Second language (L2) researchers are becoming more interested in both L2 learners’ knowledge of the target language and how that knowledge is put to use during real-time language processing. Researchers are therefore beginning to see the importance of combining traditional L2 research methods with those that capture the moment-by-moment interpretation of the target language, such as eye-tracking. The major benefit of the eye-tracking method is that it can tap into real-time (or online) comprehension processes during the uninterrupted processing of the input, and thus, the data can be compared to those elicited by other, more metalinguistic tasks to offer a broader picture of language acquisition and processing. In this article, we present an overview of the eye-tracking technique and illustrate the method with L2 studies that show how eye-tracking data can be used to (a) investigate language-related topics and (b) inform key debates in the fields of L2 acquisition and L2 processing.

Many studies on bilingual language processing have shown that lexical access is not selective with respect to language. These studies typically used nouns as word stimuli. The aim of the present study was to extend the previous findings on noun processing to verb processing. In the first experiment, Dutch-English bilinguals performed a lexical decision task in their second language and were faster to recognize cognate verbs (e.g., Dutch-English geven-give) presented out of context than control words. This verb cognate facilitation effect was not modulated by verb tense. In a second experiment, cognates and controls were presented in sentence contexts while eye movements were recorded. In contrast to the strong cognate facilitation effects on early and later reading time measures for nouns found in earlier studies, cognate facilitation was only observed on a later reading time measure (i.e., go-past time). An interpretation of the results within current models of bilingual language processing and lexical organization is provided.

Adult learners have persistent difficulty processing second language (L2) inflectional morphology. We investigate associative learning explanations that involve the blocking of later experienced cues by earlier learned ones in the first language (L1; i.e., transfer) and the L2 (i.e., proficiency). Sagarra (2008) and Ellis and Sagarra (2010b) found that, unlike Spanish monolinguals, intermediate English-Spanish learners rely more on salient adverbs than on less salient verb inflections, but it is not clear whether this preference is a result of a default or a L1-based strategy. To address this question, 120 English (poor morphology) and Romanian (rich morphology) learners of Spanish (rich morphology) and 98 English, Romanian, and Spanish monolinguals read sentences in L2 Spanish (or their L1 in the case of the monolinguals) containing adverb-verb and verb-adverb congruencies or incongruencies and chose one of four pictures after each sentence (i.e., two that competed for meaning and two for form). Eye-tracking data revealed significant effects for (a) sensitivity (all participants were sensitive to tense incongruencies), (b) cue location in the sentence (participants spent more time at their preferred cue, regardless of its position), (c) L1 experience (morphologically rich L1 learners and monolinguals looked longer at verbs than morphologically poor L1 learners and monolinguals), and (d) L2 experience (low-proficiency learners read more slowly and regressed longer than high-proficiency learners). We conclude that intermediate and advanced learners are sensitive to tense incongruencies and—like native speakers—tend to rely more heavily on verbs if their L1 is morphologically rich. These findings reinforce theories that support transfer effects such as the unified competition model and the associative learning model but do not contradict Clahsen and Felser’s (2006a) shallow structure hypothesis because the target structure was morphological agreement rather than syntactic agreement.

This study focuses on beginning second language learners’ attention to irregular verb morphology, an area of grammar that many adults find difficult to acquire (e.g., DeKeyser, 2005; Larsen-Freeman, 2010). We measured beginning learners’ eye movements during sentence processing to investigate whether or not they actually attend to irregular verb features and, if so, whether the amount of attention that they pay to these features predicts their acquisition. On the assumption that attention facilitates learning (e.g., Gass, 1997; Robinson, 2003; Schmidt, 2001), we expected more attention (i.e., longer fixations or more frequent comparisons between verb forms) to lead to more learning of the irregular verbs. Forty beginning learners of German read 12 German sentence pairs with stem-changing verbs and 12 German sentence pairs with regular verbs while an Eyelink 1000 recorded their eye movements. The stem-changing verbs consisted of six a → ä changing verbs and six e → i(e) changing verbs. Each verb appeared in a baseline sentence in the first-person singular, which has no stem change, and a critical sentence in the second- or third-person singular, which have a stem change for the irregular but not the regular verbs, on the same screen. Productive pre- and posttests measured the effects of exposure on learning. Results indicate that learners looked longer overall at stem-changing verbs than regular verbs, revealing a late effect of verb irregularity on reading times. Longer total times had a modest, favorable effect on the subsequent production of the stem vowel. Finally, the production of only the a → ä verbs—not the e → i(e) verbs—benefited from direct visual comparisons during reading, possibly because of the umlaut in the former. We interpret the results with reference to recent theory and research on attention, noticing, and language learning and provide a more nuanced and empirically based understanding of the noticing construct.

