Language Aptitude in Second Language Acquisition (SLA)

Abstract

This paper reviews how the issues of language aptitude in second language acquisition (SLA) has been investigated and developed over a period of time. To be begin with, the construct of language aptitude and how its operationalization has experienced changes over the time in SLA is addressed. Secondly, empirical studies examining the association between foreign language aptitude and instructional approach for grammar, feedback, as well as language learners’ speaking ability are respectively reviewed. Thirdly, the interaction between foreign language aptitude and early learners as well as late ones is also taken into consideration. Also, the connection between aptitude and motivation, a well-researched individual difference (ID) factor, is reviewed. Eventually, pedagogical implications and future directions are mentioned.

Key words: language aptitude, individual difference, SLA

Introduction

Among ID factors, motivation and language aptitude are the most attractive sources for SLA research. Language aptitude, a notion relating to cognition, is defined as a “predictor” of learners’ success in language learning. Wesche’s (1981) paper was one of the first articles who utilized the participants’ language aptitude to match them with appropriate training approaches and proved positive results. Conducted nearly three decades later, Hwu and Sun’s (2012) study was different from Wesche’s (1981) article in a way that Hwu and Sun (2012) were interested in examining how language aptitude mediated the effectiveness of instructional methods. Between these two papers, remarkable studies have been conducted regarding language aptitude and SLA. Therefore, with the aim to bridge the gap and clarify the development of issues of language aptitude, this paper begins by providing details of the two main studies. Then, the development of aptitude and its operationalization is explored. Next, the link between language aptitude and grammar instructional approach is examined. Additionally, the connection between aptitude and feedback is reviewed. Also, the association between language aptitude and learners’ speaking ability also received some attention from SLA researchers, which is also addressed in the paper. In addition, how aptitude differently interacts and has impacts on young learners as well as late learners is investigated. Finally, the conclusion along with pedagogical implications and possible future directions are made.

A broad picture of language aptitude: Wesche (1981) and Hwu and Sun (2012)

Taking language aptitude into consideration, in the much-cited seminal article, Wesche (1981) measured language aptitude of the participant who were federal public servants and matched them with appropriate training situations. The adult participants were asked to sit for a language aptitude test consisting of five subtests of MLAT and two subtests of LAB (the sound discrimination and sound symbol association). Afterwards, based on the criterion of language aptitude, the participants were divided into one of the three groups of Audio-visual Method, Analytical approach, or Functional approach. According to Wesche (1981), Audio-visual Method was the core method and applied with students who had high overall ability and acceptable scores on the subtests of LAB. For those who were not assigned to the core method, they would go with either one of the two alternative approaches (analytical or functional). More specifically, the participants whose scores on Word in Sentences and Spelling Cues were high were matched with analytical approach. For those who performed well on memory, auditory, and phonetic coding abilities were categorized into group receiving functional approach. The results from Wesche’s (1981) seminal article demonstrated that the participants expressed satisfaction with the teaching approach they had been assigned to in general. Particularly, the participants assigned appropriate approach did not only show positive attitude and motivation to practice the target language outside the classroom setting; they also expressed lower level of anxiety in class. Overall, the example of Wesche (1981) clearly pointed out the practical and educational purposes of language aptitude test. Also, Wesche’s (1981) study proved that matching language learners with suitable teaching approach based on their language aptitude would have positive impacts on their language learning.

Conducted many years later and sharing some differences from the seminal article of Wesche (1981), Hwu and Sun’s (2012) study was done to investigate how language aptitude could mediate the effects of language instruction. Hwu and Sun (2012) conducted the study with university student in the US who were studying Spanish as a second language. L2 Spanish learners participating in the study were asked to complete language aptitude test, including: memory for text, associate memory (MLAT), and grammatical sensitivity (MLAT). Then, they were pretested with two types of assessment, which were written sentence production and written sentence correction. After the language aptitude test and the pretest, the participants were divided into three different groups: the control group which received no treatment, deductive (DE) group, and explicit-inductive (EI) group. Regarding the target feature, the Spanish verb gustar was chosen to focus because it has been considered as a problematic language feature for Spanish learners. For the two experimental groups, they received the treatment with five lessons in which each of the lesson lasted around forty to fifty minutes, while the control group received no treatment. Afterwards, the participants were asked to finish the immediate posttest and two delayed posttests. The results from Hwu and Sun’s (2012) study pointed out four main findings. Firstly, the two experimental groups showed overall improvement in the immediate posttest although experiencing decrease in the second posttest and slight increase in the last posttest. Secondly, the results pointed out that the two instructional groups performed equally well in both assessment tasks. Thirdly, it was found that in two experimental groups, there was a significant relationship between the learners’ gain scores and their memory-for-text ability. Lastly, the results indicated that the participants of EI group showed significant improvement in comparison with their counterpart, DE group, in both production and correction tests. This finding led to the conclusion that EI approach would be beneficial for learners with high ability of memory for text.

