NAN JIANG

University of Maryland

An adequate characterization of L2 learners’ competence should include its three dimensions: accuracy, appropriateness, and automaticity. Much progress has been made to assess L2 learners’ appropriate use of an L2 in more recently years. In contrast, much less attention has been given to the assessment of automaticity. This paper attempts to achieve four goals: characterize L2 competence, explain the importance of automaticity in second language acquisition research, discuss the limitations of current assessment methods and their associated problems in research, and outline a set of test methods that may be more effective in assessing automatic competence in a second language. Studies in which these methods were employed are reviewed.

INTRODUCTION

The assessment of the learner’s competence in a second language (L2) is of primary importance in second language acquisition (SLA) research. The assessment method one uses in a specific study often affects how a research question should be formulated and how the study should be designed. Whether the learners’ L2 competence or learning outcome is adequately assessed will also determine how the findings should be interpreted and the extent to which they address the research questions involved. Considered in a broader perspective, the adequacy of L2 competence assessment is a determinant of how successful SLA as a field of inquiry is in reaching its goals, particularly in understanding the nature and characteristics of L2 learner’s acquired knowledge or competence.

A growing inventory of tools, techniques, and methods have been used for assessment purposes in SLA research. The assessment of L2 acquisition used to rely heavily on the analysis of learners’ errors in L2 production in the 1970s when SLA was just beginning to emerge as an independent field of inquiry (Dulay and Burt 1972; Krashenetal. 1977; Winogradetal. 1976). Now a variety of methods are being used depending on the need of the researchers and the research questions involved, such as the use of grammaticality judgment in UG-based research (Banks and White 1985; Bley-Vromanetal. 1988), visual-based retelling tasks frequently used in the study of the role of instruction in SLA (e.g. Muranoi 2000; Salaberry 1997; Erlam 2003), think-aloud protocols in studying L2 writing (Johnson 1992; Raimes 1987; Wang and Wen 2002), the discourse completion task in assessing pragmatic competence (Billmyer and Varghese 2000; Byon 2004; Golato 2002), and online reaction time techniques in studying L2 representation and processing (Jiang 2004; Robinson and Katayama 1997; Segalowitzetal. 1998).

With all the progress that has been made, a long-standing problem persists: the assessment of L2 learners’ achievement or L2 competence often makes no distinction between linguistic knowledge that has and has not be automatized. The purpose of this paper is to explain the limitations of the methods widely used today in assessing L2 competence, and to offer tasks and methods that are of potential use in remedying the problem. I begin by highlighting the importance of automaticity in assessing L2 competence in a three-dimensional model of L2 competence.

THREE DIMENSTIONS OF L2 COMPETENCE

The adequacy of L2 assessment depends on an adequate characterization of L2 competence. A number of characteristics of L2 competence have been the focus of attention in previous research. For example, early interlanguage and morpheme order research (Baileyetal. 1974; Schmidt 1980) emphasized its autonomous nature in that the learner’s competence was considered to be a unique, developing, and creative system in its own right, rather than a flawed L2 system contaminated by the learner’s first language (L1). The distinction of explicit and implicit knowledge in L2 (Bialystok 1978), on the other hand, highlights the fact that the linguistic knowledge L2 learners possess is much less unitary than that of native speakers. L2 learners’ linguistic knowledge consists of two different types of linguistic knowledge that differ in whether it can be used subconsciously or automatically in spontaneous speech. The fossilization literature (Han 2004) captures a major difference between L1 and L2 linguistic knowledge: invariable native-like competence in L1 and almost inevitable outcome of non-nativeness in L2. Much of the interlanguage variation research (Bardovi-Harlig 1992; Young 1988) demonstrates the incremental and context-sensitive nature of L2 competence and use.

Whereas these different approaches highlight different aspects or characteristics of L2 learners’ linguistic competence, a general framework of L2 competence can be proposed in which L2 competence consists of three dimensions: accuracy, appropriateness, and automaticity. Accuracy is the most basic component of L2 competence. From a functional perspective, inaccuracy may not always interfere with communication, but from the perspective of acquisition research, an acquired structure is a structure that can be used in its correct forms. Thus, accuracy is the minimal requirement of acquired competence. Much of SLA research has focused primarily on the accuracy dimension, from error analysis to morpheme acquisition, from interlanguage variation to language transfer, from the examination of the development of different aspects of language (i.e. pronunciation, vocabulary, morphology, and syntax) to the study of various L2 skills. Whether a learner is able to use a target structure correctly is the primary concern and often the only criterion for assessing acquisition in such research.

In real-life language use, successful communication often requires not only the use of correct forms but also the appropriate use of the correct forms in a specific context. Thus, a structure should be considered acquired only when it is used to express the user’s intention in a way acceptable in the target language community or consistent with the pragmatic and sociocultural norms of the target language community. Consequently, appropriateness should be considered as an important aspect of L2 competence. This second dimension of L2 competence emerged as sociolinguistic theories, such as the concept of communicative competence, started to influence applied linguistic and SLA research in the 1970s (Ager 1976; Hornberger 1989; Lutjen 1973; Paulston 1974; Richards 1975; Rivers 1973; White 1974). The acquisition of L2 competence defined in terms of appropriateness requires the examination of L2 use in context and with a consideration of the learner’s communicative intent. This dimension plays a prominent role in SLA areas such as L2 discourse analysis, development of pragmatic competence, and cross-cultural communication (Cohen and Mehler 1996; Kasper 1979, 2001; Lelouche and Huot 1998; Nelsonetal. 2002; Pica 1988; Safont Jorda 2003; Takahashi 1996; Walters 1980).

The third, and often neglected, dimension of L2 competence is automaticity. Whereas the second dimension deals with the sociolinguistic aspect of language competence, the dimension of automaticity is closely related to the psycholinguistic aspect of L2 competence. Automaticity can be defined in many different ways (see Segalowitz 2003 for an excellent review of various attributes of automaticity as defined in cognitive psychology). In the present context, automaticity highlights the ability to use linguistic structures without paying conscious attention to them or their well-formedness. Just like accuracy and appropriateness, automaticity is an intrinsic property of human competence in language use, because automaticity is essential if language use is to be “unaffected by such grammatically irrelevant conditions as memory limitations, distractions, shifts of attention and interest, and errors (random or characteristic)” (Chomsky 1965:3) in real-life communication. The construct of automaticity as a dimension of L2 competence can be seen in the distinction of explicit and implicit knowledge (Bialystok 1978), in the application of information processing model to SLA (Kennedyetal. 1988; McLaughlin 1990), and in some L2 processing studies (e.g. Favreau and Segalowitz 1983; Segalowitz and Segalowitz 1993; Segalowitzetal. 1998). The importance of automaticity has also been recognized in the proposal that the ultimate goal of second language learning and teaching is the development of the learners’ ability to use the target language automatically as well as accurately and appropriately (e.g. Ellis 1993; Hulstijn 2001; Segalowitz 2003; Segalowitzetal. 1998).

On the basis of above analysis, one can refer to L2 competence as an L2 learner’s ability to use an L2, or part of it, accurately, appropriately, and automatically. Consequently, the development of one’s L2 competence should not be defined only in terms of the number of structures a learner knows, but also in terms of an increase in the level of accuracy, appropriateness, or automaticity in using the structures.1The three-dimensional model of L2 competence is illustrated in Figure 1.

