Skip to main content Accessibility help
×
Hostname: page-component-77f85d65b8-lfk5g Total loading time: 0 Render date: 2026-04-20T04:33:42.078Z Has data issue: false hasContentIssue false

5 - Absence of morphological transfer in beginners: evidence from eye tracking

Published online by Cambridge University Press:  05 October 2014

By
Nuria Sagarra
Edited by
ZhaoHong Han  and
Rebekah Rast
ZhaoHong Han
Affiliation:
Columbia University, New York
Rebekah Rast
Affiliation:
The American University of Paris, France

Information

Type
Chapter
Information
First Exposure to a Second Language
Learners' Initial Input Processing
 , pp. 139 - 170
Publisher: Cambridge University Press
Print publication year: 2014

5 Absence of morphological transfer in beginners: evidence from eye tracking

Nuria Sagarra

During the past decade, literature on cognitive approaches to second language acquisition (SLA) and bilingualism has flourished spectacularly (see Grosjean & Lee Reference Grosjean and Lee2013; Herschensohn & Young-Scholten Reference Herschensohn and Young-Scholten2013a for recent overviews). Not only are second language (L2) researchers applying psycholinguistic theories to SLA, but they are also using experimental techniques historically associated with psychology such as eye tracking to better understand how humans process languages. Most psycholinguistic studies investigating L2 processing of morphology and syntax in adult populations have focused on intermediate, advanced, and near-native learners (see, e.g., Jegerski Reference Jegerski2010 for an overview of near-native studies). The few studies examining morphological processing in beginners focus on the first hours of exposure (Rast Reference Rast2008), or the cognitive and neural mechanisms guiding artificial language learning (e.g., DeKeyser Reference DeKeyser1997; de Graaff Reference de Graff1997; DeDiego-Balaguer, Fuentemilla & Rodríguez-Fornells Reference de Diego-Balaguer, Fuentemilla and Rodríguez-Fornells2011). This neglect of novice learners is particularly evident in studies on morphological processing of natural languages on account of adults’ persistent difficulty in processing inflectional markers (e.g., Hopp Reference Hopp2013; Slabakova Reference Slabakova2008; VanPatten, Keating & Leeser Reference VanPatten, Keating and Leeser2012). In effect, the few online morphological studies that include adult beginners show that they lack sensitivity to morphosyntactic agreement violations (see Sagarra & Herschensohn Reference Sagarra and Ellis2013 for a review). However, interestingly, it is also found that sensitivity increases not only with higher proficiency but also with higher working memory span (e.g., Sagarra Reference Sagarra and Han2008; Sagarra & Herschensohn Reference Sagarra and Ellis2013), less complex linguistic systems (e.g., Ellis & Sagarra Reference Ellis and Sagarra2011), and cognitively easier structures (e.g., Han & Liu Reference Han and Liu2013).

Heeding the need for more morphological processing studies with novice learners, and, in particular, the negative correlation between cognitive effort and sensitivity to inflectional markers, and Holme’s (Reference Holme2013) proposal that higher cognitive effort yields less transfer, we examine sensitivity and transfer effects in beginning learners processing morphological concord/discord involving verbal suffixes. To investigate these issues, Romanian, Spanish, and English monolinguals, and Romanian and English learners of Spanish read sentences with subject-verb (SV) agreement/disagreement, while their eye movements were recorded. In Romanian, Spanish, and English, V agrees in person and number with S, and V indicates SV agreement via verbal suffixes. However, Romanian and Spanish are pro-drop languages with a rich morphological structure, whereas English requires an explicit S and has poorer morphology (e.g., English lacks person markers in the past and only distinguishes between third person singular and the rest of persons in the present). The results of this study will shed light on the current debate concerning transfer effects in the early stage of SLA, and the underlying issue of whether first language (L1) and L2 processing share the same cognitive system.

