THE INCIDENCE OF NATIVELIKENESS IN SECOND LANGUAGE ACQUSITION AFTER THE CRITICAL PERIOD HYPOTHESIS: A COMPARATIVE RESEARCH ANALYSIS

(1)

Universiteit van Amsterdam Faculteit der Geesteswetenschappen MA Linguistics: Language and Education Jacqueline Mattie / 11668512

Supervisor: A.P. Versloot Second reader: J.H. Hulstijn Date: 26 July 2018

Master’s Thesis

THE INCIDENCE OF NATIVELIKENESS IN

SECOND LANGUAGE ACQUSITION AFTER THE

CRITICAL PERIOD HYPOTHESIS:

A COMPARATIVE RESEARCH ANALYSIS

(2)

Abstract

This paper examines the incidence of nativelikeness in second language acquisition (SLA) of late-learners after the critical period hypothesis (CPH). The CPH, proposed by Lenneberg (1967), is the period of time that is set to begin early on in life and end around puberty, during which certain behaviors or skills appear to be acquired with particular ease and speed, and without external intervention. While many studies have been conducted on nativelikeness in realm of SLA, particular studies have concluded that nativelike competence in SLA can indeed be achieved by some individuals, while others argue that nativelikeness is overall unattainable. One of the main goals of this comparative research analysis is to determine whether or not late second language learners can achieve comparable results to that of native speakers. Furthermore, this study examines the individual variables that contribute to the ultimate success, or lack of success, in a late-learner’s attainment of their L2 and why these variables are important in effecting these outcomes. Lastly, these variables are examined in the context of the relationship between the linguistic domain and the language distance of the L1-L2 pairing with regard to the overall results of nativelike competence of post-puberty learners.

(3)

1. Introduction

The incidence of nativelikeness has been a long-debated and controversial issue in the realm of second language acquisition (SLA), the scientific process by which people learn a second language. Numerous studies have been conducted on researching the incidence of nativelikeness, which refers to the quality of using a language and performing like a native speaker of that language, in adult SLA. One of the main goals of studying nativelike competence in adult SLA is to determine whether or not late second language learners can achieve

comparable results to that of native speakers. However, the results on these studies of nativelikeness have been extremely varied and have not yet given one conclusive answer to whether or not nativelikeness can actually be obtained. While many studies have been conducted on nativelikeness, particular studies have concluded that nativelike competence in SLA can indeed be achieved by some individuals, while others argue that nativelikeness is unattainable. According to linguists Niclas Abrahamsson and Kenneth Hyltenstam (2009: 249) from

Stockholm University in Sweden, “Although some researchers claim that any learner, regardless of age of acquisition, can attain nativelike levels of second language, or L2, proficiency, others hold that attainment of nativelike proficiency is, in principle, impossible”.

The study of nativelikeness has also been linked to other realms of SLA research such as the Critical Period Hypothesis (CPH), a claim made by linguist and neurologist Eric Lenneberg (1967) that has turned into a long-standing debate. In theory, the CPH questions the limits on maturation of the brain and asks if there is a critical period in the brain for when language can be acquired that begins at infancy and ends at puberty. Researchers J. Johnson and E. Newport (1989: 60) state that in its basic form, the critical period hypothesis need only have consequences for first language acquisition, but nevertheless, it is “essential to our understanding of the nature of the hypothesized critical period to determine whether or not it extends as well to second language acquisition”. The ultimate question of SLA in reference to the CPH asks the question of whether SLA is constrained by the maturation of the brain, and furthermore if it is even possible for adult learners to obtain nativelike results in their second language after the critical period.

This paper sets out to examine the individual variables that contribute to either the

successful achievement, or the lack of achievement, of nativelike competence compared to native control groups in a second or foreign language within the frame of Lenneberg’s (1967) CPH. The

(5)

aim of this research analysis is to determine which individual variables contribute to the ultimate success, or lack of success, in a late-learner’s attainment of their L2 and why these variables are important in effecting these outcomes. In addition, the study addresses the questions under which conditions nativelikeness can be achieved and in which linguistic domains.

This paper is organized into a total of 9 sections. Following is the theoretical framework section, which discusses major SLA premises including the critical period hypothesis (CPH), defining nativelikeness, the stability hierarchy and usage-based language and grammar.

Afterward is the methodology section, which breaks down each individual variable in relation to the research conducted on nativelikeness and examines and explains how the evidence is

evaluated and weighted. Next are two sections with case studies conducted by researchers on the incidence of nativelikeness that either support or reject the claim that nativelikeness can be achieved after the critical period of language acquisition. The first section, research against nativelikeness, claims that nativelike competence of late second language learners after the CPH cannot be achieved, and the second section, research supporting nativelikeness, claims that nativelike competence of late second language learners after the CPH can indeed be achieved. The next two sections of the paper, the discussion and conclusion sections, examine the overall trends from the previous data, provide an interpretation of the weight of the individual variables in the case studies, observe which variables have the most impact on late SLA, and discuss the overall outcomes and findings of this comparative research analysis. Finally, there is a short summary, or Zusammenfassung der Arbeit, of this research analysis written in German, which is then followed by the complete list of references of work used throughout this paper.

2. Theoretical framework

2.1. Critical Period Hypothesis

Discussions of nativelikeness and the role of age in successful second language acquisition (SLA) often begin with the notion of a critical period and the Critical Period

(6)

speed, and without external intervention”. This period of time, when language develops readily, begins early on in life and ends around puberty. Based on neurological and clinical evidence, Lenneberg (1967), who is credited with proposing the CPH, set the critical period of language learning between approximately two years of age and the end at puberty, or around 13 years of age (Hummel 2014: 171). Hummel (2014: 176) writes that Lenneberg had observed that “automatic acquisition from mere exposure to a given language seems to disappear [after puberty], and foreign languages have to be taught and learned through a conscious and labored effort”. Using this observation, Lenneberg proposed the Critical Period Hypothesis (CPH), which is then the theory in SLA that the critical period favors language acquisition, after which a

native-speaker-like ability cannot be attained (Hummel 2014: 170).

