THE ROLE OF L1 AND AGE EFFECT IN THE ACQUISITION OF ENGLISH ARTICLES AMONG CROATIAN YOUNG EFL LEARNERS

The category of definiteness/indefiniteness is often seen as a linguistic universal (Chomsky 2000; Silić 2000) that exists in every human language but is expressed with different linguistic means and patterns in different languages. In English, this category is grammaticalized primarily via articles (Zergollern-Miletić 2008). The Croatian language does not contain an article system, and acquiring English articles represents a major problem for most learners of English as a foreign language (Balenović and Medved Krajnović 2013; Zergollern-Miletić 2015). The present paper investigates difficulties in acquiring English articles among young learners in Croatia in written tasks. Pupils involved in the research started learning English in the first grade of primary school. A total of 76 primary school learners from six different classes (schools) and different age groups were included in this research. The initial hypothesis is that the correct use of articles increases as learners’ overall L2 proficiency increases due to exposure to instructed EFL learning. A qualitative analysis was carried out using error analysis, i.e., proper/wrong use, or omission of the article. The findings partially confirmed the initial hypothesis and showed that our participants were becoming aware of the existence of English articles (they comprehended that articles should precede nouns), but we found very frequent article omission, which, according to our research results, indicates that the process of developing and understanding the semantic and pragmatic role of English articles is long and fluctuating specifically with learners whose L1 lacks the article system.

indicates that the process of developing and understanding the semantic and pragmatic role of English articles is long and fluctuating specifically with learners whose L1 lacks the article system.

Introduction
The role of a native/primary/first language (L1)/mother tongue in second language research has been an important topic (Ellis 1994) since all humans have an "innate ability" (Chomsky 1957) to acquire a language in the first years of life.This innate ability/knowledge is, in what Chomsky calls, the language faculty, which is "a component of the human mind, physically represented in the brain and part of the biological endowment of the species" (Chomsky 2002: 1) known as a "language acquisition device" (LAD) guided by Universal Grammar (UG).Contrastive Analysis (CA), founded on the behaviouristic theory of learning, was the first approach which pointed out that elements acquired in the L1 will be transferred to the L2 resulting in a positive or negative influence on the learner's L2 competence.In other words, the primary cause of errors in the L2 is negative transfer or interference from the L1.According to this approach, a comparison of the native and second language is needed to isolate differences between the two languages (Saville-Troike 2008).The easiest L2 structures to acquire are those that exist in the L1 with the same form, meaning, and distribution, while among the most difficult to acquire are the ones with a partial overlap in form, meaning, and/or distribution (ibid.).In other words, when two languages are similar, a positive transfer occurs (Ekiert 2004;Park 1996;Snape et al. 2013); while, when they differ, a negative transfer happens (Chen 2000).Moreover, L2 learners face more problems with structures absent in their L1, such as Croatian learners of English as a foreign language (Croatian EFL learners) acquiring English articles.In the Croatian language, the category of definiteness/indefiniteness is grammaticalized differently -via adjectives, quantifiers, or determiners (Zergollern-Miletić 2014).In terms of definiteness/indefiniteness, Croatian EFL learners have full access to UG as an innate guide to L2 acquisition matched with information from an input (e.g., a teacher).In addition, Ionin, Zubizarreta, and Maldonado (2008) found that L1 transfer, L2 input, and UG as sources of knowledge were related to the acquisition of English articles by L2 learners.When examining L1 transfer in article acquisition, Sharma (2005) found that L1 transfer was obvious when L2 speech should contain the specific indefinite article, which was often omitted among learners whose L1 lacks articles.
As mentioned above, L2 learners of English often have difficulties mastering articles, especially when their L1 lacks an article system, which may suggest that they are unaware of the existence of the notion of definiteness/indefiniteness in their L1.This category is often seen as a linguistic universal (Chomsky 2000;Silić 2000) but expressed with different linguistic means and patterns.In the English language, this category is primarily grammaticalized in the articles (Zergollern-Miletić 2008), as well as in semantic notions of definiteness, specificity (Ionin, Ko andWexler 2004), uniqueness (Yang andIonin 2009), and nominal countability (Ogawa 2008).
The notion of definiteness refers to the knowledge shared between the speaker and the hearer, while specificity refers to the knowledge known only to the speaker.The notion of definiteness is crucial in determining article choice for L2 learners of English (Ionin 2008).For example, I saw a girl.The girl was walking down the street.The choice of the indefinite article indicates that there is no shared knowledge of the girl between the speaker and the hearer, while in the second sentence, the hearer realizes that the girl is the same as in the previous sentence.As mentioned above, the notion of specificity refers to the knowledge known only to the speaker, for example: (1) I want to see a film.Any recent film will be fine.(2) I want to see a film.It's on the top list.The use of the indefinite article in the first example indicates that it is the first mention of the film by a speaker in the discourse.However, the second sentence shows that the speaker has a specific film in mind that he/she is referring to.The indefinite article in this example receives a referential interpretation (Fodor and Sag 1982).According to Hawkins (1978), the category of definiteness is related to the identifiability of referents in a discourse set by multiple ways of establishing uniqueness (e.g., previous mention, association, visible situation, or modifying clause).Lyons (1999) assumes that identifiability which is seen as semantic/pragmatic definiteness is a universal cognitive category.This assumption presupposes that articleless languages such as Croatian have different linguistic means (e.g., other determiners or quantifiers) to mark something as identifiable within a discourse set.This semantic complexity of English articles makes it difficult for L2 learners to master the use of articles, especially for ones whose L1 lacks an article system.
Definiteness and specificity are notions important for the Article Choice Parameter and the Fluctuation Hypothesis proposed by Ionin, Ko and Wexler (2004).This hypothesis presupposes that L2 English learners whose L1 lacks articles fluctuate between the semantic universals of definiteness and specificity because there is no L1 transfer but full access to UG.These L2 learners fluctuate between two settings until the input leads them to the right option.In these cases, L1 interference certainly plays a significant role (Liu and Gleason 2002).However, Ionin, Zubizarreta and Maldonado (2008) showed that fluctuation overrides language transfer with L2 learners whose L1 does not have articles (e.g., Russian).L2 English learners whose L1 has articles (e.g., Spanish) should transfer article semantics from their L1.
In modern English grammar textbooks, articles are set with other determiners as a closed system, i.e., put into the same category with other grammatical elements (e.g., demonstratives, possessives, quantifiers) that express the category of definiteness/ indefiniteness in a noun phrase.They sometimes overlap in meaning; for example, the demonstrative pronoun can be used instead of the definite article (Hawkins' location theory, 1978).In the Croatian language, this category is expressed by other linguistic means such as determiners (Pranjković 2000;Silić 2000) or quantifiers (Šarić 2002).Silić (2000) also points out that, in the Croatian language, this category is expressed with definite and indefinite adjectives (e.g., visok čovjek (a tall man), visoki čovjek (the tall man).In other words, an indefinite form of an adjective precedes nouns unfamiliar to interlocutors, while a definite adjective precedes nouns familiar to interlocutors.There is still no consensus on what could be the grammatical forms by which the categories of definiteness and indefiniteness can be expressed.The lack of articles in learners' L1 presupposes constraints and difficulties in article choice in L2 acquisition.