There there are some similarities and differences between Wesche’s (1981) seminal article as well as Hwu and Sun’s (2012) study. The first point relates to language aptitude measurement. It could be noticed that the well-known language aptitude tests MLAT (Carroll & Sapon, 1959) and LAB (Pimsleur, 1966) were utilized in the two articles. However, in Hwu and Sun’s (2012) study, the memory for text was added to capture the interaction between learner’s memory for text ability and the gain scores. In other words, the measurement of foreign language ability has changed over time and the role of working memory was attached to the notion of language aptitude. Although MLAT and LAB have been known to be the most validated tools to capture learners’ language aptitude, the advent of other tests such as CANAL-F model (Grigorenko et al., 2000), and LLMA test (Meara, 2005) also brought a variety of language aptitude perceptions and took different components into consideration. The second point is that both Wesche (1981) as well as Hwu and Sun (2012) explored the interaction of foreign language aptitude and instructional approach, there is a difference between these two papers. Wesche (1981) focused on matching learners with suitable instructional approach and looked at the general results, whereas Hwu and Sun (2012) investigated how language aptitude of learners could mediate the effects of instructional methods. Regarding this area, many studies have been conducted to find out the interaction between language aptitude and second language acquisition and instruction (Erlam, 2005; Sheen, 2007; Smemoe & Haslam, 2012; Winke, 2013; Yilmaz, 2013; Hwu & Sun, 2014; and so on). For the third point, although the participants of Wesche’s (1981) article as well as Hwu and Sun’s (2012) study were all adults, the interaction between language aptitude and learners’ ages is also worth tracing. To begin with, the construct of foreign language aptitude and its operationalization will be addressed in the next part of the paper.

The construct of language aptitude and its measurement

Language aptitude is known as a predictor of language learners’ rate and success. It was once questioned whether the notion of language aptitude overlapped with intelligence. According to Skehan (1998), there was some partial overlapping between these two constructs. Taking MLAT – a type of language aptitude test as an example, Skehan (1998) stated that MLAT’s grammatical sensitivity and analytical abilities subtests overlapped with intelligence. Although there are overlaps between the two constructs, language aptitude and intelligence are still conceptualized as two different notions (Ortega, 2009). Regarding operationalization, foreign language aptitude is measured by a test with a variety of subtests to touch upon different elements of aptitude. More specifically, many tests have been developed to capture learners’ language aptitude. Owing to the high rate of validity, MLAT and LAB were not only used by Wesche (1981) but it was also utilized by Hwu and Sun (2012). Besides these two major types of tests, CANAL-F test and LLMA test are also utilized to measure language aptitude in many studies, for instance Erlam (2005) and Winke (2013). In addition, through the development of foreign language aptitude’s operationalization, working memory gradually involved in the notion of aptitude and it has been considered as a component of language aptitude, leading to the fact that tests measuring working memory capacity have been developed and applied in research.

Modern Language Aptitude Test (MLAT)

As the outcome of five-year research, Carroll and Sapon (1959) released MLAT with the expectation of distinguishing between good and bad language learners. MLAT, the first test measuring language learners’ aptitude, was designed to have five sections. In the first section of the test, number learning, test-takers were provided auditory practice to learn numbers from one to four hundred in a new language. Then, they were asked to translate fifteen numbers into English. The second section, spelling clues, measured how a test-taker would chose a printed phonetic script for a word they heard. The third part of MLAT asked test-takers to choose a word which had nearest meaning in their L1. The forth part, measuring grammatical sensitivity, required subjects to choose an option which played the same role as an underlined word. The last section, paired associates, was designed to measure leaners’ rote learning ability. Through five sections of MLAT, Carroll and Sapon (1959) mentioned that foreign language aptitude was perceived as a construct with four main components: phonetic coding ability, grammar sensitivity, inductive language learning ability, and associative memory respectively.

PLAB (Pimsleur Language Aptitude Battery)

Following the same approach as Carroll and Sapon (1959), Pimsleur’s (1996) PLAB included six sections, which were Grade Point Average (GPA), interest in foreign language learning, vocabulary, language analysis, sound discrimination, and sound-symbol association. With these six sections of language aptitude test, Pimsleur (1996) perceived language aptitude as a construct with three components: verbal intelligence (learners’ word knowledge and the ability of analytical reasoning), auditory ability (ability of receiving and information processing auditorily), and learners’ motivation. Dörnyei and Ryan (2015) pointed out some similarities between how Carroll and Sapon (1985) and Pimsleur (1996) conceptualized the notion of foreign language aptitude. In particular, Pimsleur’s (1996) verbal intelligence component was the same as what Carroll and Sapon (1985) addressed as inductive language learning ability and grammatical sensitivity. Also, auditory ability of Pimsleur (1996) was similar to Carroll and Sapon’s (1985) phonetic coding ability (Dörnyei & Ryan, 2015). The major difference was that Pimsleur (1996) conceptualized the learners’ GPA and their level of motivation as a component of language aptitude, which was not supported by Gardner and MacIntyre (1992) as Gardner and MacIntyre (1992) believed that motivation and language aptitude should be two different individual difference factors. In short, the advent of LAB test has brought a new dimension to the construct of language aptitude. Foreign language aptitude did not solely consist of four components as Carroll and Sapon (1959) has mentioned; learners’ GPA and motivation should also be considered as add-on features predicting the success of language learners.