Theoretical background

Inflected languages can mark semantic information such as temporal reference, person, and number with lexical cues (e.g., temporal adverbs, explicit subjects, quantifying adjectives) and morphological cues (e.g., suffixes) (Evans Reference Evans2003). While children use both cues in their L1, there is plenty of evidence that adult learners rely on different cues depending on their proficiency level (Klein Reference Klein1994). Thus, lower proficiency learners mostly use implicit cues, such as chronological order, and lexical cues (e.g., Bardovi-Harlig Reference Bardovi-Harlig2000; Ellis & Sagarra Reference Ellis, Sagarra, Indefrey and Gullberg2010b; Lee Reference Lee2002; Leeser Reference Leeser2004; VanPatten Reference VanPatten, Van Patten and Williams2007), whereas higher proficiency learners are able to employ morphological cues (e.g., Dietrich, Klein & Noyau Reference Dietrich, Klein and Noyau1995; Giacalone-Ramat Reference Giacalone-Ramat1992; Skiba & Dittmar Reference Skiba and Dittmar1992; Starren Reference Starren2001). A question that remains open is why beginning learners show this lexical bias. An obvious observation is that lexical cues are more salient than morphological cues (compare the man with the -s in walks, or yesterday with the -ed in walked). More salient cues are more noticeable and thus easier to learn (e.g., Ellis Reference Ellis, VanPatten and Williams2006, Reference Ellis2007; Goldschneider & DeKeyser Reference Goldschneider and DeKeyser2001; Rescorla & Wagner Reference Rescorla, Wagner, Black and Prokasy1972; Slobin Reference Slobin1992). Let’s take SV agreement in English as an example. In English, S is more salient than V (+suffix) both (a) physically (the S is longer in duration and length than is the suffix of V), appears more often in initial than medial position in a sentence, is less redundant than the V because it precedes the V, and is easier to segment because it is not bound), and (b) psychologically (the S is more reliable: walk can refer to various persons). Although salience modulates SLA (Jiang Reference Jiang2004; Sato & Felser Reference Sato and Felser2008; Zobl & Liceras Reference Zobl and Liceras1994), it affects L1 and L2 alike, and thus cannot explain child–adult differences in language acquisition. Some scholars explain this distinction in terms of previous language experience: first-learned cues modulate the learning of later ones. This study employs eye-tracking methodology to examine possible transfer effects while adult beginning learners whose L1 is morphologically rich (Romanian) or poor (English) process SV agreement/disagreement in a morphologically rich L2 (Spanish). Therefore, we focus the background on online studies exploring native language influence of morphosyntax in non-endstate adult learners (see Lardiere Reference Lardiere2007 for a review of studies on transfer in so-called endstate learners).

From the days of Contrastive Analysis (Lado Reference Lado1957) until now (Herschensohn & Young-Scholten’s [Reference Herschensohn and Young-Scholten2013b] workshop on L1 transfer), the study of transfer in SLA, as well as its relationship with usage-based factors, universal linguistic properties, and sociocultural dynamics, has been the focus of numerous investigations and has generated a debate between those who advocate for strong transfer effects and those who advocate for few or none (particularly relevant to the present study; see Han & Liu Reference Han and Liu2013 for a review of studies on transfer at the L2 initial state). Some scholars posit that adults’ persistent difficulty in processing L2 morphosyntax is due to inadequate L2 grammatical ability: native speakers rely on syntactic information, but adult learners give preference to lexical, semantic, and pragmatic information (e.g., VanPatten Reference VanPatten, Van Patten and Williams2007). According to the Shallow Processing Hypothesis (e.g., Clahsen & Felser Reference Clahsen and Felser2006), this inadequate grammatical processing affects L2 syntactic processing, where transfer is rare, but not L2 morphological processing, where transfer is possible.

In contrast, other researchers argue that the native language influences L2 morphological processing because properties similar in the L1 and the L2 depend on the same cognitive and neural mechanisms (e.g., Bates & MacWhinney Reference Bates, MacWhinney, Wanner and Gleitmann1982; Ellis Reference Ellis, VanPatten and Williams2006; Frenck-Mestre Reference Frenck-Mestre2005; Hopp Reference Hopp2013; Juffs Reference Juffs2005; MacWhinney Reference MacWhinney, Gass and Mackey2011; Schwartz & Sprouse Reference Schwartz, Sprouse, Hoekstra and Schwartz1994; Tokowicz & MacWhinney Reference Tokowicz and MacWhinney2005). Thus, learning that a particular stimulus is associated with a specific outcome makes the later learning of a different stimulus paired with the same outcome more difficult (Kamin Reference Kamin, Campbell and Church1969; Krushke & Blair Reference Kruschke and Blair2000). The blocking of later experienced cues by earlier learned ones (learned attention) determines attention to certain L2 cues in the short term (blocking) and long term (transfer), and may eventually affect L2 ultimate attainment (Ellis Reference Ellis, VanPatten and Williams2006).