The effect of l1 and the age factor in the acquisition of the English article system
Numerous researchers have addressed the problem of the acquisition of English articles by learners whose mother tongues do not contain an article system (Huebner 1983;Parish 1987;Ionin, Ko and Wexler 2004;Hawkins et al. 2006;Trenkić 2000Trenkić , 2004Trenkić , 2008;;Zergollern-Miletić 2008;Martinović and Balenović 2019, among others).In most of the studies, the major problem in mastering English articles was article omission and substitution.Zdorenko and Paradis (2008) questioned whether different L1 backgrounds (-article L1, +article L2) influenced child L2 acquisition of English articles.They found that L1 background did not exert much influence on children's accuracy with article choice, with the exception at the earliest stages of L2 acquisition, since L1 transfer was likely to be the reason for lower accuracy rates of article omission in L1 articleless group (Zdorenko and Paradis 2008).Master (1990Master ( , 1997) ) claimed that L1 articleless speakers needed more time to acquire the article system than speakers whose L1 contains articles.Park (1996) also found that students whose L1 contains articles (French, German) earned higher scores than students whose L1 lacks articles (Japanese, Korean).Trenkić (2002) found that EFL learners whose mother tongue lacks articles (Serbian) substituted the indefinite article in favour of the definite in most cases, especially with noun phrases referring to concepts with a relatively constant form.Martinović and Balenović (2019) found that most errors with article usage among Croatian university students included omission of the indefinite article.Vickov and Jakupčević (2020) also found that articles were the most frequent source of error for Croatian learners in writing thesis abstracts, especially article omission.Patekar (2017) conducted a study among Croatian primary school EFL learners aiming to describe the influence of Croatian as a native language on the written output in English as a foreign language.The research results showed that the influence was recorded at all linguistic levels, and the most errors were errors in the omission of both types of articles (Patekar 2017).The omission of the indefinite article was also found in other studies (Robertson 2000;Jarvis 2002) to be a crucial source of problems related to article usage (Martinović and Balenović 2019).Articles are frequently used in the input EFL to which learners are exposed in naturalistic and instructed settings and at all levels of EFL learning.Acquiring the proper use of articles is a very long process, even more so for learners whose L1 lacks articles (e.g., Croatian).
Studies on the acquisition of English articles in naturalistic L2 settings have found that age plays a role, giving an advantage to early starters (Zdorenko and Paradis 2008), while no early starters advantage has been found in formal instructional settings (Ekiert 2004).On the other hand, Kim and Song (2020), in their research on the acquisition of English articles by Korean EFL learners, found that late starters were less sensitive to the use and semantic features of English articles than early starters.The findings of their study suggested that consistent exposure to EFL learning at an early age positively affects learners' ability to process English articles.In addition, when it comes to formal L2 learning, children show variations in L2 acquisition depending on individual differences which include both child-internal (e.g., age factor, L1 background, cognitive capacities) and child-external (e.g., input, learning environment) factors (Paradis 2019).In other words, when discussing the role of age in L2 learning or when questioning the best onset of L2 acquisition (e.g., popular belief: the younger, the better; or the younger -better in the long run), we should take into consideration all the abovementioned factors in child L2 learning.When it comes to L2 article acquisition, some authors pointed out that the early stage of EFL learning in non-article L1 is marked by article omission (Zdorenko and Paradis 2008), followed by substitution, usually in favour of the definite article (Trenkić 2002).Furthermore, Zdorenko and Paradis (2008) found misuse of articles among children in both article L1 and non-article L1, which led to the conclusion that children, in contrast to L2 adults, do not transfer their L1 knowledge but instead use their access to UG no matter their linguistic background.Morales (2011) found that children were able to transfer their L1 semantic knowledge to the L2 since they did not show errors in article omission as a common error among learners whose L1s have no article.Taking these controversial study results into account, we might conclude that both L1 transfer and UG as sources of knowledge together with children's differences are related to the child L2 learners' acquisition of English articles.
Most of the previously mentioned studies dealt with the acquisition of English articles in the later stages of EFL learning or with different L1 backgrounds.There have been very few studies on the acquisition of English articles with the Croatian L1 background.The part that follows includes studies conducted among young Croatian EFL learners.

