There are good theoretical and educational reasons to place matters of implicit and explicit learning high on the agenda for SLA research. As for theoretical motivations, perhaps the most central issue in SLA theory construction in need of explanation is the differential success in one's first language (L1) and in one's second language (L2). Although acquisition of an L1 results in full mastery of the language (provided that children are exposed to sufficient quantities of input and do not suffer from mental disabilities), learners of an L2—even after many years of L2 exposure—differ widely in level of attainment. How can we explain universal success in the case of L1 acquisition and differential success in the case of L2 acquisition? Among the many explanations that have been proposed, including brain maturation and brain adaptation processes (critical period), access to Universal Grammar, L1 interference, and sociopsychological factors (see Hyltenstam & Abrahamsson, 2003, for a review), one finds explanations that involve the notions of implicit and explicit learning. Scholars working in different disciplines, in different theoretical schools, and sometimes using different terminology have argued that L1 acquisition (or at least the acquisition of L1 grammar) relies principally on processes of what we might now call implicit learning, whereas the acquisition of an L2 often relies on both implicit and explicit learning (Bley-Vroman, 1991; DeKeyser, 2003; N. Ellis, this issue; R. Ellis, 2004; Krashen, 1981; Reber & Allen, 2000).I am grateful to Rod Ellis for his thoughtful comments on previous versions of this text.

A problem facing investigations of implicit and explicit learning is the lack of valid measures of second language implicit and explicit knowledge. This paper attempts to establish operational definitions of these two constructs and reports a psychometric study of a battery of tests designed to provide relatively independent measures of them. These tests were (a) an oral imitation test involving grammatical and ungrammatical sentences, (b) an oral narration test, (c) a timed grammaticality judgment test (GJT), (d) an untimed GJT with the same content, and (e) a metalinguistic knowledge test. Tests (a), (b), and (c) were designed as measures of implicit knowledge, and tests (d) and (e) were designed as measures of explicit knowledge. All of the tests examined 17 English grammatical structures. A principal component factor analysis produced two clear factors. This analysis showed that the scores from tests (a), (b), and (c) loaded on Factor 1, whereas the scores from ungrammatical sentences in test (d) and total scores from test (e) loaded on Factor 2. These two factors are interpreted as corresponding to implicit and explicit knowledge, respectively. A number of secondary analyses to support this interpretation of the construct validity of the tests are also reported.This research was funded by a Marsden Fund grant awarded by the Royal Society of Arts of New Zealand to Rod Ellis and Cathie Elder. Other researchers who contributed to the research are Shawn Loewen, Rosemary Erlam, Satomi Mizutani, and Shuhei Hidaka.The author wishes to thank Nick Ellis, Jim Lantolf, and two anonymous SSLA reviewers. Their constructive comments have helped me to present the theoretical background of the study more convincingly and to remove errors from the results and refine my interpretations of them.

We used event-related brain potentials (ERPs) to investigate the contributions of explicit and implicit processes during second language (L2) sentence comprehension. We used a L2 grammaticality judgment task (GJT) to test 20 native English speakers enrolled in the first four semesters of Spanish while recording both accuracy and ERP data. Because end-of-sentence grammaticality judgments are open to conscious inspection, we reasoned that they can be influenced by strategic processes that reflect on formal rules and therefore reflect primarily offline explicit processing. On the other hand, because ERPs are a direct reflection of online processing, they reflect automatic, nonreflective, implicit responses to stimuli (Osterhout, Bersick, & McLaughlin, 1997; Rugg et al., 1998; Tachibana et al., 1999).

We used a version of the GJT adapted for the ERP environment. The experimental sentences varied the form of three different syntactic constructions: (a) tense-marking, which is formed similarly in the first language (L1) and the L2; (b) determiner number agreement, which is formed differently in the L1 and the L2; and (c) determiner gender agreement, which is unique to the L2. We examined ERP responses during a time period between 500 and 900 ms following the onset of the critical (violation or matched control) word in the sentence because extensive past research has shown that grammatical violations elicit a positive-going deflection in the ERP waveform during this period (e.g., the “P600”; Osterhout & Holcomb, 1992).

We found that learners were sensitive (i.e., showed different brain responses to grammatical and ungrammatical sentences) to violations in L2 for constructions that are formed similarly in the L1 and the L2, but were not sensitive to violations for constructions that differ in the L1 and the L2. Critically, a robust grammaticality effect was found in the ERP data for the construction that was unique to the L2, suggesting that the learners were implicitly sensitive to these violations. Judgment accuracy was near chance for all constructions. These findings suggest that learners are able to implicitly process some aspects of L2 syntax even in early stages of learning but that this knowledge depends on the similarity between the L1 and the L2. Furthermore, there is a divergence between explicit and implicit measures of L2 learning, which might be due to the behavioral task demands (e.g., McLaughlin, Osterhout, & Kim, 2004). We conclude that comparing ERP and behavioral data could provide a sensitive method for measuring implicit processing.This research was supported by a National Institutes of Health Individual National Research Service Award (NIH HD42948-01) awarded to Natasha Tokowicz and a National Institutes of Health Institutional National Research Service Award (T32 MH19102) awarded to Brian MacWhinney. We thank Beatrice DeAngelis, Dayne Grove, Kwan Hansongkitpong, Katie Keil, Lee Osterhout, Chuck Perfetti, Kelley Sacco, Alex Waid, and Eddie Wlotko for their assistance with this project. We gratefully acknowledge the comments of Rod Ellis, Jan Hulstijn, Albert Valdman, and the two anonymous SSLA reviewers on earlier versions of this manuscript. A portion of these results was presented at the 43rd Annual Meeting of the Psychonomic Society (2002, November).