Children acquiring their L1 first focus on cues that have the highest availability, regardless of their reliability (MacWhinney, Pléh & Bates Reference MacWhinney, Pléh and Bates1985), because they are at the same time learning about the world and various discourse strategies. For instance, young children do not yet know about the custom of recounting events in their usual script order of occurrence, nor do they clearly understand the exact meaning of temporal adverbs or quantifying adjectives. As a result, they use morphological before lexical means of temporal reference and number (Dale & Fenson Reference Dale and Fenson1996; Nelson Reference Nelson1996; Pawlak, Oehlrich & Weist Reference Pawlak, Oehlrich and Weist2006; Valian Reference Valian2006). Older learners, however, do know these things. For example, adult English native speakers know that temporal adverbs are more reliable than the non-salient and ambiguous verbal inflections (Ellis Reference Ellis2007).

There is mounting evidence that intermediate L2 learners tend to focus on one cue at a time (e.g., VanPatten Reference VanPatten, Van Patten and Williams2007; artificial languages: MacWhinney & Bates Reference MacWhinney and Bates1989; McDonald & MacWhinney Reference McDonald and MacWhinney1991), and that their cue selection depends on the L1 (the richer the L1 morphology, the more reliance on L2 morphological cues) until they become more advanced and are able to attend to L2 cues following L2 settings (Bardovi-Harlig Reference Bardovi-Harlig2009; Bordag & Pechmann Reference Bordag and Pechmann2007; Ellis Reference Ellis, VanPatten and Williams2006; Parodi, Schwartz & Clahsen Reference Parodi, Schwartz and Clahsen2004). If L1 transfer affects SLA, it can be in a positive (L1–L2 similar structure) or negative (L1–L2 different structure; L2 unique structure) way (see Tolentino & Tokowicz Reference Tolentino and Tokowicz2011 for a review of neurolinguistic studies on transfer effects). Next, we review some recent behavioral studies examining online processing of non-adjacent morphosyntactic dependencies.

First, we review self-paced reading studies exploring short-term (L2 proficiency) and long-term (L1 transfer) learned attention effects on L2 morphological processing. For short-term effects and cognitive demand effects, Sagarra and Herschensohn (Reference Sagarra and Herschensohn2010, Reference Sagarra and Herschensohn2013) found that English native speakers’ sensitivity to noun-adjective gender agreement violations in Spanish is influenced by L2 proficiency level (intermediate learners displayed native-like sensitivity to violations but beginners did not) and cognitive demands (greater sensitivity to violations with single-gendered inanimate than double-gendered animate nouns, with gender than number violations, and with higher than lower working memory span). Working memory effects in beginners have also been observed in the verbal domain: Sagarra (Reference Sagarra and Han2008) reported that beginning English learners of Spanish are insensitive to adverb-verb violations unless they have high working memory capacity. These studies suggest that proficiency and cognitive demands affect L2 morphological processing.1 As for evidence of long-term learned attention effects, Jiang (Reference Jiang2004, Reference Jiang2007) and Jiang et al. (Reference Jiang, Novokshanova, Masuda and Wang2011) found that native speakers of a morphologically rich language (Russian) are sensitive to noun-adjective number agreement violations in L2 Spanish (advanced proficiency), whereas native speakers of a language lacking number agreement morphology (Chinese, Japanese) are not. These studies presented the stimuli using a non-cumulative moving window paradigm and are therefore unable to determine whether native language influences attention to morphological information. This is relevant because “it is not clear whether learners will not attend to the later experienced cues at all or whether they will attend to these cues but fail to perceive their relevance” (Sagarra & Ellis Reference Sagarra and Ellis2013). To address this issue, Sagarra and Ellis (Reference Sagarra and Ellis2013) used eye-tracking methodology.