LLMA Test

Following the approach of Carroll and Sapon (1959), LLAMA test was built by Meara (2005) with the aim to measure foreign language aptitude. Specifically, the commonly-used version of LLAMA was called “LLAMA suite” which consisted of four subtests. The first subtest, vocabulary learning task, was perceived to be the same as Carroll and Sapon’s (1959) vocabulary learning task. The second subtest was to measure how a participant was able to recognize effectively segments of an oral language which they have previously provided with; this subtest was believed to be different from Carroll-tradition. The third subtest, called sound-symbol correspondence, required test-takers to match syllable sounds in an unfamiliar language. The forth subtest was grammatical inferencing task with visual stimuli with the hope of making the test suitable with participants with different L1s. Although LLAMA test provided new insights of means to measure language aptitude, the developer Meara (2005) recommends that the test should not be used in high-stake situations due to its low validity.

Cognitive Ability for Novelty in Language Acquisition-Foreign (CANAL-F test)

Approaching foreign language aptitude from a novel perspective, Grigorenko (2000) conceptualized language aptitude as the ability to handle novelty and ambiguity when it came to L2 learning. For this perception, Grigorenko (2000) proposed a new test type which captured different abilities in comparison with Carroll and Sapon (1959). According to what Dörnyei and Ryan (2015) reviewed, CANAL-FT focus on five knowledge acquisition processes which were selective encoding, accidental encoding, selective comparison, selective transfer, and selective combination. In terms of language level, this type of test was operationalized at four aspects of lexical, morphological, semantic, as well as syntactic. Despite supplying new perception of language aptitude, the validity of CANAL-FT was proved to be less validated than MLAT (Sternberg & Grigorenko, 2002).

Overall, it could be observed that the tests capturing language learners’ aptitude have considerably developed throughout the years. On the one hand, the advent of various tests gradually added meaningful components, which helped SLA researchers understand aptitude more thoroughly. Also, it provided a variety of tools for researchers to operationalize the language aptitude construct; MLAT has no longer been the only choice. On the other hand, although many different types of tests were developed and utilized in research, MLAT still holds the superiority due to the high level of validity. This could be noticed through the two studies of Wesche (1981) and Hwu and Sun (2012) in which MLAT were utilized. More specifically, although Hwu and Sun (2012) conducted the study nearly three decades after the seminal article of Wesche (1981) and no matter that there were many test types to choose from, MLAT was still Hwu and Sun’s (2012) choice. However, it is essential to mention that memory for text was used in Hwu and Sun’s (2012) paper. It is worth addressing how the link between language aptitude and memory as well as how memory has been perceived as a component of language aptitude.

Working memory as a component of language aptitude

It could be observed that in the development of MLAT, Carroll and Sapon (1959) took associative memory – a dominant memory form in psychology into investigation. However, the perspectives of memory have experienced dramatical changes over time. Baddely (2007) stated the fact that working memory was linked to the development of mentioned perspectives. Specifically, Collins and Marsden (2016) reviewed how working memory connected to language learning. Firstly, working memory enabled learners to retain the amount of sound and visual form information. Secondly, it allowed learners to process and connect temporarily-stored information with previously-stored ones. Lastly, working memory played a role of controlling where to direct perceptual attention and had influence on what had been temporarily-stored or rehearsed. In terms of means to measure language learners’ working memory capacity, different types of tests have been utilized, such as span tasks, nonword repetition tasks, and so on (Skehan, 2013). According to Miyake and Friedman (1998), working memory was conceptualized as a component of foreign language aptitude; there have been many studies conducted to examine the interaction between working memory and L2 learning. Particularly, Kormos and Safar (2008) examined the relationship between phonological short-term memory, working memory capacity, and learners’ Cambridge test performance. Kormos and Safar (2008) indicated that working memory was vital to learners at beginner level. With pre-intermediate learners, phonological short-term memory played an important role in test performance. Also, Hummel (2009) explored the impacts of phonological memory on L2 proficiency. Hummel’s (2009) results suggested that phonological memory significantly correlated with L2 vocabulary performance. With the same focus on language aptitude components, Winke (2013) found that rote memory contributes the most and working memory the least in language aptitude. In short, the previously reviewed studies have strengthened the importance of working memory, a component of language aptitude, in SLA. The interaction between aptitude and SLA will be explored in a more detailed way in the next parts

Foreign language aptitude and SLA

The purpose of Wesche’s (1981) seminal article was to match adult learners with suitable training approaches depending on their language aptitude. Conducted years later, Hwu and Sun (2012) explored the interaction between language aptitude and instructional approaches. In other words, the two papers investigated how learners’ language aptitude interacted with learning situations. To trace the development between the two mentioned articles, this part respectively reviews the connection between language aptitude instructional approaches of grammar, feedback, and L2 speaking ability.

The interaction of language aptitude and instructional approaches of L2 grammar

Between the article of Wesche (1981) in which the overall effectiveness of suitable training approaches was demonstrated and the study of Hwu and Sun (2012) which examined the mediation between language aptitude and different instructional approaches, several studies have been conducted to explore this aspect of language aptitude, providing insights into the facilitative effectiveness of L2 grammar instructional approaches and learners with different level of aptitudes.