Figure 1: The three-dimensional model of L2 competence

I hope to emphasize in the present context that the automaticity dimension is particularly important for research purposes due to a unique characteristic of adult L2 learning: the involvement of explicit learning. The learning of an L2 by most adult L2 learners or classroom L2 learners is to some extent conscious and explicit. Their initial exposure to a grammatical rule or structure is often accompanied by explicit explanation or an effort to find rules about the structure. As a result, they often have access to explicit knowledge about the target language, and their L2 use is to a great extent a process of applying and gradually automatizing explicit knowledge. However, native-like linguistic competence is much more than and qualitatively different from the application of explicit knowledge. Learning an L2 is also an enormously complicated phenomenon, of which learning, remembering, and applying explicit knowledge is a small part. The consideration of automaticity is thus obligatory for ensuring that one is assessing the learners’ ability in using the L2 spontaneously rather than their ability to apply explicit knowledge. By ignoring the automaticity dimension in assessing L2 competence, one equates L2 competence with the ability to apply explicit knowledge, reduces L2 acquisition to a memory issue, and fails to capture the complicated cognitive processes involved in L2 learning and use.

A CRITIQUE OF SOME CURRENT METHODS OF ASSESSMENT

A major threat to the assessment of automatic competence is the involvement of explicit knowledge. If an assessment method allows participants to rely on explicit knowledge, then the observed performance may be a result of applying explicit knowledge, rather than reflecting automatic competence. So to determine whether a method is adequate in assessing automatic competence, one considers whether explicit knowledge is involved in the learner’s performance. In line with this view, a structure can be considered to have been acquired only when two conditions are met. First, the non-native participants produced native-like performance. Second, these participants have no explicit knowledge about the structures under investigation, or an assessment method is adopted that is able to minimize the involvement of explicit knowledge.

There is an increasing awareness of the importance of automaticity in assessing L2 acquisition. It can be seen, for example, from the adoption of timed rather than untimed grammaticality judgment tasks (e.g. Bley-Vroman and Masterson 1989; Robinson 1997) and the use of spontaneous production tasks such as story-retelling (e.g. Muranoi 2000; Salaberry 1997; Erlam 2003) and interviews (Harley 1989; Day and Shapson, 1991). While the use of these tasks all represent attempts to discourage the reliance on explicit knowledge, they are only partly successful at most in doing so. A closer look at a few such assessment methods should help make the point clear.

The grammaticality judgment task (GJT)

The GJT first became a widely used method in the study of UG accessibility in SLA in the 1980s, and it is now used in SLA research involving a variety of topics. The traditional written, untimed GJT can be an effective method for taping the learner’s internalized intuitive knowledge about the target language when the participants have no explicit knowledge about the structure involved. That is why it is used extensively in linguistic and psycholinguistic research involving native speakers. However, when the participants do have access to explicit knowledge, as is the case with most L2 learners, this method becomes inadequate because it directs the participants’ attention to grammatical well-formedness, and motivates them to apply explicit knowledge.

Two variants of the earlier written GJT may help alleviate this problem. One is an auditory GJT such as the one used in Johnson and Newport (1989), DeKeyser (2000), and Birdsong and Molis (2001). The transient nature of material presentation in the auditory GJT makes it difficult to apply explicit knowledge. A potential problem, though, is that it requires well developed skills in the listening mode, and thus, the participants may underperform due to limited listening comprehension skills, as demonstrated in the Johnson (1992) study. This is particularly true when the target structure is one of limited physical salience, such as grammatical morphemes. The second variant is the timed GJT in which the participants are asked to perform the task as quickly as possible (e.g. Bialystok 1979; Bley-Vroman and Masterson 1989; de Graaf 1997; Ellis 1993; Han and Ellis 1998; Juffs and Harrington 1995; Robinson 1997). The data may include their accuracy rates alone, or both accuracy rates and response times. While individuals may be less likely to apply explicit knowledge under time pressure, it is very unlikely that increasing time pressure will completely eliminate the involvement of explicit knowledge. As DeKeyser suggested, the use of time pressure “merely made the use of explicit knowledge more difficult, and not impossible” (2003: 326).

Meaning-based production task (MPT)

A variety of tasks have also been used that require participants to communicate ideas through productive use of an L2. These meaning-based tasks may include picture description (Baileyetal. 1974; Wei 2000; Whiteetal. 2004; Yang and Givon 1997), retelling based on previously shown film or video clips (e.g. Muranoi 2000; Salaberry 1997; Towelletal. 1996) or slides (Lyster 1994), and interviews (Bardovi-Harlig 1992; Herschensohn 2001; Prevost and White 2000; Wei 2000). A major difference between the MPT and the GJT is that the former focuses the participants’ attention on meaning whereas the latter focuses their attention on form. In this regard, the MPT is a better assessment approach in limiting the involvement of explicit knowledge. However, adopting a spontaneous task like the ones mentioned here is no guarantee in itself that no explicit knowledge is involved in the participants’ performance, at least for two reasons. First, it is difficult to determine and verify that L2 participants are not dependent on their explicit knowledge in such a task, at least to some extent. The finding that L2 users tend to be less accurate in spontaneous speech than in self-paced writing (e.g. Salaberry 2000) seems to indicate less involvement of explicit knowledge and monitoring in spontaneous speech, but it does not mean no involvement of explicit knowledge. Second, there is much evidence showing L2 learners do monitor their output in spontaneous meaning-oriented L2 production (e.g. Ortega 1999). Some studies showed that when planning time is given, sometimes as short as one minute, accuracy can improve in L2 production that can still be characterized as spontaneous (Mehnert 1998; Skehan and Foster 1999; Yuan and Ellis 2003). It is reasonable to suggest that such improved accuracy is at least partly a result of explicit knowledge involvement in planning and monitoring L2 output. After all, in an L2 learning situation, accuracy is always a desirable goal and also something many L2 speakers are consciously aware of during L2 production, so there is a motivation to rely on one’s explicit knowledge whenever possible. In short, “free-production tasks make it difficult but not impossible for learners to perform on the basis of explicit knowledge.” (Ellis 2002:234)

Both the GJTs and MPTs may be effective in assessing L2 learners’ intuitive linguistic knowledge when no explicit knowledge is available to them. A good example is the use of the GJT in the study of L2 learners’ knowledge of the principle of subjacency (e.g. Johnson and Newport 1991; Schachter 1989, 1990; White 1985). However, when participants have access to explicit knowledge, these methods become less adequate. As a result, the findings obtained with such methods become less useful in addressing the research questions. The problem is particularly apparent in two research areas: the acquisition of grammatical morphemes and the role of instruction in SLA.