In his 2007 study “Effects of Textual Enhancement and Topic Familiarity on Korean EFL Students’ Reading Comprehension and Learning of Passive Form,” Lee demonstrated that learners were better able to correct written sentences that contained incorrect English passive forms after exposure to texts flooded with enhanced (versus nonenhanced) passive forms. But with enhanced forms, learners did worse on comprehension tests, which arguably demonstrated a trade-off: More attention to forms resulted in less to meaning. In this study, a conceptual replication of Lee’s using eye-movement data, I assessed how English passive construction enhancement affects English language learners’ (a) learning of the form (via pre- and posttest gains on passive construction tests) and (b) text comprehension. In contrast to Lee’s results, I found enhancement did not significantly increase form correction gain scores, nor did enhancement significantly detract from comprehension. There was no trade-off effect. Form learning and comprehension did not correlate. By recording learners’ eye movements while reading, I found enhancement significantly impacted learners’ noticing of the passive forms through longer gaze durations and rereading times. Thus, enhancement in this study functioned as intuitively and originally (Sharwood Smith, 1991, 1993) proposed; it promoted noticing, but, in this case, without further explicit instruction, it appeared to have done little else.

In a recent study, Lew-Williams and Fernald (2007) showed that native Spanish speakers use grammatical gender information encoded in Spanish articles to facilitate the processing of upcoming nouns. In this article, we report the results of a study investigating whether grammatical gender facilitates noun recognition during second language (L2) processing. Sixteen monolingual Spanish participants (control group) and 18 English-speaking learners of Spanish (evenly divided into high and low Spanish proficiency) saw two-picture visual scenes in which items matched or did not match in gender. Participants’ eye movements were recorded while they listened to 28 sentences in which masculine and feminine target items were preceded by an article that agreed in gender with the two pictures or agreed only with one of the pictures. An additional group of 15 Italian learners of Spanish was tested to examine whether the presence of gender in the first language (L1) modulates the degree to which gender is used during L2 processing. Data were analyzed by comparing the proportion of eye fixations on the objects in each condition. Monolingual Spanish speakers looked sooner at the referent on different-gender trials than on same-gender trials, replicating results reported in past literature. Italian-Spanish bilinguals exhibited a gender anticipatory effect, but only for the feminine condition. For the masculine condition, participants waited to hear the noun before identifying the referent. Like the Spanish monolinguals, the highly proficient English-Spanish speakers showed evidence of using gender information during online processing, whereas the less proficient learners did not. The results suggest that both proficiency in the L2 and similarities between the L1 and the L2 modulate the usefulness of morphosyntactic information during speech processing.

Eye-trackers are becoming increasingly widespread as a tool to investigate second language (L2) acquisition. Unfortunately, clear standards for methodology—including font size, font type, and placement of interest areas—are not yet available. Although many researchers stress the need for ecological validity—that is, the simulation of natural reading conditions—it may not be prudent to use such a design to investigate new directions in eye-tracking research, and particularly in research involving small lexical items such as articles. In this study, we examine whether two different screen layouts can lead to different results in an eye-tracking study on the L2 acquisition of Italian gender. The results of an experiment with an ecologically valid design are strikingly different than the results of an experiment with a design tailored to track eye movements to articles. We conclude that differences in screen layout can have significant effects on results and that it is crucial that researchers report screen layout information.