More specifically, Erlam (2005) examined the relationship between three instructional approaches (deductive, inductive, and structured input instruction) and language aptitude of the participants studying L2 French in New Zealand. To measure the participants’ language aptitude, Erlam (2005) utilized sections of MLAT (language analytic ability and phonemic coding ability), working memory test, and object pronouns in L2 French as the target feature. The learners were divided into three groups receiving different teaching methods: deductive instruction, inductive instruction, and structured input instruction. Overall, Erlam’s (2005) study indicated that deductive approach was the most superior, following by structure input approach and inductive approach. For the deductive group, no significant correlation was revealed between language aptitude and test scores. According to Erlam (2005), it could be interpreted that deductive approach, in which explicit rules and focus-on-form activities were applied into teaching, seemed to bring benefits to all language learners regardless their language aptitude. For the inductive and structured input instruction groups, Erlam (2005) proved that these two methods were beneficial to learners with high analytical ability and the evident correlations were shown in their written performance. Also, Erlam (2005) suggested that structured-input instruction was beneficial for learners with good working memory capacity regarding target feature production.

As previously mentioned, the results of Hwu and Sun (2012) indicated that the participants receiving EI instruction significantly improved when being compared with those in DE group, leading to the conclusion that EI approach would be beneficial for learners with high ability of memory for text. With the same data set of Hwu and Sun (2012), Hwu, Pan, & Sun’s (2014) was conducted with similar aim of exploring the interaction between language aptitude and instructional approaches. However, the latter study perceived language aptitude as a term with many components instead of solely focusing on memory for text ability that the former study had done. The results from Hwu, Pan, & Sun (2014) demonstrated the positive impacts of language aptitude on learners’ performance. Also, Hwu, Pan, & Sun (2014) suggested that low aptitude learners performed significantly better with deductive condition in the correction test.

Generally, the studies of Erlam (2005), Hwu and Sun (2012), as well as Hwu et al. (2014) brought some findings for effective instructional approaches regarding L2 grammar. Specifically, it was suggested that learners with high analytic ability would be beneficial from either inductive or structured-input method. Besides, structured-input approach was also proved to benefit learners with good working memory capacity. On the other hand, deductive approach was proved to be suitable with learners whose language aptitude was low. However, according to Erlam (2005), the working memory measurement applied in the study was not flawless as it only captured information processing rather than information storage and processing.

The interaction of language aptitude and feedback

Although not directly relating to the study of Wesche (1981) and that of Hwu and Sun (2012), feedback – a form of responses to language learners’ production containing an error, has always been an attractive area to SLA researchers. More specifically, some studies (Han, 2002; Lyster, 2004) have been done with the attempt to explore the impacts of corrective feedback on language acquisition. Regarding the relationship between language aptitude and feedback, remarkable studies have been conducted, providing a great deal of facilitative effectiveness of feedback on learners with different aptitude levels. The findings ranged from general benefits of oral feedback to high aptitude learners (Dekeyser, 1993; Havranek & Cesnik, 2001) to details benefits of written corrective feedback and learners with high analytic ability (Sheen, 2007; Yilmaz, 2013).

More specifically, Dekeyser (1993) explored how variables of previous achievement, language aptitude, anxiety, and motivation had impacts on oral feedback. The study was conducted in classroom setting with high school students whose L1 was Dutch and studying French as L2. The participants were asked to complete grammatical sensitivity test, questionnaires of extrinsic motivation and anxiety, as well as fill-in-the-blank Grammar test as grammatical achievement measurement. Regarding feedback, it was operationalized as general error correction in which the instructors correct the learners’ errors. Dekeyser’s (1993) study indicated that error correction was beneficial with learners who had one of the following criterion: high previous achievement, high language aptitude, high level of extrinsic motivation, or low level of anxiety. In other words, from the view of connection between language aptitude and feedback, Dekeyser’s (1993) provided some positive correlation between these two factors. However, there are some drawbacks of Dekeyser’s (1993) study. Firstly, Dekeyser (1993) did not clearly divide feedback into different categories. Secondly, no specific linguistic features were focused in Dekeyser’s (1993); instead, the study covered and analyzed all linguistic features that were corrected by the teachers.

Another study exploring the interaction between corrective feedback and language aptitude was done by Havranek and Cesnik (2001). The participants of Havranek and Cesnik’s (2001) study consisted of 207 learners at different proficiency levels, ranging from ten years old to university students; they were told to complete verbal and nonverbal intelligence test, questionnaires of motivation, anxiety, and attitudes towards speaking English as well as oral correction. Similar to Dekeyser’s (1993) study, Havranek and Cesnik (2001) operationalized feedback as general oral correction. However, in Havranek and Cesnik’s (2001) study, oral correction did not only come from the teacher; correction from peers was also counted. The results of the study demonstrated that learners with high language ability and positive attitude towards feedback were beneficial from oral correction. Havranek and Cesnik’s (2001) evidence regarding aptitude-interaction was in line with what had been proved by Dekeyser (1993). In other words, there was clear evidence that corrective feedback benefited learners with high language ability. Nevertheless, there are some critics for Havranek and Cesnik’s (2001) study. Firstly, the participants were mixed-aged and mixed-level; therefore, it was unlikely to conclude that oral correction was equally useful for the participants. Secondly, similar to Dekeyser’s (1993), this study did not categorize corrective feedback into different types. Moreover, Havranek and Cesnik (2001) did not focus on one specific target structure. Instead, all structures have been taken into analyzing.