Although it is a controversial subject in SLA research, many researchers support the existence of a critical period and the validity of the CPH. In their study, Bongaerts et al. (1997: 447) claim that “beyond this neurologically based critical period, ending around the onset of puberty, complete mastery of a language, first or second, is no longer possible”. Other

researchers, such as Snow & Hoefnagel-Höhle (1978: 1114), also support the existence of the CPH and agree that “a commonly drawn, though not absolutely necessary, corollary of the CPH is that any language learning which occurs after the age of puberty will be slower and less successful than the normal first language learning”. However, it must be made clear that not all researchers agree on this controversial issue, and there is still considerable evidence of

exceptional learners that have indeed achieved nativelike competence after the critical period. There are also other factors beyond age that are thought to influence language

acquisition. The Universal Grammar (UG) view is a linguistic theory on the genetic component of language, usually credited to American linguist Noam Chomsky (1969), that views language as unique and different from other cognitive systems. By definition, Universal Grammar is the innate principles and properties that characterize the grammars of all human languages (Hummel 2014: 14). Chomsky (1969) claims that these innate principles build a ‘language acquisition device’ (LAD) in the brain, which enable learners to construct a grammar out of 'raw input' collected from the environment. As a theoretical approach, UG looks at the biological basis of language learning and suggests that humans are born with an innate set of rules about the

grammar of a language as “a part of their genetic endowment” (Hummel 2014: 14). According to Hummel (2014: 15), there is also considerable evidence that humans from the start are

(7)

“prewired” to acquire language that support the UG view for the biological basis of language, such as how a child can develop his or her first language without receiving nearly enough

information for grammatical processing or possible sentences of the language by exposure alone. Although Chomsky’s UG and LAD theories seem accessible to learners of a first

language, the key question here asks if these theories are still accessible to learners of a second language. Linguists argue that L2 and L1 acquisition are different and that both stages of acquisition do not go through the same phases. In relation to UG under the CPH, although all of language may be governed by UG, older learners might have great difficulty in gaining access to the target language's underlying rules from positive input alone (Singleton, 2005).

Lastly, it is important to note that the CPH and UG are simply hypotheses and do not dictate the laws of SLA. Scholars have experienced and continue to experience today substantial variations in findings adhering to the CPH. While there are researchers that support the CPH in SLA, other researchers find that a specified biological timeframe for second language learning is not necessary in order to produce nativelike competence. Hummel (2014: 171) states that this is not to say that the notion of the critical period does not stipulate that adult learners cannot reach very high levels of L2 proficiency, rather, it “predicts that beyond a certain age, learners cannot reach native speaker levels of proficiency”. Other researchers, such as Long (1990), have argued that “one single post-critical-period L2 learner with an underlying competence indistinguishable from that of native speakers would suffice to reject the CPH” (Abrahamsson & Hyltenstam 2009: 251). Therefore, in accordance with Long (1990), one successful late learner alone could suffice as enough evidence to at least partially refute the CPH in its strict form.

2.2. Defining nativelikeness

Defining nativelikeness can differ between several texts and individual interpretations of what nativelike competence of a second language constitutes. For the purpose of this paper, nativelikeness is defined as non-native speakers of a second or foreign language that reach a certain level of proficiency in their second language that is at a level comparable with native speakers of that language. Otherwise stated, nativelikeness is the quality of speaking or using a language and sounding like the native speaker of that language. Abrahamsson & Hyltenstam

(8)

be understood in their paper entitled “Age of Onset and Nativelikeness in a Second Language: Listener Perception Versus Linguistic Scrutiny”. According to Abrahamsson & Hyltenstam 2009: 259), interpretation 1 of nativelikeness is to self-identify as a nativelike speaker of the target language, interpretation 2 of nativelikeness is to be perceived as a nativelike speaker by native speakers of the target language, while interpretation 3 of nativelikeness is to be a

nativelike speaker of the target language in terms of linguistic measurable behavior. This paper employs interpretation 3 of nativelikeness, where participants must be a nativelike speaker of the target language. In addition to Abrahamsson and Hyltenstam (2009), American linguist David Birdsong defines nativelikeness in his publication entitled “Nativelikeness and

non-nativelikeness in L2A research”. According to Birdsong (2007: 32), “non-nativelikeness is most often operationalized as experimental performance on (psycho-)linguistic tasks which falls within the range of performance of native controls”.

For the control group of native speakers, there are also a few requirements to determine native levels of proficiency. First, the participants must be native speakers of language in question, meaning that the language used must be their first language that was acquired from early childhood on. Additionally, the participants must possess intuitive knowledge of their first language and be able to produce fluent, spontaneous discourse with a natural accent or dialect. Lastly, the participants must be communicatively competent and able to communicate within different social settings without any foreign accent (Lee 2005). These skills may be assessed through achieving native levels on oral proficiency tests, grammaticality judgement tasks and spontaneous speech that shows correct accent, grammar and semantic referencing (Lee 2005).

2.3. Stability hierarchy

Inspired by the work published by Belgian linguist Frans van Coetsem (1988), the stability hierarchy is a gradient that categorizes the differences in stable and unstable elements within the linguistic domains of language contact in second language acquisition. In his publication entitled “Loan Phonology and the Two Transfer Types in Language Contact”, van Coetsem (1988) explains the two types of transfer in language contact and how these linguistics borrowings are understood within the greater framework of the stability hierarchy of language. The first transfer type, according to van Coetsem (1988), is borrowing, which refers to when a

(9)

speaker transfers unstable elements from an L2 (second language) into L1 (first language), while speaking the L1. In contrast, imposition, the second type of language transfer, is when a speaker transfers stable elements from an L1 into L2, while speaking the L2. Within the linguistic domain of phonology, van Coetsem (1998: 7) implies that a phonological loan indicates any form of loan in the realm of segmental or suprasegmental phonology and states that it can be defined as “an imitation, replication or reproduction in the recipient language of a foreign or source language pronunciation”. However, phonological transfer is not the only type of language transfer and borrowing. According to van Coetsem (1988: 8), “the recipient language speaker may apply adaptation while imitating, i.e. borrowing, source language lexical material; he may then also apply adaptation in other domains than phonology.” These other linguistic domains included in the stability gradient are morphology, syntax and lexicon.

The stability hierarchy itself is based on the stable and unstable elements of a language and their resistance to change within language domains. ‘Stability’ here is defined by van Coetsem (1988: 25) as “resistance to change”. According to the stability hierarchy, instable elements are easy to learn, while stable elements are uneasy to learn. Van Coetsem (1988: 25) proposes that the two types of language transfer must be viewed in connection with a

constitutional property of language, “namely the difference in stability and resistance to change between and within language domains”, which van Coetsem names the stability gradient of language. The terms stable and unstable as they are used here must be interpreted with respect to the language user: which elements are difficult to eradicate and which are acquired rapidly while learning an L2? In general, the language user is scarcely aware, if at all, of stable elements (i.e. pronunciation, accent features), while unstable elements have a much higher degree of

awareness. The main claim here is that language does not offer the same degree of stability in all its parts, “in particular that there are differences in stability among language domains, namely among vocabulary, phonology and grammar (morphology and syntax)” (Van Coetsem 1988: 25). According to this theory, then, not all language transfers equally in each linguistic domain.