Prior to Sagarra and Ellis (Reference Sagarra and Ellis2013), two eye-tracking studies revealed short-term learned attention effects. First, Ellis and Sagarra (Reference Ellis, Sagarra, Indefrey and Gullberg2010b) investigated how beginning and intermediate English learners of Spanish and English and Spanish monolinguals (control groups) processed adverb-verb/verb-adverb congruent/incongruent sentences. As noted earlier, English has poorer morphology than Spanish. As expected, the beginners were insensitive to tense incongruencies and thus were not biased by the lexical or the morphological cues; the intermediates and English monolinguals were sensitive and were biased more by the lexical cues, and the Spanish monolinguals were sensitive and were biased more by the morphological cues. LaBrozzi (Reference LaBrozzi2010) replicated the study with intermediates with and without an immersion experience, and found that, having been more exposed to verbal inflections, the immersed group was more morphological and less lexical than the non-immersed group. To investigate long-term learned attention effects, Sagarra and Ellis (Reference Sagarra and Ellis2013) asked English and Romanian intermediate and advanced learners of Spanish, as well as English, Romanian, and Spanish monolinguals, to read sentences containing the same type of tense incongruencies used in Ellis and Sagarra (Reference Ellis, Sagarra, Indefrey and Gullberg2010b) and choose one of four pictures after each sentence (two competing for meaning and two for cue type: lexical or morphological). As predicted, the intermediate and advanced learners showed sensitivity to the incongruencies. Most importantly, Romanian learners and Romanian and Spanish monolinguals looked longer at verbs than did English learners and English monolinguals, suggesting L1 transfer effects at intermediate and advanced stages of adult language acquisition.2

These L2 morphological processing studies focus on intermediate and advanced learners and cannot speak about transfer effects at early stages of acquisition. To fill this gap, Ellis and Sagarra (Reference Ellis and Sagarra2010a, Reference Ellis, Sagarra, Indefrey and Gullberg2010b, Reference Ellis and Sagarra2011), and Ellis et al. (Reference Ellis, Hafeez, Martin, Chen, Boland and Sagarra2012) conducted a series of studies with a subset of Latin and found short- and long-term learned attention effects. For the short-term learned attention effects, exposing learners to adverbs or verbs made them rely more on adverbs or on verbs when interpreting and translating adverb-verb and verb-adverb incongruent sentences. Eye-tracking data revealed that the training affected attention allocation in addition to accuracy: the adverb pretraining group looked longer at adverbs and less at verbs than the verb pretraining group, and relied more on adverbs than verbs in the past. For long-term learned attention effects (i.e., L1 transfer), native speakers of Chinese (no morphology) and English (poor morphology) relied more on lexical cues (temporal adverbs) than morphological ones (verbal inflection) in Latin (rich morphology), whereas the opposite applies to native speakers of Spanish and Russian (rich morphology). Interestingly, when the complexity of the Latin verbal paradigm increased, all participants became lexically biased regardless of their L1, suggesting that lexical information is favored over bound morphology in the case of a heavier cognitive load.

Although these studies reveal L1 transfer effects in intermediate and advanced learners of a complete natural language (online data) and in ab initio learners of a miniature language (offline data), it is unclear whether transfer occurs in beginning learners. Some offline studies attribute certain L2 production phenomena to L1 transfer (see Rast Reference Rast2008 for transfer effects of gender and number in French learners of Polish), but there is no direct evidence of L1 transfer in extant online processing data (see Sagarra & Ellis Reference Sagarra and Ellis2013). Hence, the question remains: Is the initial state of the L2 learner “a tabula rasa as for L1 infants, a version of the native grammar, or something in between?” (Herschensohn & Young-Scholten Reference Herschensohn and Young-Scholten2013b).