To address the mentioned shortcomings of Dekeyser (1993) as well as Havranek and Cesnik (2001), Sheen (2007) conducted a study in which the notion of corrective feedback and language aptitude were narrowed down. Particularly, Sheen’s (2007) study explored the connection between written corrective feedback and language analytic ability. The participants taking part in the study consisted of ESL learners at intermediate level of six classrooms; they were divided two three groups: control group and two experimental groups. Regarding feedback of the former experimental group, language error was pointed out and a correct linguistic feature was provided. For the latter experimental group, besides that the error was also pointed out as well as a correct element was supplied, the metalinguistic explanation was also given. Instead of taking all linguistic features into consideration as previous studies, Sheen (2007) focused on English articles. The finding of Sheen (2007) pointed out that corrective feedback was effective to ESL learners in general. Also, Sheen’s (2007) result suggested that high analytic ability learners were beneficial from both types of feedback. However, metalinguistic feedback benefitted them more evidently. One given explanation provided by Sheen (2007) was that language analytic ability was perceived as learner’s ability to acquire explicit knowledge; therefore, they were more beneficial from feedback with metalinguistic awareness provided.

One more study of this area was Yilmaz’s (2013), in which the connection between language analytic ability, role of working memory capacity, and two types of feedback (explicit correction and recast) were examined. The participants of Yilmaz’s (2013) study L1 English adults, they were divided into three groups (explicit-correction group, recast group, and control group). Regarding language aptitude test, the participants were asked to complete LLAMA R and operation span task. Also, oral production, comprehension, and recognition tests were applied to measure their performance. Yilmaz (2013) revealed three findings. First, both working memory capacity (WMC) and analytic ability (LAA) were found to have impacts on feedback of the two targeted features. Secondly, recasts were reported either work with low WMC learners or high LAA learners in plural feature. On the contrary, explicit correction worked for both high and low WMC or LAA. Lastly, with learners with high LAA, although recasts and explicit feedback worked, explicit correction showed superior effectiveness.

The interaction of language aptitude and L2 speaking

In one of the variables to investigate the matched versus mismatched analytical participants, Wesche (1981) conducted the achievement test with oral expression ability. For this reason, the connection between language learners’ aptitude and their L2 speaking ability is reviewed. There have not been many studies conducted in this area, and most of the below reviewed studies focus on learners in EFL context. Generally, the findings (Smemoe & Haslam, 2012; Winke, 2013; Saito & Hanzawa, 2016) pointed out an association between language aptitude and ESL/EFL learners, although the connection was not indeed clear due to some reasons such as other variable involving, and so on.

One study of this type was conducted by Smemoe and Haslam (2012) with the goal of exploring the interaction between language aptitude and pronunciation strategies as well as pronunciation achievement usage in ESL and EFL context. There were two groups of participants joining Smemoe and Haslam’s (2012) study: one in EFL context and another in ESL one. Along with PLAB to capture two groups’ language aptitude, pretest and posttest design with two tasks (three sentences with difficult phonemes for English learners and open-ended question responding task) were utilized to test the learners’ pronunciation proficiency. Smemoe and Haslam (2012) found no significant effects of language aptitude nor learning context (EFL/ESL) on learners’ pronunciation strategy usage as well as on pronunciation gains. However, Smemoe and Haslam’s (2012) study indicated several interesting findings. First, the analysis revealed that the ESL-high-aptitude learners and EFL-low-aptitude learners tended to use practicing and hypothesis testing strategies in comparison with ESL-low-aptitude and EFL-high-aptitude ones, which was explained as the different role that language aptitude played in the two learning contexts. In other words, in the EFL context lacking with target language exposure, learners with low level of aptitude used more strategies to make up for the lack of language ability.

One of the hypothesis that Winke (2013) tested was whether language aptitude had impacts on L2 learning both directly and indirectly. The participants were English native adults who were studying Chinese as L2. Each participant was asked to complete MLAT, the working memory span test, the motivation questionnaire, SILL test to measure strategy usage. The results from Winke’s (2013) study suggested that among factors of the model, aptitude, strategy use, and motivation impacted similarly on language learning. In other words, according to Winke’s (2013) research, aptitude did not show any superior effects on second language learning when comparing with other variables. However, Winke (2013) claimed that aptitude had direct impacts on speaking more than it did with other skills.