Within the framework of the stability hierarchy gradient, there have always been notable differences in stability between language domains. For example, according to van Coetsem (1999: 26), “it is well-known that vocabulary items are far more likely to be borrowed than phonological constituents.” However, a detailed scale or hierarchy of stability of the various language domains has yet to be worked out officially. Van Coetsem (1988: 26) himself considers

(10)

“vocabulary the least stable language domain, and phonology and grammar (morphology and syntax) the more stable ones.” Pronunciation features (the phonology or “accent”) and syntax are considered to be the most stable elements, as speakers are usually largely unaware of the accent and the word order of a foreign language. In the intermediary position is morphosyntax, in which inflection is related to syntax because it is governed by rules; derivation and composition, of which the rules are less clear. After these considerations, the stability gradient would go as follows, from the most stable to the least stable (see Figure 1 below): phonology (accent) → syntax → morphology → lexicon (vocabulary). According to this theory, with the stable elements being the hardest to learn and the instable elements being the easiest to learn, we can then assume that pronunciation features (phonology and accent) and syntax are the most difficult for a second language learner to obtain, while morphology and lexicon are the easiest for a second language learner to obtain. This gradient is considered when exploring which linguistic domains have the highest rate of nativelike attainment by the participants in each of the studies.

Figure 1: The stability hierarchy (Van Coetsem, 1988)

2.4. Usage-based language learning and grammar

The usage-based theory of language acquisition and grammar is an approach to linguistic communication that focuses on the unique social and pragmatic aspects of language acquisition in young children. In the chapter entitled “The usage-based theory of language acquisition” published in the Cambridge handbook of child language, American developmental psychologist

(11)

and linguist Michael Tomasello provides an account of the investigation of the usage-based approach to language acquisition in both its functional and grammatical dimensions. Here, Tomasello (2009: 86) takes the stance that “children acquire language first and foremost by understanding how others use language”. In this respect, Tomasello rejects Noam Chomsky’s Universal Grammar (UG) paradigm, which, as stated previously in section 2.1 of this paper, is the linguistic theory that proposes that all humans possess an innate and universal grammar system. Instead, Tomasello proposes the functional theory of language acquisition, also known as the usage-based theory of language acquisition and grammar, which believes that young children learn linguistic structures through their discourse and communication with others.

This approach can be summarized in two, simple aphorisms: first, meaning is use and second, structure emerges from use. Tomasello (2009: 69) proposes that “meaning is use represents an approach to the functional or semantic dimension of linguistic communication”, while “structure emerges from use represents an approach to the structural or grammatical dimension of linguistic communication”. This is different from the classic approach that formal structure and grammar are devoid of meaning, and suggests that rather, the focus of meaning itself through communication is how constructions in language use are acquired. Essentially, according to Tomasello (2009: 70), these theoretical positions represent the view that “the pragmatics of human communication is primary, both phylogenetically and ontogenetically, and that the nature of conventional languages – and how they are acquired – can only be understood by starting from processes of communication more broadly”. While Chomsky and the UG approach take the position that humans are born endowed with a biological basis for language, Tomasello (2009: 69) takes the position that children come to the process of language acquisition “around one year of age” equipped with two sets of cognitive skills, both evolved for other, more general functions before linguistic communication emerged in the human species: intention-reading (functional dimension) and pattern-finding (grammatical dimension). This approach also favors the view that children’s language skills evolve over time from the input of communication and is derived primarily from how the language they hear being used around them. However, while both the usage-based theory of language learning and grammar and the UG paradigm are opposing theories, neither one has been able to fully convey the process of grammatical language acquisition in children with complete certainty, and the exact method of structural acquisition of grammar and structure in language acquisition are minority positions held in the field.

(12)

3. Methodology

3.1. Defining individual variables

The central point of departure for this comparative research analysis is to examine the differences that contribute to an individual learner’s measurable success in acquiring a second or foreign language after Lenneberg’s aforementioned Critical Period Hypothesis (CPH). This analysis assesses a number of eminent studies from scholars and researchers that have, at least in part, either supported or rejected the statement that the attainment of nativelike competence in an L2 after the critical period is possible. To define success in these studies is to have at least one late learner, with an age of onset (AO) of the L2 after puberty, obtain results in line with that of the native speaker. Therefore, this paper takes the position that one learner alone could suffice to disprove the CPH. Lastly, this study addresses the question under which conditions

nativelikeness can be achieved and in which linguistic domains.

The differences that contribute to an individual learner’s measurable success in this study are broken down into the variables within each study that have a significant impact or effect on an individual learner’s attainment in SLA. This paper examines five variables in particular that will be outlined, in detail, as follows. The first variable that this paper explores is the age of onset (AO) of a learner, which is the age at which an individual first began to acquire a second language. A learner’s age of onset is usually categorized by “early” onset or “late” onset, but the actual age between what is early and what is late in terms of years of age differs for each

individual study. This can also be categorized in terms of “pre-puberty” or “post-puberty”, as Lenneberg draws the line of the CPH right at puberty (found across studies around age 12). The AO of a learner varies a considerable amount in the studies assessed, with the lowest AO beginning at 3 years of age (Johnson & Newport 1989, Flege, Yeni-Komshian, & Liu 1999).

The second methodological consideration in this study is the length of residence (LOR) of each individual learner. The LOR is the amount of time a participant has spent living in the country of the target language. Like the AO, the LOR of each learner in these studies varies considerably. Some mandatory length of residence of the participants includes in the studies was as little as 3 months in the target country (Colantoni & Steele 2006), while other studies had

(13)

participants with the length of residence significantly higher (26 years for learner 1 of Ioup et al. 1994 and an average of 17.4 years of Coppieters 1987).

The third methodological consideration in this study is the linguistic domain in which the participant achieves the nativelike results. The linguistic domains include the different branches of linguistics that are concerned with how languages are structured. Each of the four major linguistic domains are represented across the corresponding studies. These include phonology, or pronunciation, (the relationship among speech sounds), morphology (the study of words and how they are formed), syntax (the set of rules that govern the structure of sentences) and semantics (the study of meaning or denotation between words, signifiers and phrases). Also included is the extra domain of lexicon (the vocabulary of a person). The linguistic domain in which an

individual learner achieves, or does not achieve, nativelike results is a central focus of this study. Furthermore, this study determines whether or not a specific linguistic domain is easier to obtain nativelike results in when compared to other linguistic domains.

The fourth methodological consideration in this study is the L1-L2 pairing of the participants. The L1-L2 pairing is the language combination of the participants’ L1 (first or native language) with the participants’ L2 (second or foreign language). Like the linguistic domain, the L1-L2 pairing is another central focus of this study. Generally, L1-L2 pairings can cause two types of interference or transfer between language combinations, namely positive transfer and negative transfer. Positive transfer occurs when the learner creates correct language production due to the transfer of a unit or language structure that is the same in both languages, while negative transfer occurs when the learner creates an error due to the transfer of items or structures that are not the same in either language. An example of positive transfer in the field of lexicon would be cognates between languages and an example of negative transfer would be “false friends”. In order to get a wide representation of L1 and L2 pairings, this study uses a wide array of L1-L2 pairings that are examined to determine the influence of this variable in the study of nativelike competence. A complete list of the languages involved in these studies is as

follows: Arabic, Chinese, Dutch, English, French, German, Korean, Spanish and Swedish. For further classification, the L1-L2 pairings are categorized into three groups (small, intermediate and large), based on the distance between each language combination with respect to the linguistic family tree, a group of languages that are related through the decent of a common ancestral language or parental language. The ‘small’ distance represents any two languages that

(14)

have a close genealogical relationship, meaning that they both belong to the same language family and are very closely related within that tree (i.e. English and German are both classified as Germanic languages, so the distance is between them is small). The ‘intermediate’ distance represents any two languages that share the same language family but occupy opposing ends of that tree (i.e. Swedish and Spanish are on the opposite spectrums of the Indo-European tree). The ‘large’ distance represents any two languages that do not have a close genealogical relationship and occupy completely different family trees, sharing no genealogical relationship to each other whatsoever (i.e. Mandarin Chinese is a part of the Sino-Tibetan language family tree, while English is a part of the Indo-European family tree).