We cannot assume that the transfer effects obtained in intermediate learners apply also to beginners because the latter have specific characteristics that make L2 morphological processing particularly challenging. First, working memory limitations are obvious at low but not advanced levels for L2 morphological processing (e.g., Sagarra Reference Sagarra and Mackey2007, Reference Sagarra and Han2008; Sagarra & Herschensohn Reference Sagarra and Herschensohn2010). Second, instructors overuse lexical cues when communicating with beginning learners, even when the target language is morphologically rich, such as Spanish (Dracos Reference Dracos2009; Goodall Reference Goodall2008; Gurzynski-Weiss, Daidone & Long Reference Gurzynski-Weiss, Daidone and Long2014; Santilli Reference Santilli1996; Sanz Reference Sanz, Lee and Valdman1999). Finally, the beginners’ peers overuse lexical cues when speaking in class (Bardovi-Harlig Reference Bardovi-Harlig2000). In the present study, to investigate whether there are L1 transfer effects in beginners, native speakers of a morphologically rich L1 (Romanian, Spanish) and a morphologically poor L1 (English) read sentences with SV agreement/disagreement in their L1 (monolinguals) or Spanish (learners), and selected one of four pictures after reading each sentence (two semantically congruent, two grammatically congruent).

The study

The specific research questions and predictions of the study are as follows:

  1. RQ1: Congruency effects. Are English, Romanian, and Spanish monolinguals and English and Romanian beginning learners of Spanish sensitive to SV agreement violations? We hypothesize that the monolinguals, but not the learners, will show sensitivity because SV agreement with third person verbs in the present tense is morphologically marked in English, Romanian, and Spanish, regardless of whether the language is morphologically rich or poor. As for the learners, we predict they will not be sensitive, based on previous studies reporting lack of congruency effects in low proficiency learners (see Sagarra Reference Sagarra and Han2008 for adverb-verb agreement, and Sagarra & Herschensohn Reference Sagarra and Herschensohn2010 for noun-adjective gender and number agreement).

  2. RQ2: Transfer effects. Are Spanish and Romanian monolinguals more morphological than English monolinguals? If yes, are Spanish monolinguals and Romanian learners of Spanish more morphological and less lexical than English learners of Spanish? In other words, does previous experience with a morphologically rich or poor language modulate the learning of a morphologically rich L2 in beginners? First, for monolinguals, we expect Romanian and Spanish monolinguals (rich morphology) to look longer at V than do English monolinguals (poor morphology), and English monolinguals to look longer at S and be more S biased in picture-selection data than Romanian and Spanish monolinguals. Second, for beginning learners, we hypothesize that English and Romanian learners will show lexical processing patterns that are similar to those of English monolinguals, because processing an L2 is cognitively strenuous at early stages of development and learners prioritize lexical over morphological cues. This is based on the findings that learners prefer to process lexical before morphological information when both are present in a sentence (VanPatten Reference VanPatten, Van Patten and Williams2007) and that beginning learners are insensitive to morphosyntactic incongruencies unless they have high working memory capacity (Sagarra Reference Sagarra and Han2008; Sagarra & Herschensohn Reference Sagarra and Herschensohn2010).

  3. RQ3: Learner effects. Do beginning learners process SV agreement like monolinguals? Apart from the obvious fact that learners read more slowly than monolinguals and that lower proficiency learners read in turn more slowly than higher proficiency ones (e.g., Sagarra & Ellis Reference Sagarra and Ellis2013), we predict that both learner groups will be more lexical than the monolingual groups, because morphological cues seem to consume more attentional resources than lexical cues. We based this prediction on Sagarra’s (Reference Sagarra and Han2008) and Sagarra and Herschensohn’s (Reference Sagarra and Herschensohn2010) findings that beginners with more working memory are more sensitive to grammatical violations of non-adjacent dependencies in the verbal and nominal domain than those with less working memory.