Previous studies on the acquisition of English articles by Croatian l1 primary school learners
As mentioned above, understanding and mastering the English article system represents a significant problem for learners whose mother tongue does not contain articles.This problem is even more significant in early EFL learning when young learners start learning both Standard Croatian and English as a foreign language at the same time, i.e., in the first grade of their education (ages six/seven).
Zergollern-Miletić (2015) explored the acquisition of the markers of definiteness/ indefiniteness by Croatian primary school learners who started learning English in the fourth grade.The study was conducted longitudinally within the project Early acquisition of English as a foreign language: the analysis of the learner's interlanguage.A sample of 24 Croatian EFL learners from four classes (two different schools) was included in the research.Participants' oral and written production was observed and analysed throughout four years, i.e., from Grade 5 (ages 10/11) to Grade 8 (ages 13/14).Oral production was elicited by pupils' responses to questions or by describing a picture presented to them.They were also asked to write a composition on a given task by which their written production was analysed.In the study, the author expected that Croatian young learners' production would show a general omission of articles in the early stages of EFL learning.The research results showed fewer problems than expected.Learners omitted articles, but not in as many cases as assumed.They seemed to understand that, in the English language, determiners should precede nouns, so they even used English articles with Croatian nouns when they could not remember the word in English.Some learners did not use any article before countable singular nouns and used the indefinite article before plural nouns.There was also a higher rate of article omission in the written tasks than in the oral tasks.However, the results in the written tasks were slightly better than those in the oral tasks in Grade 8, i.e., the proper use of articles has improved over the years of EFL learning.The analysis also showed that the notions of familiarity, countability, and uniqueness had been internalized to a certain extent, but the author concluded that young Croatian EFL learners had not entirely mastered the use of English articles, for example the use of articles before abstract nouns.This was in line with the previous findings (Zergoller-Miletić 2008, 2009) with very advanced learners of English, who also had problems with the use of articles with abstract nouns.
Another study (Balenović and Medved Krajnović 2013) conducted within the framework of the abovementioned project included two groups of Croatian L1 primary school learners.The first group consisted of 12 learners (mean age 11-13) who started learning English in the fourth grade (ages 9/10).The second group in the study (Balenović 2014) consisted of 93 learners (mean age 7-9) who started learning English in the first grade (ages 6/7).They were observed over three years.The research was based on the recordings of individual oral production tasks.The oral productions were recorded, transcribed, and analysed by using CLAN (Computerized Language Analysis), the tools developed for the analysis of oral production as part of the CHIL-DES (Child Language Data Exchange System) project (for details, see MacWhinney 2008MacWhinney , 2010)).The research aimed to analyse the process of the acquisition of English articles in primary school learners.The authors expected that the correct use of the English article system would gradually increase as learners' overall L2 proficiency increased, i.e., over the years of EFL learning.
The quantitative analysis showed an increase in participants' overall use of articles and the correct use of them.However, there was also an increase in the omission of articles as well.Overall, the indefinite article was used with higher accuracy (in 65% of obligatory cases) than the definite article (in 59% of obligatory cases) The higher accuracy of the indefinite article was also found among young learners who started learning English in the first grade (Balenović 2014).This is in contrast with some other studies (e.g., Ionin 2003) that claim that the definite article is acquired before the indefinite article.This could be explained by the earlier exposure to the indefinite article in the classroom context.Qualitative analysis showed that the learners mastered the use of articles in a familiar context, which aligns with Zergollern-Miletić's (2015) findings in which she claimed that the notion of familiarity was internalized to a certain context.Both groups of learners showed a frequent omission of the use of the indefinite article in front of singular countable nouns mentioned for the first time, especially the ones not familiar to them, which pointed to the conclusion that there was an interdependence of lexis and grammar.This interdependence was also confirmed in Trenkić's (2002) and Ionin's et al. (2008) studies.Sometimes, the number one was used instead of the indefinite article by both groups of learners.The result could be explained by the transfer from learners' L1, i.e., the number one, in some contexts, can be used to express indefiniteness in learners' L1.There was also article omission in the adjective + noun constructions, probably due to learners' perception that the noun is already defined.Another interesting example of the indefinite article omission is at the beginning of the sentence, but the correct use of the definite article in the same sentence (e.g., * book is on the shelf).This was also confirmed in Trenkić's (2002) and Jarvis ' (2002) studies.The finding might be explained by the learners' concentration only on one linguistic problem since this is the beginning of EFL learning.
From the analysis of the abovementioned studies on the acquisition of English articles in the Croatian context, we might conclude that there is a gradual but very dynamic and variable progress in the acquisition of the English article system in early EFL learning.The analysed studies mostly dealt with article acquisition among Croatian L1 learners who started learning English in the fourth grade, so there has been a need to explore the use of English articles by early starters.Moreover, there have been very few studies on L2 article acquisition among Croatian L1 learners who started learning English in the first grade so far.