The fifth methodological consideration in this study is the educational level of the participants. This study defines the educational level as the highest level of education each participant has received during their years of schooling. While this variable is not highly controversial in SLA research, it could still play a role in a learner’s ultimate attainment of a second language. The level of education within these studies is constant, with each study all having participants completed some form of secondary education. The lowest level of education required for participants is no less than senior in high school (Abrahamsson & Hyltenstam 2009), while other studies have participants that are Masters students at a graduate university (Moyer 1999) or participants at the PhD level (Ioup et al. 1994).

The sixth and final methodological consideration in this study is the level of training of the participants. The level of training is defined as the amount of instruction a participant has received in the target language. While some studies have participants that do not specify or require any training at all in the L2 (Ioup et al. 1994, Bongaerts et al. 1997, Flege,

Yeni-Komshian, & Liu 1999), others have more rigorous levels of training involved, such as scoring ‘superior’ on the ACTFL oral proficiency scale (Coppieters 1987). It is still unclear if there is a correlation between the level of training in an L2 and the overall level of the L2 competence.

3.2. How the evidence is evaluated and weighted

This section of the paper outlines how each factor is evaluated and weighted within the framework of this research analysis. This is to answer the questions of which factors are evaluated, how they are evaluated, why they are significant for this research, and lastly, which

(15)

factors are the most important for consideration in the final conclusion and discussion. For a complete representation of this data, a table (Table 1) is employed in section 6 of this paper. The first factor to be evaluated is age of onset (AO) of a learner. As mentioned earlier, AO refers to the age at which an individual learner first began to acquire a second or foreign language. Since this paper is focused on the successful or unsuccessful acquisition of a second language after the critical period, which Lenneberg (1967) places at the age of puberty, the AO is divided into two sections: pre-puberty and post-puberty. While some studies employ participants that have begun acquiring their L2 before puberty, other studies employ participants that have begun acquiring their L2 after puberty, which calls for the distinction between the two groups within the AO of the participants. This distinction between the two groups of early and late learners remains constant in order to make a distinction between pre-puberty and post-puberty learners, with emphasis later on for the overall success of the post-puberty learners.

Conclusively, the AO is evaluated by comparing the overall results (did the participant achieve nativelike results or not) with the age of when the learner began acquiring the L2. If any individual learner in the post-puberty category achieves nativelike results, then that is counted for evidence against the critical period hypothesis.

The second factor for evaluation is the length of residence (LOR) of a learner. As mentioned earlier, the LOR is the amount of time a participant has spent living in the country of the target language. According to Bialystok & Miller (1999: 142), “evidence for the

critical period hypothesis requires that the behavioral effects be attributable to the age at which learning began, not the length of time engaged in practicing the behavior”. Ideally, a study should be conducted in which the AO is independent of the LOR of participants, with each having spent precisely the same amount of time practicing the language with the same LOR. While this research does not mandate that the LOR remain constant for each participant, and because it is virtually impossible to have each participant practice the language for the exact same amount of time, the LOR factor is not critically evaluated for direct correlations between the result nativelikeness and age of the learner. Rather, this analysis looks further into the linguistic domain and the L1-L2 pairing involved from the participants, and the LOR is taken into consideration when investigating the success, or lack of success, in these two categories. For example, if participants who achieved nativelikeness had a long LOR and participants who did

(16)

not achieve nativelikeness had a short LOR, does the overall LOR of the individuals in the target country then directly contribute to success, and if so, why?

The third factor for the evaluation is the linguistic domain of a learner. As mentioned earlier, the linguistic domains include the different branches of linguistics that are concerned with how languages are structured and are comprised of four categories: phonology, or pronunciation, morphology, syntax and semantics, with the extra inclusion of lexicon. The linguistic domain in which an individual learner achieves, or does not achieve, nativelike competence is a central point of focus for this study. In order to evaluate this factor, each of the major linguistic domains are represented across the studies. This analysis determines in the conclusion the exact number of studies with participants that either achieved nativelike results or did not achieve nativelike results and in which linguistic domain, as well as which domain are the easiest or hardest to obtain nativelike competence in. This is compared also with the L1-L2 pairing and language distance. Using the stability hierarchy from Frans van Coetsem (1988), this paper hypothesizes that pronunciation and syntax should be the most difficult for a learner to acquire, while morphology and lexicon should be the easiest for a learner to acquire.

The fourth factor for evaluation is the language pairing, or L1-L2 pairing. As mentioned earlier, this is the language combination of the participants’ L1 (first or native language) with the participants’ L2 (second or foreign language). Looking in detail at the L1-L2 language

combination in comparison to the overall results of the participants (did the participant achieve nativelike results or not) provides answers on whether or not it is easier for an individual learner to achieve nativelike results in respect to the distance between languages. This paper employs the categorization of three language distances (small, intermediate and large; see section 3.1. for further clarification). The point of departure for this analysis holds that it will be easier for an individual to achieve nativelike results with a small linguistic distance, where as it will be much more difficult and perhaps even impossible, for an individual learner to achieve nativelike results with a large overall linguistic distance.

The fifth and sixth factors for this study include the educational level and the level of training of the participants involved. As mentioned earlier, the educational level is defined as the highest level of education each participant has received during their years of schooling, while the level of training is defined as the amount of instruction a participant has received in the target language. These factors are grouped together as both factors are consistent throughout the studies

(17)

and labeled as constant variables in this analysis. Having these two variables as constant allows this study to look further into the linguistic domains and language pairings in relation to the overall possibility of achieving nativelikeness. Generally speaking, a higher educational level and level of training could indeed contribute to higher likelihood of nativelikeness after the critical period, however, a separate study on the educational level and level of training would have to be conducted to prove this theory.

4. Research against nativelikeness

4.1. Coppieters 1987

In 1987, researcher René Coppieters from Pomona College examined native and near-native speakers of French in a variety of underlying grammatical structures in his study entitled "Competence differences between natives and near-native speakers.” The goal of the Coppieters (1987) study was to determine if native and near-native speakers would develop essentially identical underlying grammars of the same language. In order to test this, Coppieters (1987) conducted extensive interviews on 21 near-native speakers (NNS) of French.