Morpheme acquisition studies

The acquisition of grammatical morphemes by L2 learners has received much attention in SLA research. At least four distinct areas can be identified in which the acquisition of grammatical morphemes is examined in relation to their respective research questions. They are morpheme acquisition or acquisition order (Baileyetal. 1974; Ghrib 1987; Pica 1983), tense and aspect (Andersen and Shirai 1994; Bardovi-Harlig 1992; Collins 2002), critical period research (e.g. Birdsong and Molis 2001; Johnson and Newport 1989; DeKeyser 2000), and recent UG-based SLA research (Beck 1997; Lardiere 1998; Prevost and White 2000). Such research inevitably involved the assessment of whether the target grammatical morphemes have been acquired. Most of such studies employed one or more of the methods discussed earlier. The likelihood of explicit knowledge involvement in the participants’ performance in these tasks makes the findings very difficult to interpret. For example, when a group of L2 learners show an accuracy rate of 85% in morpheme use, which is common in studies that relied on spontaneous L2 production for data collection, or when the same participant produces an near-native accuracy rate in one task, e.g. in writing, and a much lower accuracy rate in spontaneous speech, such as the case with Krashen and Pon’s (1975) subjectP, two alternative interpretations become possible. One can argue, as Krashen (1981, 1982), Johnson and Newport (1989; Newport 1990) did, that the less-than-native like performance demonstrates the lack of authentic acquisition of the target structure. The accurate uses of the structure are attributed to the application of explicit knowledge. The alternative interpretation is to consider the structure to have been acquired based on the better-than-chance performance, and attribute the errors to temporal processing difficulties, like Sharwood Smith (1986) and Prevost and White (2000) did.

The role of instruction in SLA

One faces a similar problem in the study of the role of instruction in SLA, a topic that has received much attention in SLA research (see Norris and Ortega 2000). While these studies may vary in the type of participants tested, the target structure involved, the treatments provided, and the specific methods of assessment used, many of them (e.g. De Graaff 1997; Robinson and Katayama 1997; Scott 1989, 1990; Rosa and Leow 2004) shared three commonalities. First, while the specific treatments may vary across studies, the focus was often on treatments that differed in the availability or degree of explicit instruction and/or feedback. As a result, they may be collectively considered studies comparing explicit and implicit learning. Two, one or more of the tasks reviewed above were used to assess the learning outcomes. Third, the more explicit group typically outperformed the other group(s) in terms of grammatical accuracy. The findings of such studies are often considered as evidence in support of the greater effectiveness of explicit knowledge or instruction in SLA.

However, these findings do not provide any conclusive evidence in support of the superiority of explicit learning over implicit learning because of the assessment methods used. When explicit knowledge is provided in one group, but not in the other group, and when the assessment method allows one to use explicit knowledge, as the above discussion suggests, the only reasonable interpretation of the “explicit” group’s better performance is that explicit learning is superior to implicit learning in helping develop explicit knowledge. These findings say very little about the relative effectiveness of explicit vs. implicit learning in developing anything beyond explicit knowledge in L2.

The above review is intended to make two points. First, the current methods often fail, to a varying degree, to assess the learner’s automatic L2 competence because they often allow learners to use their explicit knowledge. Second, the use of such methods often does not allow one to address the research question one intends to explore.

ALTERNATIVE METHODS FOR ASSESSING AUTOMATIC COMPETENCE IN L2

Better methods for assessing automaticity are available. Before a more detailed examination of these methods, two points should be made about automaticity as a theoretical construct. First, automaticity should be defined as a dichotomy as well as a continuum. The continuum definition of automaticity is predominant in SLA research, and for a good reason (e.g. DeKeyser 1996). A learner’s L2 skill, or his or her control over an explicitly taught structure, may develop from a lower to a higher level of automaticity. Many SLA researchers are concerned with issues related to such development of automaticity, such as what factors may affect the automatic use of explicit knowledge (e.g. Hu 2002).

However, it is equally important, if not more so, to view automaticity as a dichotomous construct. If automaticity is an intrinsic characteristic of native-like linguistic competence, then one has to determine whether one has or has not become automatic in using a structure, rather than whether one has become more automatic than before. This is particularly important in understanding many issues related to the cognitive processes involved in L2 learning. For example, one has to determine whether a structure has been automatized in order to understand whether or not explicit knowledge can be developed into automatic competence through practice. One has also to find out whether a structure has been automatized in order to determine whether a particular structure is acquirable in adult L2 acquisition. In both cases, the continuum definition of automaticity is not helpful. Thus, a comprehensive view of automaticity should incorporate both the continuum and the dichotomy definitions, as illustrated in Figure 2.

Figure 2: Automaticity as both a continuum and a dichotomy

Second, automaticity is different from fluency. Fluency emphasizes observable qualities in speaking and how such qualities affect listeners. Thus, fluency is measured often in terms of speed of articulation, the number of pauses, and the location of pauses. Automaticity is a concept more related to indiscernible mental processes than observable qualities, and it is concerned with how much attention is needed while performing a task, not only in speaking but also in other modes of language use such as listening or reading. A fluent speaker may be also automatic (Schmidt 1992), but fluency and automaticity may not always go hand in hand. An L2 learner can be fluent but not automatic, for example, when he or she takes time to prepare and practice and stays highly concentrated in order to read a passage aloud at a normal speed. Similarly, automatic use of language does not always means fluent use of language. Many factors other than automaticity can contribute to hesitant non-fluent speeches in one’s native language. Thus, what is used to measure fluency in speaking may not be appropriate for measuring automaticity in language processing.

The methods to be discussed below deal with automaticity as a dichotomy, rather than a continuum. They assess automaticity by comparing participants’ reaction times and accuracy rates under different conditions, rather than performance speed per se. Some of these methods have already been used to explore what knowledge can be automatized in adult L2 learning, and some others are yet to be employed for such purposes (see Jiang 2011 for more detailed description of these and other tasks that may be used to assess automatic language use).

Shadowing

In performing a shadowing task, participants are presented with auditory input, which can be a phrase or a sentence, and are asked to repeat a designated word as quickly as possible. The target word to be repeated is indicated by various means, such as the change of voice. For example, if the auditory input is “I forgot to mail the letter” and the target word is “letter,” then the context section “I forgot to mail the” is read by a male voice and the target word is read by a female voice. The participants have to repeat the word read by a female voice as quickly as possible. Their repetition latency, i.e., the time interval between the onset of the target word in the input to the onset of articulation of the target word by the participants, is measured and analyzed as primary data.

The test materials usually consist of grammatical (or congruent) and ungrammatical (or incongruent) versions of the same phrase or sentence, such as “I bought several books yesterday” and “I bought several book yesterday.” The learners’ automatic competence is assessed by comparing their shadowing latencies for the grammatical and ungrammatical versions. The assumption underlying such comparison is that if one has developed automatic competence for a particular structure, one is able to automatically notice the error in the ungrammatical version even when they are not asked to pay attention to well-formedness of the test material. As a result, there would be a delay in their shadowing. Thus, a delay in shadowing time on the ungrammatical version is taken as the availability of automatic competence.

The usefulness of the task for assessing automaticity can be illustrated with a study by Guillelmon and Grosjean (2001) who investigated the processing of French gender marking in second language learners. In earlier research, it had been found that congruent gender marking, between a modifier and the following noun for example, would speech up the word recognition process in tasks such as gating (Grosjeanetal. 1994), lexical decision (Grosjeanetal. 1994; Jakubowicz and Faussart 1998), and shadowing (Batesetal. 1996). Incongruent gender marking, on the other hand, would produce an inhibitory effect (Batesetal. 1996).