The present study -aim and methodology
Our research aimed to analyse the use of English articles by primary school learners in written tasks and to compare our research results with the previous studies on article usage in learners' oral production (Balenović and Medved Krajnović 2013;Balenović 2014) aiming to find out whether there are any distinctions in the use of English articles in different learner's performances (oral vs. written) and in different age groups (fifth vs. eighth grade learners; ages 10/11 vs. ages 13/14).The following questions are to be answered: 1. Is there any difference between fifth and eighth-grade learners in article acquisition due to the exposure to instructed EFL learning?2. What are the most frequent errors in using English articles by young Croatian EFL learners in written tasks?

In what ways does the L1 without articles influence article acquisition?
The initial hypothesis is that the correct use of articles increases as learners' overall L2 proficiency increases, i.e., due to exposure to instructed L2 learning for a longer period, the correct use of articles also increases.slightly higher accuracy (in 48% of obligatory cases) than learners in Grade 5 (in 46% of obligatory cases/N=119).A one-way ANOVA confirmed a significant effect of the grade students attended on correct article usage, F(1,74) = 21.858,p < 0.001.The effect size, eta squared (η²), was 0.23, indicating a large effect.Students attending the higher grades also tended to have better overall scores in terms of correct article usage.Learners in Grade 8 omitted articles less (in 42% of obligatory cases/N=148) than learners in Grade 5 (47% of obligatory cases/N=122).They misused articles (article substitution) in very few (10%/N=35) obligatory cases in Grade 8 as well as in Grade 5 (7%/N=18).As seen from the results, learners in higher grades had better overall scores on the correct usage of both types of articles, which had been expected since they were exposed to instructed EFL learning for a longer period.The total distribution of English article usage is given in the Figure that follows.
Multiple regression analysis was used to test if the error types and the grades students attended significantly predicted their correct usage of articles (see Table 1).The results of the regression indicated that the two predictors explained 60.4% of the variance (R2=.604,F(3, 72) = 36.56,p < .01).It was found that the omission error type significantly predicted correct article usage (β = .75,p<.001), as did the grade which students attended (β = .24,p<.001).The error type of substitution was not found as a significant predictor of how accurately students employed articles in their writing composition.We might conclude that students in both grades preferred omitting articles to using them incorrectly, which is in line with other studies on child L2 article acquisition (e.g., Ionin 2009).Furthermore, in our research, learners in Grade 8 also used the indefinite article (Figure 2) with slightly higher accuracy (in 65%/N=109 of obligatory cases) than learners in Grade 5 (in 59%/N=70 of obligatory cases) with very few cases of wrong usage (30%/ N=11).Learners in Grade 5, when using, used the indefinite article incorrectly in 21% (N=4) of the obligatory cases.Learners in Grade 5 omitted using the indefinite article more (in 79%/N=96 of obligatory cases) than learners in Grade 8 (in 81%/N=120 of obligatory cases).When it comes to indefinite article usage, the result is expected and confirms our hypothesis, i.e., learners who are exposed to instructed EFL learning for a longer period become aware of the need to use articles before nouns.

Figure 2 Error analysis of the indefinite article (a/an) usage
In terms of proportions, the definite article (Figure 3) seems to have been used with slightly higher accuracy (in 55 %/N=49 of obligatory cases) by learners in Grade 5 than by learners in Grade 8 (53%/N=59).Learners in Grade 8 omitted the definite article in very few cases (in 25%/N=28 of obligatory cases), and so did learners in Grade 5 (in 29%/N=26 of obligatory cases).Learners in both grades used the definite articles incorrectly in very few cases (16%/N=14 of obligatory cases in Grade 5; 22%/N=24 of obligatory cases in Grade 8).The result confirmed that our participants in Grade 5 were becoming aware of the existence of English articles, especially the definite one, which proved that learners in Grade 5 acquired notions of familiarity and uniqueness to a certain level, but, unexpectedly, learners in Grade 8 made more errors, which contrasted with our hypothesis.Overall, a one-way ANOVA revealed a significant effect of the grade on correct indefinite article usage, F (1,74) = 33.286,p < 0.001, indicating a large effect size, eta squared (η²= 0.31), whereas when it comes to the correct use of definite articles, the ANOVA analysis showed no significant differences with respect to the grade students attended.This finding might be related to definiteness and specificity, the notions important for the Article Choice Parameter, by which Ionin (2004;2007) explains variability and inconsistency in L2 learners' production of articles.