As a prerequisite of the study, the Coppieters mandated that the NNS must have a high level of proficiency of French and also must have all acquired French as adults. This puts the AO for each participant at or above the age of 18. For the length of residence (LOR) of the

participants involved in the study, according to Coppieters (1987: 551), “the number of years spent in France, usually Paris, varied between 5.5 and 37, for an average of 17.4; this was certainly enough time to develop a fairly complete knowledge of the language”. For the

educational level of the participants, Coppieters (1987: 551) used a group of “highly educated” NNS, consisting of 17 professors, 2 students/researchers at various universities, one archeologist, one sociologist, and eleven professors of language or literature, one engineer and one secretary. Lastly, for the level of training, the NNS participants all received highest rating, that of 'superior’ in terms of the ACTFL oral proficiency scale, revealing that the participants each had a high level of training in the L2 before having any involvement in the study (Coppieters 1987: 550).

(18)

For the procedure of the study, extensive interviews that averaged about 50 minutes were conducted and taped for later further analysis. Coppieters (1987) gave the participants sentences to judge for grammaticality and structure and compared their judgements with the native

speakers (NS) of French. Some of the grammar being tested here included correctly using the third person pronoun in predicative sentences, proper placing of the adjective between pre-posed and postposed adjectives, tense/aspect distinctions for the imperfect and compound passive, complex French syntax using to the object + predicate construction, the correct use and placement of prepositions, and the correct use of articles in French.

Figure 2: % of diverge from native majority norm (Coppieters 1987)

The outcomes of the Coppieters (1987) study (see Figure 2 above) reveal that the results of extensive interviews indicated that native and near-native speakers of French have strikingly different intuitions on French sentences. In particular, according to Coppieters (1987: 554), “the two groups have markedly divergent interpretations of sentences involving basic grammatical

(19)

contrasts such as the two past tenses, the 3rd person pronouns and the placement of the adjective before or after the noun”. Additionally, Coppieters (1987: 565) clarifies that “the NS's and NNS's studied here have developed significantly different grammars for French. In other words, a given language – French, in this case – does NOT impose a specific underlying grammar on its

speakers”. Overall, such results provide a clear illustration of the relative independence of the two levels of language: “on the one hand, language use, and on the other hand, underlying grammar as reflected by speakers' intuitions” (Coppieters 1987: 544).

4.2. Johnson & Newport 1989

In 1989, researchers Jacqueline S. Johnson and Elissa L. Newport from the University of Illinois addressed the hypothesized critical period to determine whether or not it exists by comparing the proficiency of L2 English of native Korean and Chinese speakers in their study entitled “Critical period effects in second language learning: The influence of maturational state on the acquisition of English as a second language.” Johnson & Newport (1989: 60) predict that in accordance with Lenneberg (1967), “young children are better language learners than adults and should consequently reach higher levels of final proficiency in the second language”. In order to test their theory, Johnson & Newport (1989) designed the present study to test the L2 of native Chinese or Korean speakers on a wide variety of structures of English grammar and morphosyntax using a grammaticality judgement task.

The participants of the study consisted of 46 native Chinese or Korean speakers who learned English as a second language. Johnson & Newport (1989: 68) explain their choice of the L1-L2 pairings in the present study by maintaining that “Chinese and Korean were chosen as the native languages because of their typological dissimilarity to English”. The primary criterion for the researchers in regard to participant selection was that they varied in age at which they moved to the United States and that all subjects were exposed to English by native speakers in the United States. As for the age of arrival or age of onset (AO) of the selected participants, Johnson & Newport (1989: 68) report that the 46 subjects varied in age of arrival in the United States from ages 3 to 23. The participants were divided into two groups, early arrivals and late arrivals, with each group having exactly half of the total (23 subjects) number of participants. The LOR

(20)

for an unbroken stay of at least three years prior to the time of testing. For the level of education, Johnson & Newport (1989: 68) report that “to ensure ample exposure to English and to ensure some homogeneity of social background, all subjects were selected from the student and faculty population at an American university (University of Illinois).” Finally, the level of training for the participants for the early arrivals was unspecified, and the level of training for the late arrivals required between 2 and 12 years of mandatory formal English instruction in their native country (Johnson & Newport 1989: 68).

For the procedure section of the study, the participants were tested on their knowledge of “English syntax and morphology by being asked to judge the grammaticality of spoken English sentences of varying types” (Johnson & Newport 1989: 70). The subjects evaluated the

grammaticality of each sentence, which was recorded on tape by a native English speaker, to give a yes or no (Y or N) response on whether or not they believed the prior sentence to be grammatically correct. According to Johnson & Newport (1989: 71), following the

grammaticality judgment test, subjects were interviewed for approximately half an hour about their language background. This included gathering information such as the type and amount of exposure they had to the English language.

(21)

The results obtained by Johnson & Newport (1989) were quite apparent. In answering their first question of whether there was a relation between age of arrival (AoA) and ultimate performance in the grammar of a second language (see Figure 3 above), Johnson & Newport (1989: 89) reported that “the results of this study clearly show such a relation, and therefore support the notion that children have an advantage over adults in acquiring a second language”. Johnson & Newport (1989: 60) conclude the findings of their results:

Both correlational and r-test analyses demonstrated a clear and strong advantage for earlier arrivals over later arrivals. Test performance was linearly related to age of arrival up to puberty; after puberty, performance was low but highly variable and unrelated to age of arrival. This age effect was shown not to be an inadvertent result of differences in amount of experience with English, motivation, self-consciousness, or American

identification. The effect also appeared on every grammatical structure tested, although the structures varied markedly in the degree to which they were well mastered by later learners. The results support the conclusion that a critical period for language acquisition extends its effects to second language acquisition.

4.3. Ioup et al. 1994

In 1994, researchers Georgette Ioup of the University of New Orleans, Elizabeth

Boustagui of the Arab Maritime Transport Academy, and Nal El Tigi and Martha Moselle from the American University in Cairo conducted a case study on two exceptional adult learners of Arabic in their study entitled “Reexamining the critical period hypothesis: A case study in a naturalistic environment”. The Ioup et al. (1994) study focuses on the ability of adults to achieve nativelike competence in a second language when the acquisition context lacks any sort of formal instruction. The study itself consisted of testing the nativelike competence of an adult named Julie, a woman who claimed to have acquired nativelike proficiency of in Egyptian Arabic (EA) through informal exposure. The overall goal of this study, according to Ioup et al. (1994: 73), was to “determine to what extent Julie’s linguistic competence matched that of native speakers”. Julie moved from Britain to Cairo, Egypt at the age of 21, when she married an

Egyptian man. Ioup et al. (1994: 77) report, “Julie acquired Arabic very quickly. Within 45 days she was using simple sentences and fixed expressions. By the end of 6 months, she was

(22)