Also investigating speaking ability of second language learning, Saito and Hanzawa (2016) examined how the variables of length and focus of instruction, frequent L2 conversation, aptitude, and motivation can predict L2 learning outcome of late learners. The data of Saito and Hanzawa’s (2016) study was collected from Japanese EFL learners. The participants, who were university students, were asked to complete a speaking task, questionnaires of length and focus FL instruction as well as motivation, and LLAMA language aptitude test. The results of Saito and Hanzawa’s (2016) study indicated that all the variables (length and focus of instruction, frequent L2 conversation, aptitude, and motivation) related to Japanese EFL learners’ speaking performance, which meant that aptitude did not solely demonstrate any clear evidence relating to oral ability in foreign language learning.

To summarize, the three reviewed studies confirmed the connection between learners’ language aptitude and their L2 speaking performance. The findings pointed out that the interaction between aptitude and L2 speaking did exist; however, aptitude was not the only factor influencing L2 speaking performance. Moreover, three reviewed studies were conducted in EFL context, only one group of Smemoe and Haslam (2012) was learners in ESL context. For this reason, although it has been proved that strategies were used more in EFL context to compensate the lack of natural exposure, there was still lack of evidence to conclude that ESL learners used fewer strategies than their counterparts.

Language aptitude and age

Although the participants of both Wesche’s (1981) article as well as Hwu and Sun’s (2012) study, learners’ age has been a controversial issue in SLA. Particularly, critical period (Lenneberg, 1967) has been known as an important factor in SLA. Critical period hypothesis (reviewed by Ortega, 2009) suggested biology constraint on L2 learning outcomes. Whereas, language aptitude has been known as a predictor to the success of language learners. Taken these two variables into considerations, several studies have been conducted to examine the connection between learners’ age, language aptitude, and their L2 performance. On the one hand, some studies (Harley and Hart, 1997; Dekeyser, 2000; Abrahamsson & Hyltenstam, 2009) pointed out the interaction between language aptitude and late learners. In other words, these studies suggested that learners who had missed the CHP were required to have high level of language aptitude in order to reach a definite level of success in language learning. On the other hand, a part of results from Abrahamsson and Hyltenstam (2009) as well as the study of Granena (2014) explored that there also was an interaction between language aptitude and early learners when it came to language features that were sensitive to CPH.

More specifically, Harley and Hart’s (1997) study aimed to investigate the interaction between language aptitude and L2 outcomes of learners with different language immersion age. Harley and Hart (1997) conducted the study with eleventh-grade English students of L2 French early continuing and late immersion. Associate memory, memory for text, and analytical ability test were applied to measure the learners’ language aptitude. Also, the participants’ L2 proficiency was measured with five types of tests (vocabulary, listening, cloze texts and written production, as well as individual oral test). From the findings, Harley and Hart (1997) pointed out the connection between learners’ language aptitude and their L2 proficiency. Also, Harley and Hart’s (1997) result suggested that early immersion learners tended to depend on memory, while their counterpart – late immersion group showed clear evidence on analytic ability. With the result that analytic ability played a role in late learners’ L2 success, Harley and Hart (1997) mentioned that it could be the case of different teaching programs. More specifically, English was taught in a holistic view to a group of early learners. However, for the late immersion group, English was considered as a code and taught more explicitly, which explained for the result of the study. Overall, Harley and Hart (1997) pointed out positive influence of analytic ability on L2 learning. It shed a light to the fact that late learners with high level of analysis ability could perform significantly better as early learners.

In order to examine the interaction between analytic and leaners’ age, Dekeyser’s (2000) study focused on how aptitude and age affected learners to reach L2 near-nativeness competence. The participants consisted of L1 Hungarian who have various length of residence in a speaking-English country. Dekeyser (2000) utilized Words in Sentences part of MLAT to capture the participants’ analytic ability. Furthermore, the participants were asked to complete Grammatical Judgement Task (GJT). Dekeyser (2000) indicated that the participants whose age on set after fifteen and high GJT scores were reported to have high level of analytic ability. In other words, it could be concluded from Dekeyser’s (2000) study that high analytic ability, a component of language aptitude, played a role to help late learners reach near-nativeness level of English. This finding of Dekeyser (2000) was also in line with which was proved by Harley and Hart (1997).

With the aim to test Dekeyser’s (2000) results, Abrahamsson and Hyltenstam (2009) focused on investigated the association between learners’ language aptitude and their level of near-nativeness in L2 proficiency. The participants recruited for Abrahamsson and Hyltenstam’s (2009) study were L1 Spanish speaker who moved to Sweden for many years; they considered themselves as potentially native in Swedish. Based on age onset, the participants were divided into two groups. In order to measure language aptitude, LTAT (Meara, 20050 was utilized. Besides that, the participants also completed GJT task to record their L2 performance. Abrahamsson and Hyltenstam (2009) pointed out that the late-learner group had higher aptitude in comparison with young learner group, suggesting that aptitude was necessary for late-learners to reach near-nativeness level in L2. However, Abrahamsson and Hyltenstam (2009) also noted that when it came to a demanding GJT task, late-learners’ performance did not reach the level of near-nativeness level. In other words, their scores were below the level of native-speaker range in a demanding GJT. This finding leads to a conclusion that although there was association between late-learners’ aptitude and their L2 proficiency in terms of being near-nativeness, the effects of language aptitude could not be generalized to all actual, nativelike intuition. Also, in demanding GJT task, Abrahamsson and Hyltenstam (2009) pointed out that even young learners felt to be “short of actual nativelikeness”, suggesting the possibility of the interaction between language aptitude and young learners.