As seen from the above data analysis of article usage, we might conclude that learners used the indefinite article with slightly higher accuracy than the definite article.Our results obtained from analysing written tasks are in line with previous findings  (Balenović and Medved Krajnović 2013;Balenović 2014) on the acquisition of articles in learners' oral production.The findings could be related to higher frequency use of the definite article, which presupposes a higher rate of errors.That is probably why the result is in contrast with some studies that claim that the definite article is acquired before the indefinite article (e.g., Martinović and Balenović 2019;Hawkins et al. 2006;Ionin 2004).The omission of both articles was also found in our study, as it was found in other studies on article acquisition in learners whose L1 lacks articles (e.g., White 2003).
Article substitution was present with learners in both grades, for example, when learners used the definite article instead of the indefinite before nouns mentioned for the first time (e.g., *The boy is playing football) or the indefinite article instead of the definite when placing a referent in the context (e.g., There are nine rocks in *a park).
The qualitative analysis showed article omission before singular countable nouns mentioned for the first time (e.g., There is *boy in the picture; * Girl is skating; There is *book on the bench), as well as the use of the number one instead of the indefinite article (e.g., *One boy is driving a bike).As seen from the example, the learner correctly used the definite article.Still, in some instances, learners omitted definite articles when placing a referent in a defined space, or they substituted articles (e.g., *The man is in *park).This incorrect usage/substitution is in line with the previous findings (Balenović and Medved Krajnović 2013;Balenović, 2014).
Demonstrative pronouns were also used in some contexts instead of articles (e.g., *This boy is riding a bike; *That boy is playing football).This could be attributed to the influence of learners' L1, which lacks articles, and in which a demonstrative pronoun may be used to express definiteness.Incorrect use of the definite article was found in front of the nouns denoting family members (e.g., *The mum is sitting on the bench).Learners probably do not understand that the word mum already implies definiteness and uniqueness, and it is superfluous to further define it with the definite article.The omission of the definite article was also found when placing a referent (object) in a discourse set (e.g., Children are talking together and sitting in *a park), which was found in previous findings (Balenović and Medved Krajnović 2013).This can be explained by the influence/interference of learners' first language since in the Croatian language there is no need for expressing/placing a referent (object) into a defined context.
The research questions addressed by our study concerned the age effect on the acquisition of English articles, the most frequent errors in using English articles, and the ways in which the articleless L1 influences the L2 article acquisition.When it comes to the age effect, the study revealed that learners attending higher grades tend-ed to have better overall scores in terms of correct article usage.Learners attending lower grades omitted articles more often (especially the indefinite article) than those who were at more advanced levels of language acquisition.The indefinite article was mostly omitted in front of singular countable nouns mentioned for the first time.As far as the definite article usage is concerned, the study showed no significant differences related to the grades learners attended.This result contrasted with our hypothesis which assumed that the correct use of articles would increase due to longer exposure to L2 learning.The use of the number one instead of the indefinite article was found to be very frequent, as well as the use of demonstrative pronouns, which might be related to the influence of learners' L1 which lacks articles, so learners resort to the grammatical words in English whose counterparts exist in Croatian to express definiteness or indefiniteness.Article substitution was also frequent in our study, and so was a higher frequency of the definite article in the indefinite context.
The results from this study are very much in line with the results of article usage in learners' oral production from the previous studies (e.g., Balenović and Medved Krajnović 2013) among learners who started learning English in the fourth grade.That leads us to the conclusion that the early age of L2 onset was not associated with faster rates of L2 acquisition of English articles.In addition to an early start, other factors (e.g., motivation, attitudes, aptitudes, learning environment, exposure to instructed learning, etc.) might lead to the result that late starters achieve as good results in L2 learning as early starters (Hyland 2019).This brings us to the crucial question of whether it is beneficial to start L2 learning early in childhood.In their study on early EFL learning, Erk and Ručević (2021) questioned the benefits of an early start.They found no significant difference in learners' performance between very young (pre-primary beginners) and young (primary beginners) learners.When it comes to the acquisition of English articles, our study does not prove that starting to learn English early necessarily brings good results.Frequent exposure to both types of articles and constant repetition of structures containing articles might help.