To investigate if such gender marking effects existed in L2 learners, Guillelmon and Grosjean (2001) tested early and late English-French bilinguals in a shadowing task, or an auditory naming task, as they called it. The early bilinguals reported using their two languages regularly in childhood. The late bilinguals learned French as their second language in school and they became regular user of French at a much older age than the first group (average ages of onset of bilingualism being 5;4 and 24;8 for the two groups, respectively).

Thirty-six French nouns and their neutral, congruent, and incongruent modifiers were used to create four conditions: gender congruent and their matching neutral conditions, and gender incongruent and their matching neutral conditions. The congruent, incongruent, and neutral conditions are illustrated in the following phrases:

a. congruent:la jolie glace (the nice mirror)

b. incongruent: le joli glace

c. neutral: leur jolie glace

The participants heard the phrases over headphones and were asked to repeat the nouns as quickly as they could. Their reaction times were recorded as primary data. A group of monolingual French speakers served as controls.

The results showed both a facilitative and inhibitory gender marking effects in monolinguals, who named the target words in the congruent condition faster than those in the neutral condition, and named incongruent target words slower than neutral words. However, only the earlier bilinguals showed the same pattern as native French speakers. The late bilinguals showed no difference between the three conditions, suggesting “a total insensitivity to gender marking in perception.” (Guillelmon and Grosjean 2001: 509)

Note that in performing the shadowing task, the participants were not asked to check gender agreement. They were not asked to generate their own phrases either. All they did was to listen to the phrase and repeat the last word as quickly and accurately as they could. There was little need, reason, or time for them to utilize their explicit knowledge while completing the task. Thus, the task allows one to assess one’s internalized automatic competence with little involvement of explicit knowledge. If gender agreement knowledge, or other morphosyntactic knowledge, has been automatized, such knowledge will be automatically activated in the process of completing the shadowing task. This automatic activation will lead to a difference in reaction time for different conditions. However, if a participant has explicit knowledge about certain morphosyntactic rule, but this knowledge is not part of his or her automatic competence, then it will not be automatically activated, and they should show no reaction time difference for the grammatical and ungrammatical items.

Sentence matching

Sentence matching is a task first adopted by Freedman and Forster (1985) in studying mental representation in sentence processing within the UG framework. A test item in the sentence matching test begins with a sentence presented on the computer monitor. It remains on the monitor for a brief duration, for example 2500 milliseconds, before a second sentence appears. The participants’ task is to determine whether or not the two sentences are identical. Their response time, i.e., the interval between the appearance of the second sentence and the onset of their response, is recorded as primary data. The critical stimuli typically include matching items that may be grammatical or ungrammatical. Non-matching items are fillers. The rational underlying the use of the sentence matching task is similar to that of the shadowing task. If participants have developed automatic competence for a structure, they would automatically and often subconsciously catch the related error in the ungrammatical sentences, which would then lead to a delay in generating a mental representation for the sentence to be used for matching, and thus a delay in performing the matching task. Otherwise, their response times for the grammatical and ungrammatical sentences would be similar.

Even though a number of L2 studies have employed this method (Eubank 1993; Duffieldetal. 2002; Duffield and White 1999; Gass 2001), its usefulness is yet to be determined for a number of reasons. First, in the study by Freedman and Forster (1985), native speakers of English demonstrated their sensitivity to only certain types of errors, not all errors. Whether the same method can be used to assess automatic competence in areas other than those investigated in Freedman and Forster (1985) is yet to be determined. Second, L2 studies have produced conflicting results regarding its usefulness in assessing L2 learners’ internalized linguistic competence. Duffield and White (1999), for example, found the expected reaction time differences between grammatical and ungrammatical sentences in sentence matching while studying Spanish clitic replacement among both native and non-native speakers of Spanish. Gass (2001), however, failed to observe reliable differences between grammatical and ungrammatical items. As pointed out by Gass, a number of procedural factors may affect the test results. In our own research, for example, participants were found to match sentences on a word-by-word basis, rather than read the two sentences first and then make the comparison based on their mental representations. An effort is yet to be made to pilot and standardize the procedures if this potentially useful paradigm is to become an effective tool for assessing automatic competence in L2.

Self-paced word-by-word reading

In a self-paced word-by-word reading experiment, a test item begins with the first word of a sentence presented at the left side of the computer monitor. Participants are instructed to press a key to obtain the next word as soon as they finish reading a word, until he or she reaches the end of a sentence. Each following word appears to the right of the preceding word (which is known as the moving-window presentation). A word disappears as soon as the next word appears so that participants never see a sentence presented on the monitor in its entirety. This word-by-word reading allows the computer to record the participants’ reading time for each word. The task is self-paced in the sense that the participants decide how fast they want to proceed through the sentence. The focus of the task, as specified in the instructions given to the participants, is on the comprehension of meaning, and not on grammatical accuracy. Comprehension questions are given to assess the participants’ reading comprehension.

The self-paced word-by-word reading method has been widely used in sentence processing studies (Thorntonetal. 2000; Weyertsetal. 2002). Jiang (2004, 2007) was among the first to use the method to assess automatic competence in L2, and its use was inspired by the finding of a study by Pearlmutter, Garnser, and Bock (1999) that native speakers of English would take longer to read an ungrammatical sentences such as “The words on the screen was hard to recognize” in relation to its grammatical counterpart “The word on the screen was hard to recognize” even when they were not instructed to pay attention to grammatical well-formedness of the sentences. Jiang interpreted this finding as evidence demonstrating native English speakers’ automatic competence in processing the plural morpheme.

This led him to use the method for assessing automatic competence in L2 learners. The rationale is very similar to that underlying the use of shadowing and sentence matching. If L2 learners possesses the automatic competence for a structure, they are expected to show a sensitivity to errors involving this structure even when the task does not require them to pay attention to grammatical accuracy. This sensitivity can be assessed by comparing their reading times on grammatical and ungrammatical sentences. A delay in reading ungrammatical sentences is considered an indication of such sensitivity and in turn of the presence of automatic competence.

In two self-paced reading studies, Jiang (2004, 2007) compared native and nonnative English speakers’ reading time on grammatical and ungrammatical versions of sentences such as “The word(s) on the screen was hard to recognize” and “John saw one of his friend(s) in the room.” Native speakers showed a reliable difference in reading time between the grammatical and ungrammatical items, demonstrating the automatic activation of their linguistic knowledge involving the plural morpheme. However, Chinese ESL speakers showed little reliable difference. These findings were interpreted as evidence against the possession of automatic competence for the plural morpheme by these ESL speakers.