Method

Participants

The sample pool was formed by 142 right-handed 18- to 40-year-olds with normal or corrected-to-normal vision, with high school or more education, and with at least 90 percent accuracy on selecting a semantically congruent picture in the eye-tracking tasks. Participants received extra credit or monetary compensation and belonged to one of five groups: English learners of Spanish (n=23), Romanian learners of Spanish (n=19), English monolinguals (n=37), Romanian monolinguals (n=31), and Spanish monolinguals (n=32). The groups were comparable in terms of working memory and inhibitory control (see the Materials section for a description of the tests used to measure working memory and inhibitory control): a series of one-way analyses of variance (ANOVAs) revealed no between-group differences in the working memory test, F(4, 135)=1.819, p>0.05, or the inhibitory control test: accuracy inhibition effects, F(4, 135)=1.690, p>0.05, accuracy facilitation effects, F(4, 135)=1.102, p>0.05, reaction times (RTs) inhibition effects, F(4, 135)=0.328, p>0.05, and RT facilitation effects, F(4, 135)=0.986, p>0.05. Three subjects are missing from the working memory and inhibitory control tests due to technical difficulties. See Table 5.1 below for the means and standard deviations for these tests.

Table 5.1 Descriptive statistics for the cognitive tests

The learner data were collected in a large American university (English learners) and at the Instituto Cervantes in Bucharest (Romanian learners). To avoid possible transfer effects of other foreign languages on Spanish, the learners had no knowledge of other foreign languages, or minimal knowledge of languages morphologically similar to their L1 (some Romanian learners knew some French, but this is not problematic because they had a low proficiency level, and because both Romanian and French are similar in richness of morphology). In addition, all the learners began studying Spanish after puberty, were enrolled in a basic Spanish course based on a placement test, scored significantly lower than intermediate and advanced learners reported in a follow-up study, and scored above 90 percent in the vocabulary and grammar tests to ensure that lack of knowledge of the target lexicon and form did not affect the results. They also took an international standardized Spanish proficiency test, and the results of a t-test for independent samples revealed that the English learners (M=49.10, SD=5.68) and the Romanian learners (M=52.52, SD=5.83) scored similarly: t(1,38)=–1.883, p>0.05. Additionally, the Romanian learners, the Romanian monolinguals, and the Spanish monolinguals took a basic English proficiency test to control for possible English transfer effects, and none scored higher than 85 percent, which was taken as the cut-off point. Finally, the monolingual data were collected in the United States, Romania, and a monolingual central region of Spain. The monolinguals had basic or no knowledge of other morphologically rich languages, had only lived in predominantly monolingual communities, and had not spent more than four months abroad.

Procedure

Participants took between 1 hour 10 minutes (monolinguals) and 2 hours (learners) to complete the tests in this order: language background test, Spanish proficiency test (learners), English proficiency test (English non-natives), eye-tracking task 1 (reading or listening), working memory test, inhibitory control test, eye-tracking task 2 (listening or reading), grammaticality judgment test, vocabulary test (learners), and grammar test (learners). The Spanish sentences of the eye-tracking tasks and the grammaticality judgment tasks were translated into English and Romanian for the English and Romanian control groups. Also, scoring of all the tests, with the exception of the eye-tracking tasks and the cognitive tests, was 1 point for a correct answer and 0 points for an incorrect answer. Finally, half of the participants did the listening eye-tracking task before the reading eye-tracking task, and half the reading eye-tracking task before the listening eye-tracking task, to control for possible practice effects. We describe these tests in the Materials section, in which we will focus on the results of the reading eye-tracking task. Before that, we justify the use of eye-tracking over other online techniques to examine L1 and L2 visual morphological processing in sentence reading in the next section on eye-tracking methodology.

Eye-tracking methodology

Eye tracking is an ideal methodology to explore online morphological processing for several reasons (see Dussias Reference Dussias2010, as well as Sagarra & Seibert Hanson Reference Sagarra and Seibert Hanson2011, for more information on the use of this technique in linguistic research and on its benefits investigating L2 morphological and syntactic processing). First, eye tracking is superior to self-paced reading because the former measures early and late processing mechanisms whereas the latter only provides data about late processing. This is important because sometimes early processing data are more robust than late processing data (e.g., see Wilson & Garnsey Reference Wilson and Garnsey2009). Second, eye tracking is preferable to self-paced reading and event-related potentials (ERPs) because the former allows participants to read complete sentences, whereas the other two techniques force participants to read sentences word by word or region by region, which requires constant attention to the text and consumes a great deal of cognitive resources (Rayner & Clifton Reference Rayner, Clifton and Medin2002). In addition, eye tracking allows participants to return to previously read text. Regressions are relevant because they account for 10 percent of all eye fixations during sentence comprehension (Rayner & Clifton Reference Rayner, Clifton and Medin2002), because first pass viewing times can reflect nonstructural processing (e.g., Trueswell, Tanenhaus & Garnsey Reference Trueswell, Tanenhaus and Garnsey1994), and because there may be late but not early processing effects or early processing effects evident in regressions but not first pass viewing times (e.g., Pynte & Colonna 2000). Third, eye tracking is better than techniques that use words at a fixed rate, such as ERPs with rapid serial visual presentation, because word presentation rate affects sentence comprehension (see Ditman, Holcomb & Kuperberg Reference Ditman, Holcomb and Kuperberg2009 for a review).

