1 and L2. The present study confirms that Greek DYS children demonstrate significant reading difficulties in L1 and in learning English as L2. This finding adds more data to the Greek DYS children’s profile in the field of SLA. However, the assessment of memory skills of DYS and TD Greek-speaking children in English makes this study innovative. To our knowledge, this is the first study, at least in Greek, which examines memory tasks in both L1 and L2 (English). Results reveal that, except for reading and PA deficits, DYS children experience simultaneous cognitive difficulties in RAN, STM, and WM. Therefore, such an outcome can be considered crucial for future research on dyslexia in L1 and L2, with researchers paying equal attention to L2 memory skills as well.

4.5. Limitations and Future Implications

Certain limitations of the present study are worth considering. One issue with the data collection methods was using some non-standardized tests and the lack of equivalence between the Greek and English tasks. Additionally, participants were given only oral language assessments in the domains under investigation, while the written abilities of DYS children are also crucial. The current study addressed the PA, reading, and memory skills of children with and without dyslexia. Still, a wide array of skills should be considered for a broader view of their overall performance in both L1 and L2. A larger sample size could also yield more reliable results. Nevertheless, important implications for research and practice arise from the present study. Our findings demonstrated that Greek DYS children who exhibit significant deficits in L1 continue experiencing relevant difficulties while learning English as L2. In particular, we examined the reading and memory skills of DYS children, which proved impaired in both languages. Lack of prior research on memory and cognitive skills in a second language necessitates further research to examine STM, WM, and RAN in both languages. Future studies, including adolescents and adults with higher proficiency levels, may help us gain greater insight into the potential co-occurrence of reading and memory difficulties in L1 and L2. In a nutshell, our findings support cross-linguistic transfer between L1 and L2 and indicate that children’s cognitive skills influence the acquisition of two languages with different orthographic systems. Apparently, learning English is very challenging for DYS children. When considering dyslexic profiles in both languages, we should also take into account factors beyond the transition of difficulties from L1 to L2.

However, DYS children should not be discouraged from learning English. Implementing tailored interventions as early as possible is critical, including both linguistic and cognitive training. Therefore, in addition to individual intervention programs, in the English L2 classroom, DYS children should follow methods and techniques that combine linguistic and cognitive skills to increase their self-confidence in English lessons. It is common, especially in the English language, to encounter many learning strategies based solely on phonological skills, as they represent the core causes of dyslexia. We thus recommend incorporating activities that help develop reading and memory skills (such as language and memory games), allowing DYS learners to improve in both areas, which are critical for learning an L2. The interdependence of language and memory predicts that improvement in both domains may determine DYS children’s future performance in L2. Further, alternative methods, such as the multi-sensory approach, which engages all senses with multiple practices, can also be included in similar curriculums and become effective for DYS students’ performance in SLA.

5. Conclusions

In the present study, we sought to investigate how children with and without dyslexia perform in a series of tasks in two different language systems: Greek as their first language and English as their second. Our initial objective was to compare DYS participants’ performance in L1 and L2. The results revealed significant differences in most tasks, including word and nonword decoding, reading accuracy and fluency, PA, STM, and WM. The DYS group, however, obtained similar scores on the RAN and VSTM tasks in both languages. Secondly, we examined TD participants’ performance in the aforementioned tasks in L1 and L2. Results indicated that the TD group did not differ in L1 and L2 reading, STM, WM, and RAN measurements. Significant differences for the TD groups characterized all the remaining domains. Last but not least, we separately evaluated DYS’s performance compared to TD’s on the abovementioned reading and memory domains in L1 and L2. Between-group comparisons indicated that DYS children face essential difficulties in both languages across all tasks.

Overall, we conclude that DYS children struggle with reading in both Greek (L1) and English (L2), despite the different orthographic systems, while they also exhibit significant memory deficits in both languages. However, it is important to acknowledge that additional factors could have contributed to substantial differences both between and within groups’ performance, such as socioeconomic status (SES), exposure to English (L2), and individual cognitive differences.