The self-paced reading task has several important features that can help minimize the involvement of explicit knowledge. First, it is a meaning-based reading comprehension task. The participants’ attention is focused on the meaning of the sentences, and not on their forms. Second, in this receptive task, participants are not required to produce sentences, which takes away the motivation to rely on their explicit knowledge in generating and monitoring output. Furthermore, the presentation of the sentences is such that the participants never see the entire sentence. They see one word at a time. This transient nature of material presentation, and the emphasis on reading speed and reading for comprehension make it very difficult for them to check the grammaticality of the sentence, let alone to retrieve explicit knowledge in checking grammaticality. Thus, it is reasonable to suggest that only automatized linguistic knowledge is involved in sentence processing in the self-paced reading task. The sharp contrast between the participants’ perfect performance in the pencil-and-paper test and their lack of sensitivity to number agreement violations in the self-paced reading task also provide confirmation for the lack of explicit knowledge involvement in the task. Because of these advantages, the task has been used widely since Jiang (2004) in studying the acquisition of grammatical morphemes, such as Jackson and Dussias (2009), Tokowicza and Warrena (2010), Foote (2011), and Jiangetal. (2011), to name just a few.

Word monitoring

In a word monitoring task, a participant is given a target word and asked to monitor this word while listening to sentences for comprehension. The auditory input consists of grammatical and ungrammatical sentences that involve a particular structure under investigation. To use the same examples shown earlier, the two sentences may be “The words on the screen were hard to recognize” (grammatical) and “The words on the screen was hard to recognize” (ungrammatical). The target word may be “hard” in this example. A participant is asked to press a button as quickly as possible after they hear the target word. Their reaction times are recorded as primary data. Previous research demonstrated that native speakers of English showed a delay in responding to the target word when it was presented in a ungrammatical sentence as compared to the same word presented in a grammatical sentence (Peelleetal. 2007). This finding suggested that native speakers of English were able to automatically detect an error in auditory input while listening for comprehension, which resulted in a delay in response time.

Jiangetal. (2010) applied this method in studying whether adult L2 learners were able to develop such native-like automatic competence. A group of ESL speakers with Chinese, Russian, and Spanish background were tested, along with a group of English native speakers serving as controls. The structure was tense marking, with sentences such as “Last month we purchased another company in the same area” (grammatical) and “Last month we purchase another company in the same area” (ungrammatical) as stimuli. Native speakers showed a reliable delay of 98 ms and 111 ms delay for regular and irregular verbs, respectively. However, non-native speakers showed only mathematical differences in the right direction; the differences were not statistically reliable.

We are just beginning to use the word monitoring task in the study of L2 acquisition. Many methodological details have to be worked out. For example, we found that a phrase such as “last month” or “yesterday” was more effective than a clause such as “after they went home” or “they agreed that” for indicating a past time event, and thus bringing out the effect in word monitoring data. It was also possible that the effectiveness of this method may be compromised by the fact that auditory input was used, and thus this method has the same limitation as the auditory grammaticality judgment task discussed earlier. That is, it places much demand on a participant’s listening comprehension skills. It has yet to be determined whether a visual version of the task can be more effective.

Tracking eye movement

Another method that can be used to examine one’s sensitivity to morphosyntactic violations in sentence processing is tracking eye movement. When one is engaged in reading, one’s eyes do not move smoothly in one direction (from left to right) over a sentence. Instead, the eyes may fixate on one word briefly before jumping to another word, engaged in so-called saccadic eye movements. Sometimes we may also look back at an earlier section of a sentence. An eye tracker allows one to monitor a participant’s eye movements in reading and record information such as eye fixation durations (how long one fixates on a word), regressions (when one looks back), and end points of regressions (where one looks when one looks back).

Similar to self-paced reading, the eye-tracking method also allows one to examine a participant’s reading process. Thus, it is often used together with the self-paced reading paradigm, and they often produce the same pattern of results (e.g. Kennison 2002; Pearlmutteretal. 1999). Compared to the self-paced reading paradigm, the eye-tracking paradigm has the advantages of producing much richer data and allowing the reading task to be more similar to normal reading activities.

One can thus use the eye-tracking paradigm to assess the development of automatic competence in the same way as one uses self-paced reading, i.e., by comparing an individual’s reading time on grammatical and ungrammatical sentences of various types. A morphosyntactic error in a sentence may lead to a longer fixation time often without an individual’s awareness if he or she has developed automatic competence for the related structure. This is clearly demonstrated in the study by Pearlmutteretal. (1999). Their eye movement data showed that native speakers of English took significantly longer to read ungrammatical sentences such as “The key to the cabinet were rusty from many years of disuse” compared to its grammatical counterpart. Similar findings were obtained in another eye track study by (Nietal. 1998).

Whereas many studies in L2 research have employed the eye-tracking method (see Frenck-Mestre 2005 for a review), these studies have focused on topics such as ambiguity resolution in L2 and bilingual lexical organization, rather than on the automatization of L2 knowledge. The use of this promising method for assessing automatic morphosyntactic competence in L2 is yet to be seen.

The dual-task paradigm

The employment of two concurrent tasks has been used to study a variety of cognitive processes such as skill development, working memory, and attention. In many studies where the dual-task method is used, a participant is required to perform a primary task in two conditions: with and without the simultaneous performance of a secondary task. For example, in the dual-task condition, participants may be asked to write an essay or complete semantic judgment while performing a secondary task such as listen to and recall a set of random digits, detect or count certain type of auditory signals, or count numbers at a particular interval.

Assuming that an automatic skill requires little attentional resources, the performance of such a skill would be affected by performing a concurrent task. In contrast, a less automatized skill would be more likely to be affected by the secondary task. Thus, one can compare an individual’s performance of a linguistic task with and without performing a secondary task to assess the extent to which their performance of the task has been automatized. A lack of an inhibitory effect, in terms of reaction time or accuracy, in the presence of a secondary task can be taken as an indication of an automatized skill.

Only a small number of studies have adopted the dual-task approach to assess automatic competence in L2. In one such study, DeKeyser (1997) explored the automatization of L2 rules through practice. To assess the level of automaticity, both comprehension and production tasks were performed under two conditions: single-task and dual-task. In the latter condition, the participants had to remember a three-digit number presented on a computer monitor, e.g. 236, count the number of beeps they heard, e.g., 3, subtract that beep number from the written number they saw, and report the results (e.g. 236-3=233) after they completed the linguistic task.

A challenge one faces in adopting the dual-task approach is the selection of an adequate secondary task. A task that is too easy may fail to distinguish automatic from less automatic skills because it may have little inhibitory effect even on less automatic skills. A concurrent task with a high level of complexity and difficulty may produce the same result because it utilizes so much attentional resources as to interfere with the performance of automatic skills. The adequacy of the task will also depend on what primary linguistic task is involved. Thus, while the dual-task method is a potentially useful one for assessing automaticity, its effective use may require careful consideration of the design of an individual study and much pretesting and piloting. A further limitation of the task involves the difficulty in determining how much interference from the secondary task would make the performance of a primary task non-automatic.

CONCLUSION

The assessment method one uses in a study often affects the formulation of the research question or hypothesis, the design of the study, and the interpretation of the results. Considered in a broader perspective, assessment adequacy is a determinant of how successful SLA scholars will be in their attempt to understand the nature and characteristics of L2 learners’ competence. While there is clear indication that this problem is beginning to be addressed, the emerging approaches have limitations of their own, too. For example, it is yet to be seen whether shadowing and self-paced reading can also be employed to assess automatic competence in areas other than grammatical morphemes. Event related potentials have been shown to be affective in assessing automatic competence in L1, but its application in L2 have produced less consistent findings (e.g. Hahne and Friederici 1999; Hahneetal. 2006; Muelleretal. 2005; Ojimaetal. 2005), which is why the use of this method is not included in this review. Much more collective efforts are needed from both language testing, SLA, and cognitive science scholars before a variety of such methods become available for assessing L2 learners’ automatic competence in all aspects of language.