In order to test if Julie was an exception of the CPH, Ioup et al. (1994) assessed her quality of speech, accent, and knowledge of syntactic rules by comparing her to both native speakers of Arabic and a proficient L2 learner who began learning Arabic in a formal setting. This proficient L2 learner was named Laura, an American living in Cairo who commenced in the formal study of standard Arabic for many years and teaching standard Arabic at the university level. While Julie had no specified level of education or formal training in EA, Laura possessed a Master’s degree in modern Arabic, where she finished at the American University of Cairo with a Ph.D. in standard Arabic. The first procedure in the study was speech production, where Ioup et al. (1994: 79) “evaluated the two participants spoken language to determine if they would be considered as native speakers by the Egyptian judges”. The speech task involved spontaneous speech that was recorded, where the participants were asked to explain their favorite recipe in detail. The second procedure was accent identification, where the two subjects were tested on their ability to discriminate among various accents and regional dialects of EA that were

recorded and played for them on tape. The last procedure was grammatical intuitions, where the two subjects were “administered a translation task, a grammaticality judgement task, and task to measure the interpretation of anaphora” (Ioup et al., 1994: 81). Aside from Laura, the study used eleven native speakers of EA that were evaluated for comparison.

Although both Julie and Laura could be perceived as successful, nativelike L2 speakers both immersed in Egyptian language and culture, the results of the Ioup et al. (1994) study revealed that both participants did not perform on the level of native-like speakers on the grammaticality judgement task. According to their review of the Ioup et al. (1994) study,

Abrahamsson & Hyltenstam 2009: 257) conclude that “Although both Julie and Laura performed extremely well on all these tasks”, in which case Julie even performed better than many native speakers on the accent identification task, “both performed significantly below the native-speaker range on aspects of grammatical intuition”. While both Laura and Julie did not match nativelike measures in every aspect, they can still be considered as exceptional late L2 learners with a considerable amount of success. Ioup et al. (1994: 91) suggest that further research be conducted in other linguistic domains, “such as syntax and semantics that nonnative speakers fail to reach native norms”, to further evaluate Julie and Laura’s linguistic competence before a final evaluation of their internalized grammar can be completed.

(23)

4.4. Bialystok & Miller 1999

In 1999, researchers Ellen Bialystok and Barry Miller from York University conducted a study on the issue of a critical period for second language acquisition in their study entitled “The problem of age in second-language acquisition: Influences from language, structure, and task”. In this study, Bialystok & Miller (1999) perform a partial replication of the Johnson & Newport (1989) study, in which three groups of participants were given a grammaticality judgement test based on five structures of English grammar in both an oral and written form. The three groups of participants that were involved in this study consisted of one group of native Chinese

speakers, one group of native Spanish speakers, and one group of native English speakers. The participants of this study, which were divided into three groups, were all either university students or university graduates. The first group of participants consisted of 33 adults who spoke Chinese (Mandarin) as a first language and learned English as a second language, with a mean age of participants in this group was 24 years (range 19 to 37 years) (Bialystok & Miller 1999: 131). The second group of participants consisted of 28 adults, with a mean age of 25 years (range 18 to 47 years), whose first language was Spanish and the third group of

participants consisted of 38 native speakers of English with a mean age of 28 years (range 21 to 43 years) (Bialystok & Miller 1999: 131). Additionally, the participants were divided into older and younger learners, where the cutoff age was 15 years at which they arrived to Canada. The mean LOR for the total number of participants in Canada was 7.5 years (Bialystok & Miller 1999: 131). Lastly, the level of training mandated that a comprehension test must be given by the researchers that each participant must pass (Bialystok & Miller 1999: 131).

As for the tasks and procedures involved, the participants who were not native speakers of English were asked to take a first-language proficiency test to assure that they were

reasonably competent in their first language (in this case, either Mandarin or Spanish). Following the first-language proficiency test was the grammaticality judgement test. According to Bialystok & Miller (1999: 132), the “participants were presented with a set of 160 sentences and asked to decide whether each sentence was correct or contained a grammatical error” and the errors were based on one of five structures: plurals, determiners, future tense, present progressive, and collocation restriction. There were “16 examples of each sentence structure, and each appeared

(24)

in its grammatical and ungrammatical form. All sentences contained between 6 and 9 words and were composed of between 10 and 15 syllables” (Bialystok & Miller 1999: 132).

For the results, Bialystok and Miller (1999) (see Figure 4 above) found participants in two learner groups (L1 Spanish, L1 Chinese) “who performed like English native speakers on the GJT until AO 8 years, whereas no participants with AO beyond this age were reported to score within the range of native speakers” (Abrahamsson & Hyltenstam 2009: 252). Bialystok and Miller (1999: 127) report themselves that “the results showed that performance patterns that were different for the two learner groups, that the linguistic structure tested in the item affected participants' ability to respond correctly, and that task modality produced reliable response differences for the two learner groups”. Furthermore, although there were proficiency differences in the grammaticality judgement task between the younger and older Spanish learners, there were no such differences for the Chinese group. Bialystok and Miller (1999: 127) conclude that “the age of learning influenced achieved proficiency through all ages tested rather than defining a point of critical period”, and that “the results are interpreted as failing to provide sufficient evidence to accept the hypothesis that there is a critical period for second language acquisition.”

(25)

4.5. Abrahamsson & Hyltenstam 2009

In 2009, researchers Niclas Abrahamsson and Kenneth Hyltenstam from Stockholm University conducted a large-scale study on the incidence of nativelikeness using native Spanish speakers in L2 Swedish with differing AO in their study entitled “Age of Onset and

Nativelikeness in a Second Language: Listener Perception versus Linguistic Scrutiny.” According to Abrahamsson & Hyltenstam (2009: 258), the study is aimed at “identifying individuals who would potentially constitute the evidence necessary to reject the CPH”.

The participants of the study consisted of a total of 195 L2 speakers of Swedish who had begun their acquisition at various ages and who “identified themselves as advanced and

potentially nativelike L2 speakers” (Abrahamsson & Hyltenstam 2009: 262). All participants had to meet certain criteria, which included having Spanish as their L1, be 19 years of age or older, have lived in Sweden for 10 years or more, speak Swedish fluently without a foreign accent or any obvious grammatical deviations, have an educational level of no less than senior high school (i.e., minimally 12 years of schooling) and to have primarily been exposed to and acquired the variety of Swedish spoken in the greater Stockholm area (Abrahamsson & Hyltenstam 2009: 262). As for the AO of their L2 Swedish, the participants were divided into two groups, with AO either before age 12 or after age 12; of the 195 participants, 107 began their acquisition of Swedish before age 12 years and 88 began to learn Swedish at the age of 12 years or later (Abrahamsson & Hyltenstam 2009: 263). Lastly, for the level of training involved in this study, the researchers reported that the participants all identified themselves as advanced and

potentially nativelike L2 speakers.

