This study extends the evidence for implicit second language (L2) learning, which comes largely from (semi-)artificial language research, to German. Upper-intermediate L2 German learners were flooded with spoken exemplars of a difficult morphological structure, namely strong, vowel-changing verbs. Toward the end of exposure, the mandatory vowel change was omitted, yielding ungrammatical verb forms (compare Leung & Williams, 2012). Two pre- and posttests—word monitoring and controlled oral production—gauged the development of learners’ implicit and explicit knowledge, respectively.

Interviews revealed 33 out of 38 L2 learners remained unaware of the ungrammatical verbs in the input flood; however, they showed significant sensitivity during listening as evidenced by a reaction time slowdown on ungrammatical trials. The unaware learners also improved significantly from pretest to posttest on the word-monitoring task, but not the oral production measure, unless the verbs’ salience in the input flood had resonated with them. Thus, implicit instruction affected implicit knowledge primarily, although prior knowledge and memory could potentially account for interactions between implicit processing, implicit knowledge, and explicit knowledge.

This article investigates the learning of two syntactic structures in Italian—Pre-possessive Determiner Article (PPDA) and Differential Object Marking—that each pose specific learning difficulties for Spanish-speaking learners of Italian (SSLI), resulting from the asymmetrical relationship with corresponding structures in the learners’ L1 (Spanish). The article also examines to what extent these difficulties can be alleviated by means of a relatively implicit type of form-focused instruction (Textual Enhancement, or TE). SSLI tend to omit the article before the possessive determiner (e.g., *mia macchina) and overuse the Prepositional Accusative (PA, the Romance languages’ instance of Differential Object Marking) in canonical SVO sentences (e.g., *conosco a Marco). Although both structures pose learning difficulties to SSLI, the learning path that leads SSLI to acquire these two features is different: SSLI have to learn to add PPDA to their interlanguage and they have to unlearn PA, a property of their L1. Sixty-eight instructed SSLI were divided in two groups. Group A read five texts where the absence of PA was textually enhanced; group B read the same texts where the presence of PPDA was enhanced. The participants took a timed grammatical judgment test before, shortly after, and two months after the instructional treatment. The results show that the effects of the TE are different for PA and PPDA, which indicates that features to be unlearned are more difficult and require more explicit instructional intervention.

This study investigates the relationship between foreign language aptitude and the learning of two English structures defined as easy or difficult to learn. Using a quasiexperimental design, 66 secondary-level learners of English as a foreign language from three intact classes were provided with four hours of instruction on the passive (a difficult structure) and the past progressive (an easy structure). Language aptitude was measured using the LLAMA Aptitude Test (Meara, 2005). Language outcomes were measured with a written grammaticality judgment and an oral production task. The results revealed that one of the aptitude components, grammatical inferencing, contributed to learners’ gains on the passive but not the past progressive on the written measure. Another component of aptitude, associative memory, contributed to learners’ gains on the past progressive on the oral measure. The results provide support for the claim that different components of aptitude contribute to the learning of difficult and easy L2 structures in different ways. There is also support for the proposal that different components of aptitude may be involved at different stages of language acquisition (Skehan, 2002).

Meta-analytic research suggests an edge of explicit over implicit instruction for the development of complex L2 grammatical structures, but the jury is still out as to which type of explicit instruction—deductive or inductive, where rules are respectively provided or elicited—proves more effective. Avoiding this dichotomy, accumulating research shows superior results for guided induction, in which teachers help learners co-construct rules by directing their attention to relevant aspects in the input and asking guiding questions. However, no study has jointly investigated the effects of guided induction on both learning outcomes and processes, or whether guided induction can prove effective outside classroom settings where teacher mediation is not possible. In this study, which targeted the complex Spanish gustar structures, 70 English-speaking learners of beginning Spanish received either guided induction via a videogame, deductive instruction in a traditional classroom setting, or no instruction. Learning outcomes were measured via one receptive and two controlled production tasks (oral and written) with old and new items. Results revealed that while both instruction groups improved across time, outperforming the control group, the guided induction group achieved higher learning outcomes on all productive posttests (except immediate oral production) and experienced greater retention. Additionally, the think-aloud protocols of the guided induction group revealed high levels of awareness of the L2 structure and a conspicuous activation of recently learned knowledge, which are posited to have contributed to this group’s superior performance. These findings thus illustrate, quantitatively and qualitatively, the potential of guided induction for the development of complex L2 grammar in online learning environments.

Second language learning outcomes are highly variable, due to a variety of factors, including individual differences, exposure conditions, and linguistic complexity. However, exactly how these factors interact to influence language learning is unknown. This article examines the relationship between these three variables in language learners.

Native English speakers were exposed to an artificial language containing three sentence patterns of varying linguistic complexity. They were randomly assigned to two groups—incidental and instructed—designed to promote the acquisition of implicit and explicit knowledge, respectively. Learning was assessed with a grammaticality judgment task, and subjective measures of awareness were used to measure whether exposure had resulted in implicit or explicit knowledge. Participants also completed cognitive tests.

Awareness measures demonstrated that learners in the incidental group relied more on implicit knowledge, whereas learners in the instructed group relied more on explicit knowledge. Overall, exposure condition was the most significant predictor of performance on the grammaticality judgment task, with learners in the instructed group outperforming those in the incidental group. Performance on a procedural learning task accounted for additional variance. When outcomes were analyzed according to linguistic complexity, exposure condition was the most significant predictor for two syntactic patterns, but it was not a predictor for the most complex sentence group; instead, procedural learning ability was.

This article draws on an approach that conceptualizes L2 learning difficulty in terms of implicit and explicit knowledge. In a study with first language Mexican Spanish university-level learners (n = 30), their teachers (n = 11), and applied linguistics experts (n = 3), we investigated the relationship between (a) these groups’ difficulty judgements of 13 selected L2 English structures and (b) perceived learning difficulty and learners’ actual performance on measures of implicit and explicit knowledge. Our findings show that experts’ learning difficulty judgements did not lead to significant predictions, whereas the learners’ own difficulty rankings correlated significantly with their performance on the measure of explicit knowledge. Although correlations based on teachers’ difficulty rankings did not reach statistical significance, the judgements of this group were the only ones that showed trends toward successful prediction of learners’ performance on both the implicit and the explicit L2 measures. Thus the teachers exhibited a trend toward the best overall prediction ability.

In this short article I try to present a taxonomy of sources of difficulty, in particular complexity. I emphasize how these factors interact with each other and with individual differences and treatments, before illustrating some of these interactions and the methodological issues involved in studying them with examples from the empirical studies in this issue. Finally, I make some suggestions for future research.