It could be noticed that the studies of Harley and Hart (1997), Dekeyser (2002), as well as Abrahamsson and Hyltenstam (2009) indicated the connection between language aptitude, age, and level of near-nativeness. More specifically, Harley and Hart (1997) as well as Dekeyser (2002) touched upon a specific component of language aptitude, analytic ability, to explain for its effects on late-learners’ L2 performance. In a more general view, Abrahamsson and Hyltenstam (2009) used the whole aptitude test to give explanation for the interaction between late-learners’ aptitude and L2 proficiency. Besides, Abrahamsson and Hyltenstam (2009) also brought up an important factor of aptitude, saying that it impacts on language performance was undeniable; however, the effects could not be generalized to all extent. Also aiming to examine the connection between language aptitude and L2 domain, Granena and Long (2012) did a study with L1 Chinese moving to Spain. The participants were divided into three groups, depending on their age on set (3-6, 7-15, and 16-29). Regarding language aptitude, LLAMA was utilized. Also, phonology test, lexis and collocation test, as well as morphosyntax test (through GJT task) were conducted. The findings from Granena and Long (2012) suggested that with late group (age on set rani from 16 to 29), there was an interaction between aptitude and L2 pronunciation, as well as lexical and collocation. In other words, a conclusion that aptitude played a role in late-learners’ pronunciation, lexis and collocation features was drew, contributing one more evidence to the previous studies regarding aptitude, age, and L2 performance. However, despite the fact that the connection between aptitude, pronunciation, lexis and collocation of late-learners was reported, the association between aptitude and morphosyntax of young learners was not mentioned by Granena and Long (2012). Therefore, the question of relevance between these two factors remained unclear.

With the aim to fill the gap, Granena (2014) explored how language aptitude had impacts on morphosyntax attainment of early childhood learners. The learners taking part in Granena’s (2014) study were L1 Chinese living in Spain; whose age of arrival ranged from three to six and length of residence varied from elven to twenty-eight). Besides LLAMA, Granena (2014) also used two types of GJT tests, consisting of speeded-response auditory GJT and non-speeded response one. Granena’s (2014) study demonstrated several findings. Firstly, the connection between aptitude, test, and language structure was found. Secondly, the result of the study showed that the participants performed significantly better on structures which involve no grammatical agreement. Also, according to Meisel (2009), early maturational changes affected on some structures relating to inflectional morphology (for instance gender, number, and subject-verb agreement). For these two reasons, Granena (2014) suggested that language aptitude played a role for early learners in this case. This finding of Granena (2014)’s study was also in line with that of Abrahamsson and Hyltenstam (2009)’s article, in which language aptitude was proved to affect not only late learners but also young ones.

In short, it could be observed that research of connection between language aptitude and age has brought a great deal of findings. Initially, it was reported that early learners showed more evidence in memory while late learners strongly showed the ability of language analysis (analytic ability). Also, it was suggested that aptitude only played a role in late learners’ L2 learning (Dekeyser, 2000; Abrahamsson & Hyltenstam, 2009; Granena & Long, 2012). Specifically, for late learners who missed the sensitive period, language aptitude was believed to help them reach near-nativelikeness to some extent, or perform well on L2 pronunciation, lexis, and collocation. In other words, at first, language aptitude was believed to interact with only late learners; no vivid evidence was found how this ID factor associated with young learners. However, the findings of Abrahamsson and Hyltenstam (2009) and Granena (2014) pointed out evidence which language aptitude also interacted with young learners in some special grammar features, although more studies would be necessary to clarify this point. Abrahamsson and Hyltenstam’s (2009) and Granena’s (2014) at least suggested that the association of language aptitude did not only occur after sensitive period. Instead, it showed the relevance before the critical period as well.

Language aptitude and motivation

The findings from some reviewed studies suggested a connection between these two ID factors: language aptitude and motivation. Particularly, Pimsleur (1996) included the question regarding learners’ interest in language learning into PLAB, which meant that motivation was considered either to be a part of aptitude or influence aptitude. However, this view was not acceptable as according to Gardner and MacIntyre (1992), who believed that language aptitude and motivation were two distinguishable factors. However, Winke (2013) pointed out a negative correlation between language aptitude and motivation. In other words, according to Winke (2013), learners with low language aptitude would need more motivation, which was in line with Dörnyei’s (2005) view. There are several studies in which motivation and language aptitude were reported as factors influencing language learning. Specifically, Dekeyser (1993) noted that motivation, language aptitude, along with low anxiety would be beneficial for error correction. In addition, Saito and Hanzawa (2016) mentioned that in the hypothesized model, motivation and language aptitude are factors impacted speaking performance of Japanese EFL learners. From the opposite viewpoints of Pimsleur (1996) and Gardner and MacIntyre (1992), it was controversial whether motivation should be a component of language aptitude or not. However, the findings of Dekeyser (1993), Winke (2013), as well as Saito and Hanzawa (2016) noted that language aptitude and motivation interacted with each other. Although it has been proved that there was an interaction between motivation and language aptitude, the level of association between these two factors were not answered clearly.