NOTE

1 It has to be pointed out, however, that the appropriateness dimension becomes relevant only when one deals with sociolinguistic issues of L2 acquisition and use. When one’s primary concern is the development of grammatical competence, as is the case in most SLA studies, accuracy and automaticity become most relevant.

REFERENCES

Ager, D. E. 1976. ‘Language-learning and sociolinguistics,’IRAL,InternationalReviewofAppliedLinguisticsinLanguageTeaching14/3: 285-97.

Andersen, R. W. and Y. Shirai. 1994. ‘Discourse motivations for some cognitive acquisition principles,’StudiesinSecondLanguageAcquisition16/2: 133-56.

Bailey, N.,C. Madden, and S. D. Krashen. 1974. ‘Is there a "natural sequence" in adult second language learning,’LanguageLearning24/2: 235-43.

Banks, W. P. and H. White. 1985. ‘Semantic congruity and expectancy as separate processes,’Memory&Cognition13/6: 485-93.

Bardovi-Harlig, K. 1992. ‘The relationship of form and meaning: A cross-sectional study of tense and aspect in the interlanguage of learners of English as a second language,’AppliedPsycholinguistics13/3: 253-78.

Bates, E., A. Devescovi, A. Hernandez, and L. Pizzamiglio. 1996. ‘Gender priming in Italian,’Perception&Psychophysics58: 992-1004.

Beck, M.-L. 1997. ‘Regular verbs, past tense and frequency: Tracking down a potential source of NS/NNS competence differences,’SecondLanguageResearch13/2: 93-115.

Bialystok, E. 1978. ‘A theoretical model of second language learning,’LanguageLearning28/1: 69-83.

Bialystok, E. 1979. ‘Explicit and implicit judgements of L2 grammaticality,’LanguageLearning29: 81-103.

Billmyer, K. and M. Varghese. 2000. ‘Investigating instrument-based pragmatic variability: Effects of enhancing discourse completion tests,’AppliedLinguistics21/4: 517-52.

Birdsong, D. and M. Molis. 2001. ‘On the evidence for maturational constraints in second-language acquisition,’JournalofMemoryandLanguage44/2: 235-49.

Bley-Vroman, R. W., S. W. Felix, and G. L. Ioup. 1988. ‘The accessibility of Universal Grammar in adult language learning,’SecondLanguageResearch: 4/1: 1-32.

Byon, A. S. 2004. ‘Sociopragmatic analysis of Korean requests: Pedagogical Settings,’JournalofPragmatics36/9: 1673-704.

Chomsky, N. 1965.AspectsoftheTheoryofSyntax. Cambridge, Mass.: MIT Press.

Cohen, L. and J. Mehler. 1996. ‘Click monitoring revisited: An on-line study of sentence comprehension,’Memory&Cognition24/1: 94-102.

Collins, L. 2002. ‘The roles of L1 influence and lexical aspect in the acquisition of temporal morphology,’LanguageLearning52/1: 43-94.

Day, E. M. and S. M. Shapson. 1991. ‘Integrating formal and functional approaches to language teaching in French immersion: An experimental study,’LanguageLearning41/1: 25-58.

De Graaff, R. 1997. ‘The Experanto Experiment: Effects of explicit instruction on second language acquisition,’StudiesinSecondLanguageAcquisition19/2:249-76.

DeKeyser, R. M. 1996. ‘Exploring automatization processes,’TESOLQuarterly30/2: 349-57.

DeKeyser, R. M. 1997. ‘Beyond explicit rule learning: Automatizing second language morphosyntax,’StudiesinSecondLanguageAcquisition19/2: 195-221.

DeKeyser, R. M. 2000. ‘The robustness of Critical Period effects in second language acquisition,’StudiesinSecondLanguageAcquisition22/4: 499-533.

DeKeyser, R. M. 2003. ‘Implicit and explicit learning’ in C. Doughty and M. H. Long (eds):TheHandbookofSecondLanguageAcquisition. Malden, MA: Blackwell, pp. 313-48.

Duffield, N. and L. White. 1999. ‘Assessing L2 knowledge of Spanish clitic placement: Converging methodologies,’SecondLanguageResearch15/2: 133-60.

Duffield, N., L. White, J. B. De Garavito, S. Montrul, and P. Prévost. 2002. ‘Clitic placement in L2 French: Evidence from sentence matching,’JournalofLinguistics38: 487-525.

Dulay, H. C. and M. K. Burt. 1972. ‘Goofing: An indicator of children’s second language learning strategies,’LanguageLearning22/2: 235-52.

Ellis, N. 1993. ‘Rules and instances in foreign language learning: Interactions of explicit and implicit knowledge,’TheEuropeanJournalofCognitivePsychology5/3: 289-318.

Ellis, R. 2002. ‘Does form-focused instruction affect the acquisition of implicit knowledge?’StudiesinSecondLanguageAcquisition24: 223-36.

Erlam, R. 2003. ‘Evaluating the relative effectiveness of structured-input and output-based instruction in foreign language learning: Results from an experimental study,’StudiesinSecondLanguageAcquisition25/4: 559-82.

Eubank, L. 1993. ‘Sentence matching and processing in L2 development,’SecondLanguageResearch9/3: 253-80.

Favreau, M. and N. S. Segalowitz. 1983. ‘Automatic and controlled processes in the first-and second-language reading of fluent bilinguals,’Memory&Cognition11/6: 565-74.

Jackson, C. N. and P. E. Dussias. 2009. ‘Cross-linguistic differences and their impact on L2 sentence processing,’Bilingualism:LanguageandCognition12: 65-82.

Foote, R. 2011. ‘Integrated knowledge of agreement in early and late English-Spanish bilinguals,’AppliedPsycholinguistics32/1: 187-220.

Freedman, S. E. and K. I. Forster. 1985. ‘The psychological status of overgenerated sentences,’Cognition19/2: 101-31.

Frenck-Mestre, C. 2005. ‘Eye-movement recording as a tool for studying syntactic processing in a second language: A review of methodologies and experimental findings,’SecondLanguageResearch21/2: 175-98.

Gass, S. M. 2001. ‘Sentence matching: A re-examination,’SecondLanguageResearch17/4: 421-41.

Ghrib,E.M. 1987. ‘AlongitudinalstudyofaTunisianadolescent’slearningofEnglishmorphemesinaclassroomcontext,’ IRAL, International Review of Applied Linguistics in Language Teaching 25/1: 41-54.

Golato,P. 2002. ‘WordParsingbylate-learningFrench-Englishbilinguals,’ Applied Psycholinguistics 23/3: 417-46.

Grosjean,F.,J.-Y.Dommergues,E.Cornu,D.Guillelmon,andC.Besson. 1994. ‘Thegender-markingeffectinspokenwordrecognition,’ Perception & Psychophysics 56: 590-98.