Conclusion and further implications
The study aimed to explore the role of the learner's L1 and age effect in the acquisition of the L2 English article system.Study results indicated that L2 English learners whose L1 lacks articles exhibited problems in their acquisition.The study revealed the frequent omission of the indefinite article and the evidence for the transfer from L1 to L2 in expressing indefiniteness (e.g., the use of the number one instead of the indefinite article).Furthermore, the study also revealed differences in the correct use of the indefinite article in favour of learners in higher grades, while no significant differences in the correct use of the definite article were indicated concerning the grades students attended.We also found article substitution and omission of the definite article when placing a referent in a defined place, which is in line with previous studies (Balenović and Medved Krajnović 2013;Balenović, 2014).Furthermore, Croatian young EFL learners master the indefinite article with higher accuracy than the definite article in the present as well as in previous studies (ibid.), and it is in line with the results found in Sun's study (2016) conducted among college students learning English as a second language in the United States.Such research results are in contrast with other studies that claim the definite article is commonly acquired before the indefinite article (e.g., Ionin 2004).The reason why Croatian young EFL learners acquire the indefinite article before and with higher accuracy might be in the curriculum according to which they are exposed to the indefinite article before the definite one.The analysis of the studies on article acquisition in early EFL learning revealed inconsistency and difficulty in article usage, which indicated a long and variable acquisition process at all levels of EFL learning, specifically when learners' L1 lacks articles (Robertson 2000;Jarvis 2002;White 2003;Ekiert 2007;Sarko 2008).Therefore, there is a need to develop Croatian learners' awareness of the existence of the notions of definiteness/ indefiniteness in their first language (Martinović and Balenović 2019) and to compare Croatian and English structures dealing with these notions in explicit classroom teaching/learning settings for the more effective theoretical and practical acquisition of English articles.

Figure 1
Figure 1 Error analysis of overall English article usage

Figure 3
Figure 3 Error analysis of the definite article (the) usage

taBle 1
Regression analysis