Conclusion

To summarize, between Wesche’s (1981) seminal article and Hwu and Sun’s (2012) study, there was a development of language aptitude measurement. Firstly, a variety of tests to capture language learners’ aptitude have been developed, providing deeper understanding of the construct as well as more choices on the tool to operationalize foreign language aptitude. Secondly, the studies conducted between Wesche (1981) and Hwu and Sun (2012) have shed light to specific effectiveness of different instructional methods. Thirdly, it has been clarified that there was an interaction between language aptitude and speaking in both ESL and EFL contexts. Fourthly, the studies examining the association of aptitude and young learners as well as late learners have supplied some remarkable findings. It was commonly known that late language learners who missed CPH required to have high level of aptitude to succeed in learning a foreign language, whereas the impacts were not clearly found in young learners. However, the connection between aptitude and young learners were also found when it came to linguistic features that were sensitive to CPH. Last but not least, the evidence of interaction between language aptitude and motivation was found. Thus far, it could be observed that a number of findings have been found regarding language aptitude, drawing the line for pedagogical implications and future directions.

Pedagogical implications

Through the findings of foreign language aptitude research, insights into L2 teaching and acquisition have been offered. As mentioned in the initial part of the paper, Wesche’s (1981) article, although not explicitly, was one of the first papers offered the facilitative effects of matching learners’ language aptitude with suitable instructional approaches. More latter studies started focusing on exploring the mediation between language aptitude and different effectiveness of instructional approaches. Although the reviewed studies sharing things in common as well as differences, it is undeniable that understanding language learners’ aptitude profile will be beneficial for L2 instructors in terms of choosing appropriate instructional approaches. For instance, for a L2 group which the majority of learners has high analytic ability, the teacher could either go with either inductive approach or structured-input method, as these two methods were proved to be beneficial with this type of language learners (Erlam, 2005). Also, in a group with mixed level of language aptitude, it is safe to choose deductive approach as Erlam (2005) mentioned that this method worked with learners regardless their language aptitude.

Regarding feedback, Dekeyser (1993) as well as Havranek and Cesnik (2001) proved the overall effectiveness of error correction in second language learning. However, it cannot be interpreted that any types of feedback and all error correction will be beneficial to all types of language learners. More specifically, according to Yilmaz (2013), when it comes to learner groups with either high working memory capacity or high language analytic ability, explicit feedback does work better than recast. Also, Sheen (2007) noted that it would be more efficient if corrective feedback focused on one specific grammar feature instead of covering all of the errors that learners made. Moreover, the finding of Sheen’s (2007) study offered a pedagogical implication, which is that corrective feedback is facilitative when correction and metalinguistic explanation are provided to respond learners’ errors.

In view of L2 speaking ability, Smemoe and Haslam (2012) suggested that no matter the learning context was ESL or EFL, teaching pronunciation strategies would be beneficial to language learners. More importantly, the finding of Smemoe and Haslam’s (2012) paper mentioned that low aptitude learners in EFL context tended utilize more strategies than their counterparts in ESL context so as to compensate for the lack of target language exposure. This finding could offer an essential implication in pedagogy for EFL teachers. It is considered that teaching speaking strategies is necessary; however, providing low aptitude learners with a variety of speaking strategies can be beneficial to help them improve the skill in a context which they do not have a lot of opportunities to communicate with native speakers.

Future directions

Although research of language aptitude has brought considerable findings and offered new insights into SLA, there have been still aspects remaining unanswered, which provides directions for future research. First and foremost, despite the fact that the construct of language aptitude and its operationalization was well-understood in some extent, the validity of the tests capturing learners’ language aptitude is still challenging. In particularly, although a variety of tests have been developed and consisted of more components, MLAT still holds the superiority owing to its validity. Therefore, validity of other test types should be taken into consideration in the future.

Secondly, there have not been many studies examining the interaction between language aptitude and L2 speaking; therefore, this area could be a direction for future research. This reminds me of the Vietnamese EFL context that I am familiar with. It might be possible to conduct a study relating to Vietnamese EFL learners’ aptitude and their usage of speaking strategy in order to explore whether in the environment lacking target language exposure, strategies used in speaking would help learners to make up for the lack of language aptitude. Moreover, the reviewed studies mainly examined the connection between language aptitude and speaking strategy use in EFL context, while which of ESL context was not thoroughly explained. Hence, more studies investigating the interaction between aptitude and L2 speaking are necessary.

Thirdly, reviewed studies showed that language aptitude plays a role in late language learners’ success. Additionally, the interaction between language aptitude and young learners was also found. Although the researchers realized that language aptitude interacted with young learners when it came to some language features which were sensitive to CPH, the issue needs more future research focusing on examining this type of structure to clarify whether language aptitude can compensate critical period and if aptitude is helpful for young learners to absorb these structures. Last but not least, as language aptitude was believed to correlate with motivation and there has been a controversy regarding motivation as a component of language aptitude, the association between these two ID factors worth further examination.