Guillelmon,D.andF.Grosjean. 2001. ‘Thegendermarkingeffectinspokenwordrecognition:Thecaseofbilinguals,’ Memory & Cognition 29/3: 503-11.

Hahne,A.andA.D.Friederici. 1999. ‘Electrophysiologicalevidencefortwostepsinsyntacticanalysis:earlyautomaticandlatecontrolledprocesses,’ Journal of Cognitive Neuroscience 11/2: 194-205.

Hahne,A.,J.L.Mueller,andH.Clahsen. 2006. ‘Morphologicalprocessinginasecondlanguage:Behavioralandevent-relatedbrainpotentialevidenceforstorageanddecomposition,’ Journal of Cognitive Neuroscience 18/1: 121-34.

Han,Z. 2004. Fossilization in Adult Second Language Acquisition.ClevedonEngland;Buffalo,NY:MultilingualMatters.

Harley,B. 1989. ‘FunctionalgrammarinFrenchimmersion:Aclassroomexperiment,’ Applied Linguistics 10/3: 331-59.

Herschensohn,J. 2001. ‘MissinginflectioninsecondlanguageFrench:Accidentalinfinitivesandotherverbaldeficits,’ Second Language Research 17/3: 273-305.

Hornberger,N.H. 1989. ‘Tramitesandtransportes:TheacquisitionofsecondlanguagecommunicativecompetenceforonespeecheventinPuno,Peru,’ Applied Linguistics 10/2: 214-30.

Hu,G. 2002. ‘Psychologicalconstraintsontheutilityofmetalinguisticknowledgeinsecondlanguageproduction,’ Studies in Second Language Acquisition 24/3: 347-86.

Hulstijn,J.H. 2001. ‘Intentionalandincidentalsecond-languagevocabularylearning:Areappraisalofelaboration,rehearsalandautomaticity’inP.Robinson(ed.): Cognition and Second Language Instruction.Cambridge:CambridgeUniversityPress,pp. 258-86.

Jakubowicz,C.andC.Faussart. 1998. ‘GenderagreementintheprocessingofspokenFrench,’ Journal of Psycholinguistic Research 27: 597-617.

Jiang,N. 2004. ‘Morphologicalinsensitivityinsecondlanguageprocessing,’ Applied Psycholinguistics 25/4: 603-34.

Jiang,N. 2007. ‘Selectiveintegrationoflinguisticknowledgeinadultsecondlanguageacquisition,’ Language Learning 57/1: 1-33.

Jiang,N. 2011. Conducting Reaction Time Research in Second Language Studies.London:Routledge.

Jiang,N.,E.Novokshanova,K.Masuda,andXin.Wang. 2011. ‘MorphologicalcongruencyandtheacquisitionofL2morphemes,’ Language Learning 61: 940-67.

Jiang,N.,G.Hu,A.Lukyanchenko,andY.Cao. 2010. ‘Insensitivity to Morphological Errors in L2: Evidence from Word Monitoring,’PresentedatSecondLanguageResearchForum,CollegePark,MD,USA,October14-17, 2010.

Johnson,J.S.andE.L.Newport. 1989. ‘Criticalperiodeffectsinsecondlanguagelearning:TheinfluencesofmaturationalstateontheacquisitionofEnglishasasecondlanguage,’ Cognitive Psychology 21: 60-99.

Johnson,J.S. 1992. ‘Criticalperiodeffectsinsecondlanguageacquisition:Theeffectofwrittenversusauditorymaterialsontheassessmentofgrammaticalcompetence,’ Language Learning 42/2: 217-48.

Johnson,J.S.,andE.L.Newport. 1991. ‘Criticalperiodeffectsonuniversalpropertiesoflanguage:Thestatusofsubjacencyintheacquisitionofasecondlanguage,’ Cognition 39/3: 215-18.

Johnson,K.E. 1992. ‘Cognitivestrategiesandsecondlanguagewriters:Are-evaluationofeentencecombining,’ Journal of Second Language Writing 1/1: 61-75.

Kasper,G. 1979. ‘Errorsinspeechactrealizationanduseofgambits,’ The Canadian Modern Language Review/La Revue Canadienne des Langues Vivantes 35/3: 395-406.

Kasper,G. 2001. ‘FourperspectivesonL2pragmaticdevelopment,’ Applied Linguistics 22/4: 502-30.

Kennedy,A.,A.Wilkes,L.Elder,andW.S.Murray. 1988. ‘Dialoguewithmachines,’ Cognition 30/1: 37-72.

Kennison,S.M. 2002. ‘Comprehendingnounphraseargumentsandadjuncts,’ Journal of Psycholinguistic Research 31/1: 65-81.

Krashen,S.D. 1981. Second Language Acquisition and Second Language Learning.Oxford:PergamonPress.

Krashen,S.D. 1982. Principles and Practice in Second Language Acquisition.Oxford:PergamonPress.

Krashen,S.,N.Houck,P.Giunchi,S.Bode,R.Birnbaum,andG.Strei. 1977. ‘Difficultyorderforgrammaticalmorphemesforadultsecondlanguageperformersusingfreespeech,’ TESOL Quarterly 11/3: 338-41.

Krashen,S.D.andP.Pon. 1975. ‘AnerroranalysisofanadvancedlearnerofESL:Theimportanceofthemonitor,’ Working Papers on Bilingualism 7: 125-29.

Lardiere,D. 1998. ‘Caseandtenseinthe"fossilized"steadystate,’ Second Language Research 14/1: 1-26.

Lelouche,R.andD.Huot. 1998. ‘Influenceofcommunicativesituationvariablesonlinguisticform,’ Computer Assisted Language Learning 11/5: 523-41.

Lutjen,H.P. 1973. ‘Oncommunicativecompetenceintheforeignlanguage,’ IRAL, International Review of Applied Linguistics in Language Teaching 11/1: 81-91.

McLaughlin,B. 1990. ‘Restructuring,’ Applied Linguistics 11/2: 113-28.

Mehnert,U. 1998. ‘Theeffectsofdifferentlengthsoftimeforplanningonsecondlanguageperformance,’ Studies in Second Language Acquisition 20/1: 83-108.

Mueller,J.L.,A.Hahne,Y.Fujii,andA.D.Friederici. 2005. ‘Nativeandnonnativespeakers’processingofaminiatureversionofJapaneseasrevealedbyERPs,’ Journal of Cognitive Neuroscience 17/8: 1229-244.

Muranoi,H. 2000. ‘Focusonformthroughinteractionenhancement:IntegratingformalinstructionintoacommunicativetaskinEFLclassrooms,’ Language Learning 50/4: 617-73.

Nelson,G.L.,J.Carson,M.AlBatal,andW.ElBakary. 2002. ‘Cross-culturalpragmatics:StrategyuseinEgyptianArabicandAmericanEnglishrefusals,’ Applied Linguistics 23/2: 163-89.

Newport,E.L. 1990. ‘Maturationalconstraintsonlanguagelearning,’ Cognitive Science 14: 11-28.

Ni, W., J. D. Fodor, S. Crain, and D. Shankweiler. 1998. ‘Anomaly detection: Eye movement patterns,’JournalofPsycholinguisticResearch27/5: 515-39.