As for the procedures involved, the present study was segmented into two parts, referred to as Part I and Part II. The goal of Part I was to investigate the number of these advanced learners who are perceived as nativelike speakers by native listeners, where native judges used recorded speech samples to interpret 40 randomly selected L2 speech samples. The goal of Part II was to investigate “how many of these perceived nativelike speakers, in fact, behave

linguistically like native speakers of Swedish on a broad selection of language proficiency measures and tasks” (Abrahamsson & Hyltenstam 2009: 260). As an overall goal of the study, the researchers’ goal was to produce an “across the board” representative measurement of nativelikeness (Abrahamsson & Hyltenstam 2009: 261). The tasks involved in testing the

(26)

nativelikeness of each participant included aspects of language processing, in which the present study employed 10 instruments for L2 scrutiny, “covering phonetic production and perception (voice onset time), perception of words and sentences in white noise and babble noise,

grammaticality judgements, grammatical, lexical and semantic inferencing, and formulaic language” (Abrahamsson & Hyltenstam 2009: 260-1). Accordingly, all four of the linguistic domains (phonology, morphosyntax, lexicon and semantics) are represented in this study.

Figure 5: Aptitude scores as function of AO (Abrahamsson & Hyltenstam 2009)

The results of the Abrahamsson & Hyltenstam (2009: 288) study were quite decisive (see Figure 5 above). In terms of age of acquisition (AoA), the researchers found that a strong

negative correlation between perceived nativelikeness and the age of which the L2 acquisition began, where an overall group comparison “revealed a significant difference in perceived nativelikeness between early and late learners”. Additionally, they discovered that the average perceived nativelikeness “began to decrease at or around AoA 12 years, but this decrease leveled

(27)

out at some point after adolescence” (Abrahamsson & Hyltenstam 2009: 288). The main point of departure for these results here is that the AoA of L2 acquisition “stands out as the variable that best predicts ultimate perceived nativelikeness” (Abrahamsson & Hyltenstam 2009: 288). As for the results of the 10 different measures that scrutinized L2 processing representing various aspects of speech production and perception, morphosyntax, and formulaic language, results showed that “none of the late learners exhibited actual, overall linguistic nativelikeness when scrutinized in detail” (Abrahamsson & Hyltenstam 2009: 289). In conclusion, the overall results of the present study support the position held namely that nativelike L2 proficiency is, in

principle, never attained by adult learners and furthermore, the results summarized by the researchers “point more in the direction that absolute nativelikeness in late learners, in principle, does not occur” (Abrahamsson & Hyltenstam 2009: 294).

5. Research supporting nativelikeness

5.1. Birdsong 1992

In 1992, David Birdsong from the University of Texas at Austin tested late foreign language learners of French in a replication of the Coppieters (1987) study in his study entitled “Nativelike pronunciation among late learners of French as a second language.” Birdsong (1992: 707) first briefly explains the Coppieters (1987) study, which he states compared grammaticality judgements of French natives with near-native French speakers from a variety of backgrounds, all having acquired French as adults. In his own study, Birdsong (1992: 707) asks two

fundamental questions for this study in regard to the question of ultimate attainment; the first question Birdsong asks is whether there are “complete differences between native speakers (NS) of a language and those exceptional learners who appear to perform like natives, i.e. near native speakers (NNS)”. The second question of the Birdsong (1992: 707) study question focuses on the differences between native speakers and near native speakers, asking the question “in what grammatical domains do the groups appear to diverge?”.

(28)

judging the acceptability of French sentences. The ENS were all college-educated with an average age of 40 years old and all had been exposed to French or had begun learning French at or post puberty. Similarly, the FNS were all college-educated with an average age of 35-40 years old and all were speakers of standard French that had lived or grown up in major metropolitan areas of France (Birdsong 1992: 717). The average cumulative LOR for the ENS was 11.8 years, while the average age of arrival (AoA) in France was 28.5 years, with a range of 19-48 years (Birdsong 1992: 717). As for the level of training involved, Birdsong (1992: 717) reports that all but 2 subjects had formal instruction in French, with the average length of study was 6 years, with a total range of 1-14 years.

For the procedure, both subject groups were presented with a list of 76 grammatical and ungrammatical French sentences in written form and asked to give each a score of 1-5 for acceptability judgements. The French sentences that were being judged for grammatical

structural features included sentences using adjacency, prenominal past participle + modifier and middle voice (Birdsong 19922; 720). In addition, following Long’s (1990) call for more

demanding tests of ultimate attainment, “it was important to exclude structures that were ‘too easy’ and to exemplify instead ‘subtle’ areas of the target-language grammar, i.e. syntactic phenomena more complex and less-frequently occurring” (Birdsong 1992: 720).

As for the results of the Birdsong (1992) study, (see Figure 6 on next page), divergence in answers between the ENS and FNS only occurred for 22% of overall items for the acceptability judgement tasks. In terms of the cumulative deviance for these judgement tasks, according to Birdsong (1992: 739), the deviance figures for individual participants shows that 15 of 20 ENS are within the FSN range. Additionally, Birdsong (1992: 739) states that “this outcome is at odds of Coppieters (1987), who found that not a single near native speaker performed within the native speaker range”. Speaking on the overall results of the present study, Birdsong (1992: 742) suggests that post-pubertal learners can acquire nativelike competence. Birdsong looks toward future research in SLA to investigate many of the unanswered questions in this field, such as the ‘trainability’ issue, in asking “what can be discovered from exceptional learners that could be applied to improve other learners’ chance of attaining native norms?”.

(29)

(30)

5.2. Bongaerts et al. 1997

In 1997, researchers Theo Bongaerts, Chantal Van Summeren, Brigitte Planken and Erik Schils from the University of Nijmegen conducted two studies in their paper entitled “Age and ultimate attainment in the pronunciation of a foreign language.” In these studies, Bongaerts et al. addressed the issue of ultimate attainment in SLA by late second language learners. The

Bongaerts et al. (1997: 447) studies included a carefully screened group of highly successful Dutch learners of English in their designs, where the aim of studies was to determine “whether or not late second language learners who had achieved a nativelike performance in the

pronunciation of a second language could be identified”. Two separate studies were conducted in order to test if it is possible to have participants, which were late second language learners, to achieve scores within a nativelike range after a specified biological period of time.

In the first study, three groups of subjects participated. Group 1 consisted of a control group of 5 native speakers of British English, Group 2 consisted of 10 L1 Dutch learners of L2 English designated as highly successful learners of English, and Group 3 consisted of 12 L1 Dutch university students of L2 English at various levels of proficiency. With regard to age, none of the Dutch subjects had received instruction in English before or around the age of 12. The LOR of the participants in each group is unspecified. For the educational level of the

participants, Bongaerts et al. (1997: 452) report that across all 3 groups, the participants involved were either university lecturers and/or university students. The level of training for group one is unspecified, although Bongaerts et al. (1997: 452) mentions that the learners had been designated “by EFL experts as highly successful learners with an excellent command of (British) English”. All subjects were assigned to do 4 different tasks: to provide spontaneous speech samples, to read an English text of 84 words, to read 10 English sentences, and to read a list of 25 English words. According to Bongaerts et al. (1997: 452), “The most striking result of the study was that the judges could not distinguish the group of highly successful learners from the native speaker control group on the basis of the four speech samples they were presented with”. Bongaerts et al. (1997: 452) interpreted these findings as “constituting evidence against the claim that only younger learners can attain such a good pronunciation of a second language that they can pass for native speakers of that language”.