This study examines whether aural processing of input in a situation of implicit instruction can build a knowledge base that is available for both comprehension and production tasks. Fifty-five Dutch students learned a miniature linguistic system based on Spanish. Three training conditions were compared in which noun-adjective gender agreement was the learning target. The first group of participants received receptive training, the second group received receptive and productive training, and a third group served as a control. The control group received no training of the target structure and only read an explanation of the target structure rule. Receptive knowledge was assessed with a self-paced listening test, a match-mismatch test, and a grammaticality judgment task. Productive knowledge was tested with a picture description task in single- and dual-task conditions. A postexperimental questionnaire tested whether any explicit knowledge had been induced. Results suggest that the receptive and receptive + productive training programs succeeded in building a knowledge base that was used in comprehension but much less so in production. These results will be interpreted in light of processing and the distinction between implicit and explicit knowledge.I would like to express my gratitude to Jan Hulstijn and Rob Schoonen from the University of Amsterdam for their supervision. I would also like to thank the two anonymous SSLA reviewers for their comments and Nick Ellis for taking time to discuss this project with me.

This paper reports replications of studies of implicit artificial grammar (AG) learning and explicit series-solution learning with experienced second language learners in order to examine their population and content generalizability. As found by Reber, Walkenfeld, and Hernstadt (1991), there was significantly greater variance in explicit compared to implicit learning. In contrast to Reber et al.'s findings, intelligence quotient (IQ) was significantly negatively related to implicit learning. As found by Knowlton and Squire (1996), chunks that appeared with high frequency (high chunk-strength) in AG training influenced incorrect acceptance of ungrammatical transfer test items containing them but did not affect the judgments of grammatical items. In a third experiment, learners semantically processed sentences in Samoan, a novel language for this population. This experiment found little evidence for the content generalizability of these AG findings to the incidental learning of Samoan. Implicit AG and incidental Samoan learning had different patterns of correlation with cognitive abilities (IQ, working memory, and aptitude) and differed in sensitivity to chunk-strength. As found for AG learning, high chunk-strength negatively affected correct rejection of ungrammatical Samoan transfer test items. Additionally, high chunk-strength negatively affected correct acceptance of grammatical items. For these grammatical items, the number of chunks they contained—not their frequency during training—positively influenced grammaticality judgments.I gratefully acknowledge the helpful comments and advice on this paper given by the editors of this special issue, Jan Hulstijn and Rod Ellis, and also by Nick Ellis, Barbara Knowlton, and two anonymous SSLA reviewers.

Two experiments examined the learning of form-meaning connections under conditions where the relevant forms were noticed but the critical aspects of meaning were not. Miniature noun class systems were employed, and the participants were told that the choice of determiner in noun phrases depended on whether the object was “near” or “far” from the subject of the sentence. What they were not told was that the choice of determiner also depended on the animacy of the noun. Most participants remained unaware of this correlation during the training and test tasks; yet when faced with a choice between two determiners for a noun, they chose the one that was appropriate to the noun's animacy at significantly above-chance levels, even though that combination had never been encountered during training. This ability to generalize provided evidence of learning form-meaning connections without awareness. In both experiments, there was a correlation between generalization test performance and knowledge of languages that encode grammatical gender. This points to the importance of prior knowledge in implicit learning.Many thanks to Ronald Leow, Daphnée Simard, and the anonymous SSLA reviewers for their helpful comments on an earlier draft of this article.

This paper considers how implicit and explicit knowledge are dissociable but cooperative. It reviews various psychological and neurobiological processes by which explicit knowledge of form-meaning associations impacts upon implicit language learning. The interface is dynamic: It happens transiently during conscious processing, but the influence upon implicit cognition endures thereafter. The primary conscious involvement in SLA is the explicit learning involved in the initial registration of pattern recognizers for constructions that are then tuned and integrated into the system by implicit learning during subsequent input processing. Neural systems in the prefrontal cortex involved in working memory provide attentional selection, perceptual integration, and the unification of consciousness. Neural systems in the hippocampus then bind these disparate cortical representations into unitary episodic representations. These are the mechanisms by which Schmidt's (1990) noticing helps solve Quine's (1960) problem of referential indeterminacy. Explicit memories can also guide the conscious building of novel linguistic utterances through processes of analogy. Formulas, slot-and-frame patterns, drills, and declarative pedagogical grammar rules all contribute to the conscious creation of utterances whose subsequent usage promotes implicit learning and proceduralization. Flawed output can prompt focused feedback by way of recasts that present learners with psycholinguistic data ready for explicit analysis. Other processes of acquisition from output include differentiation, analysis, and preemption. These processes of conscious construction in working memory underpin relationships between individual differences in working memory capacities and language learning aptitude.Thanks to Rod Ellis for first suggesting that I try to write this and to the staff and students at Department of Applied Language Studies and Linguistics University of Auckland (2003), the TESOL Program Temple University Japan (2003), the Chester Language Development Reading Group, and the LOT winter school (2004) for helping me think it through. I am particularly grateful to Michel Paradis, Michael Swan, Karen Roehr, Anne Feryok, and Tamar Keren-Portnoy for pointing their giant biological cameras of consciousness at a prior draft, and for sharing their awareness with me in kindly and constructive fashion. I have learned a lot from it.

Publications received.