Norris, J. M. and L. Ortega. 2000. ‘Effectiveness of L2 instruction: A research synthesis and quantitative meta-analysis,’LanguageLearning50/3: 417-528.

Ojima, S., H. Nakata, and R. Kakigi. 2005. ‘An ERP study of second language learning after childhood: Effects of proficiency,’JournalofCognitiveNeuroscience17/8: 1212-228.

Ortega, L. 1999. ‘Planning and focus on form in L2 oral performance,’StudiesinSecondLanguageAcquisition21/1: 109-48.

Paulston, C. B. 1974. ‘Linguistic and communicative competence,’TESOLQuarterly8/4: 347-62.

Pearlmutter, N. J., S. M. Garnsey, and K. Bock. 1999. ‘Agreement processes in sentence comprehension,’JournalofMemoryandLanguage41: 427-56.

Peelle, J. E., A. Cooke, P. Moore, L. Vesely, and M. Grossman. 2007. ‘Syntactic and thematic components of sentence processing in progressive nonfluent aphasia and nonaphasic frontotemporal dementia,’JournalofNeurolinguistics20: 482-94.

Pica, T. 1983. ‘Adult acquisition of English as a second language under different conditions of exposure,’LanguageLearning33/4: 465-97.

Pica, T. 1988. ‘Interlanguage adjustments as an outcome of NS-NNS negotiated interaction,’LanguageLearning38/1: 45-73.

Prevost, P. and L. White. 2000. ‘Missing surface inflection or impairment in second language acquisition? Evidence from tense and agreement,’SecondLanguageResearch16/2: 103-33.

Raimes, A. 1987. ‘Language proficiency, writing ability, and composing strategies: A study of ESL college student writers,’LanguageLearning37/3: 439-68.

Richards, D. R. 1975. ‘In search of relevant foundations—towards a more integrated ‘Applied Linguistics?’,’ITL,ReviewofAppliedLinguistics29: 1-18.

Rivers, W. M. 1973. ‘From linguistic competence to communicative competence,’TESOLQuarterly7/1: 25-34.

Robinson, P. 1997. ‘Generalizability and automaticity of second language learning under implicit, incidental, enhanced, and instructed conditions,’StudiesinSecondLanguageAcquisition19/2: 223-47.

Robinson, D. H. and A. D. Katayama. 1997. ‘At-lexical, articulatory interference in silent reading: The "upstream" tongue-twister effect,’Memory&Cognition25/5: 661-65.

Rosa, E. M. and R. P. Leow. 2004. ‘Awareness, different learning conditions, and second language development,’AppliedPsycholinguistics25/2: 269-92.

Safont Jorda, M. P. 2003. ‘Metapragmatic awareness and pragmatic production of third language learners of English: A focus on request acts realizations,’InternationalJournalofBilingualism7/1: 43-69.

Salaberry, M. R. 2000. ‘The acquisition of English past tense in an instructional setting,’System28/1: 135-52.

Schachter, J. 1989. ‘A new look at an old classic,’SecondLanguageResearch5/1: 30-42.

Schachter, J. 1990. ‘On the issue of completeness in second language acquisition,’SecondLanguageResearch6/2: 93-124.

Schmidt, M. 1980. ‘Coordinate structures and language universals in interlanguage,’LanguageLearning30/2: 397-416.

Schmidt, R. 1992. ‘Psychological mechanisms underlying second language fluency,’StudiesinSecondLanguageAcquisition14/4: 357-85.

Scott, V. M. 1989. ‘An empirical study of explicit and implicit teaching strategies in French,’TheModernLanguageJournal73/1: 14-22.

Scott, V. M. 1990. ‘Explicit and implicit grammar teaching strategies: New empirical data,’TheFrenchReview63: 779-89.

Segalowitz, N. 2003. ‘Automaticity and second languages’ in C. Doughty and M. H. Long (eds):TheHandbookofSecondLanguageAcquisition. Malden, MA: Blackwell, pp. 382-408.

Segalowitz, N. S. and S. J. Segalowitz. 1993. ‘Skilled performance, practice, and the differentiation of dpeed-up from automatization effects: Evidence from second language word recognition,’AppliedPsycholinguistics14/3: 369-85.

Segalowitz, S. J., N. S. Segalowitz, and A. G. Wood. 1998. ‘Assessing the development of automaticity in second language word recognition,’AppliedPsycholinguistics19/1: 53-67.

Sharwood Smith, M. 1986. ‘The competence/control model, crosslinguistic influence and the creation of new grammars’ in E. Kellerman and M. Sharwood Smith (eds):Cross-LinguisticInfluenceinSecondLanguageAcquisition(pp. 10-20). New York: Pergamon Press.

Skehan, P. and P. Foster. 1999. ‘The influence of task structure and processing conditions on narrative retellings,’LanguageLearning49/1: 93-120.

Takahashi, S. 1996. ‘Pragmatic transferability,’StudiesinSecondLanguageAcquisition18/2: 189-223.

Tokowicza, N. and T. Warrena. 2010. ‘Beginning adult L2 learners’ sensitivity to morphosyntactic violations: A self-paced reading study,’EuropeanJournalofCognitivePsychology22: 1092-106.

Thornton, R., M. C. MacDonald, and J. E. Arnold. 2000. ‘The concomitant effects of phrase length and informational content in sentence comprehension,’JournalofPsycholinguisticResearch29/2: 195-203.

Walters, J. 1980. ‘Grammar, meaning, and sociocultural appropriateness in second language acquisition,’CanadianJournalofPsychology/RevueCanadiennedePsychologie34/4: 337-45.

Wang, W. and Q. Wen. 2002. ‘L1 use in the L2 composing process: An exploratory study of 16 Chinese EFL writers,’JournalofSecondLanguageWriting11/3: 225-46.

Wei, L. 2000. ‘Unequal election of morphemes in adult second language acquisition,’AppliedLinguistics21/1: 106-40.

Weyerts, H., M. Penke, T. F. Munte, H.-J. Heinze, and H. Clahsen. 2002. ‘Word order in sentence processing: An experimental study of verb placement in German,’JournalofPsycholinguisticResearch31/3: 211-68.

White, L. 1985. ‘The acquisition of parameterized grammars: Subjacency in second language acquisition,’SecondLanguageResearch1/1: 1-17.

White, L., E. Valenzuela, M. Kozlowska-MacGregor, and Y.-K. I. Leung. 2004. ‘Gender and number agreement in nonnative Spanish,’AppliedPsycholinguistics25/1: 105-33.

White, R. V. 1974. ‘Communicative competence, registers, and second language teaching,’IRAL12/2: 127-41.

Winograd, E., C. Cohen, and J. Barresi. 1976. ‘Memory for concrete and abstract words in bilingual speakers,’Memory&Cognition4/3: 323-29.

Young, R. 1988. ‘Variation and the interlanguage hypothesis,’StudiesinSecondLanguageAcquisition10/3: 281-302.

Yuan, F. and R. Ellis. 2003. ‘The effects of pre-task planning and on-line planning on fluency, complexity and accuracy in L2 monologic oral production,’AppliedLinguistics24/1: 1-27.