(31)

In the second study, three separate groups participated. Group 1 consisted of 10 native speakers of British English with a neutral, non-regional accent, Group 2 consisted of 11 native speakers of Dutch regarded as highly successful learners of English, and Group 3 consisted of 20 native Dutch speakers with a wide range in their command of English. Again, all of the Dutch participants were considered as late learners of English beginning around the age of 12 with an unspecified LOR. All participants in the second study were also reported as University lecturers and/or university students and were designated as highly proficient learners of British English. For the processes involved, speech samples were elicited from all subjects by having them read out-loud 6 sentences, 3 times. The speech samples were recorded on a tape at the sound studio at the University of Nijmegen, where 13 native speakers of British English rated the speech

samples (Bongaerts et al. 1997: 455). As a conclusion, Bongaerts et al. (1997: 462) stated that the main result of this study was that “some, but not all, individuals in the group of highly successful learners received ratings from experienced and inexperienced English judges that were comparable to the ratings assigned to the native speaker controls”.

Of the speech samples assessed by the inexperienced native English judges from the first study, none of the judges were able to make a distinction between the native speaker control group and the group of highly successful learners of English (see Figure 7 below).

(32)

Bongaerts et al. (1997) agrees with the conclusion made by Scovel (1988), in that “such a result clearly does not support the claim that it is impossible to achieve an authentic, nativelike pronunciation of a foreign language after a specified biological period of time” (Scovel, 1988). Future research for Bongaerts et al. (1997: 463) asks the empirical question “whether the results of this research can be replicated in studies that involve typologically less closely related

languages than Dutch and English”. In conclusion of the Bongaerts et al. (1997: 464) study, it seems as if the researchers may have identified “at least some individuals who have attained a nativelike level of performance in the pronunciation of a language, in spite of a late start”. They also claim that a “major challenge in the future will be to identify the combinations of a learner, learning context, and language variables (L1-L2 pairings) that make such a feat possible” (Bongaerts et al. 1997: 464).

5.3. Moyer 1999

In 1999, Alene Moyer from Georgetown University conducted a study that set out to challenge the Critical Period Hypothesis (CPH) on ultimate attainment in L2 phonology by examining 24 highly motivated American graduate students of German in her study entitled “Ultimate Attainment in L2 Phonology: The Critical Factors of Age, Motivation, and

Instruction”. In her study, Moyer (1999: 81) focused on examining the phonological performance among highly motivated subjects who use German daily as graduate student instructors and “who have been immersed in the language through in-country residence, augmented by years of instruction in both language- and content-based courses”.

The participants of Moyer’s study consisted of 24 Graduate students in German at the University of Texas in Austin in the United States of America. Additionally, the students were all employed as teachers of the first four semesters of the German program. Moyer (1999: 86) presumes these students as high-motivated learners and speakers, “given their attention to mastering the language, necessitated through their program of study and their teaching”. In addition to being highly motivated, the students were labeled as late-learners and non-native speakers (NNS) of German, due to the fact that “all students had no measurable exposure to the target language until the critical period of language learning had passed” (Moyer 1999: 86). This puts the AO of the participants at the post-puberty level. For the LOR, Moyer (1999: 86) reports

(33)

that the exposure to German immersion time in Germany, ranging from several months to several years, with a mean of 2.7 years.

For the procedures and tasks involved, four speech elicitation techniques were used, representing four different speech modes. The subjects were asked to read a list of 24 words, to read a list of 8 sentences, to read a paragraph of text, and lastly, to participate in a free-response section in which the subjects could speak freely for several minutes on any one of five possible topics to choose from. According to Moyer (1999: 86), specific read aloud items were chosen for their inclusion of phones that, according to contrastive analysis, are typically difficult for native speakers of English”. In addition to the 24 non-native speakers (NNS), 4 native-speaker (NS) judges of German who had recently arrived in the United States were asked to judge the speech samples and to rate each speech sample for authenticity. The task of the German NS was to first assign each speech sample on a native versus non-native status, then use an effective six-point scale of judgment (1 = definitely native, 2 = native, 3 = perhaps native, 4 = perhaps nonnative, 5 = nonnative, 6 = definitely nonnative) on the speech samples, where the average scores of each subject was calculated by the judges (Moyer 1999: 87).

On a review of the Moyer (1999) pronunciation study, the results revealed that the word-list task produced the highest incidence of nativelikeness, as judged by a panel of four native German listeners. This was then followed by “sentence reading, paragraph reading, and free speech production, on which most learners failed to pass for native speaker” (Abrahamsson & Hyltenstam (2009: 254). Overall, only one individual among the 24 advanced learners when tested in the Moyer 1999 pronunciation study passed as a native speaker in all four speech modes (see Figure 8 below on the following page). Moyer (1999: 98) calls this single outlier an

“exceptional learner” and argues that this outlier “should be mentioned as a reminder of those uncommon cases in which a late learner’s speech reveals no maturational effects”. This

exceptional learner had a performance that was consistent with the native controls across all four tasks and was not exposed to German before the age of 22, suggesting that nativelike

competence can be achieved in late learners of a second language. This single successful learner proves that just one outlier, or exceptional learner, can prove to suffice evidence that nativelike competence in late learners can still be achieved.

(34)

Figure 8: Mean rating of score with age of immersion (Moyer 1999)

*Note: “In performing the statistical tests, initial results were considerably skewed due to the

one outlier who performed clearly within the NS category (below 3 on mean rating). Because of the exceptional outcome and circumstances under which this subject learned German, this subject’s score was excluded (still visible on the chart—outlier at bottom of Figure. The overall

performance results support previous evidence for a negative correlation between age of exposure and attainment” (Moyer 1999: 90).

5.4. Flege, Yeni-Komshian, & Liu 1999

In 1999, researchers James Emil Flege from the University of Alabama at Birmingham, Grace H. Yeni-Komshian from the University of Maryland and Serena Liu from the University of Alabama at Birmingham evaluated the critical period hypothesis for second language

acquisition in their study entitled “Age Constraints on Second-Language Acquisition”. In this study, Flege, Yeni-Komshian, & Liu (1999) examined the age variable in SLA by assessing the effect of the age of arrival (AOA) on the L2 performance of Korean learners of English through a series of tests that evaluated their foreign accent ratings and morphosyntax scores.

The participants of the study consisted of 240 native speakers of Korean that arrived in the United States between the ages of 1 and 23. To be included in the study, the native Korean participants were required to speak “no language other than English and Korean and to have