Materials

Language background test. This screening test asked questions about participants’ experience with language during and after childhood, such as age of onset, number of years the foreign language was studied, location and length of time living abroad, and current contact hours per week and context (at home, in school, at work, with friends).

Proficiency tests. The Spanish proficiency test was adapted from the Diploma de Español como Lengua Extranjera (DELE) exam and consisted of a multiple-choice section with sixty items assessing grammatical knowledge at the beginner, intermediate and advanced levels, and another with twenty-four items evaluating reading and listening comprehension. The English proficiency test was taken from the grammar section of the Test of English as a Foreign Language (TOEFL) and had twenty multiple-choice items.

Eye-tracking tasks. The eye-tracker was an EyeLink 1000 (SR Research), with sampling rate of 1k Hz, spatial resolution of .32o horizontal and .25o vertical, and averaged calibration error of .01o. For the listening eye-tracking task, participants listened to sentences while they looked at four pictures and indicated their selection with a mouse-click. Because this chapter focuses on the reading task data, we will not describe the listening task in detail. For the reading eye-tracking task, participants saw a dot to recalibrate before reading each sentence; once they fixated on the dot, the sentence appeared, they read it and fixated on a grey box in the lower right corner of the sentence screen; then, a third screen appeared with four pictures, and they clicked the mouse on the one which best represented the sentence they just read (see Figures 5.1a and 5.1b).

Figure 5.1a Sample trial of the SV agreement condition in the eye-tracking task

Figure 5.1b Sample trial of the SV disagreement condition in the eye-tracking task

Both eye-tracking tasks were in Spanish (L2 learners, Spanish monolinguals), English (English monolinguals), or Romanian (Romanian monolinguals). Each task contained eighty-five sentences (5 practice, 16 experimental [8 per condition], 64 distractors). To avoid presenting two experimental sentences of the same type successively, all sentences were distributed in blocks, and each block contained one sentence of each type, following a Latin square design. Additionally, the blocks were randomized and the sentences within each block pseudorandomized. All the sentences used basic words and structures that beginning learners would understand, but had less than 20 percent of cognates to reduce possible lexical priming effects. Also, all the sentences were 9–11 (experimental) or 10–13 (distractors) words long, and contained nouns and verbs that did not appear more than twice in the entire experiment to avoid practice effects. Finally, all the sentences had two- to three-syllable regular transitive verbs. The verbs were selected from the participants’ Spanish textbook, following McLaughlin, Osterhout, and Kim’s (Reference McLaughlin, Osterhout and Kim2004) suggestion that textbook word lists correlate with input frequency in beginning learners with minimum exposure to the target language outside of the classroom.

The experimental sentences had two conditions (\ indicates the break onto the next line):

(1) SV agreement:

(2) SV disagreement:

The experimental sentences followed a fixed syntactic order: adverbial phrase – subject – verb – object – prepositional phrase. The adverbial phrases were always temporal (por la mañana ‘in the morning,’ por la tarde ‘in the evening,’ por la noche ‘at night’). The subjects were animate and human, and half were in singular and half in plural. The verbs were regular, transitive, with two or three syllables, and in the present tense of the indicative mood. Finally, the objects were inanimate. The experimental sentences in Spanish, their English and Romanian translations, and a list of the target words (S and V) can be found in the Appendix.

The picture verification task following the sentence had four pictures: (1) semantically congruent and grammatically congruent with S (following the example in Figure 5.1, a man drinking juice), (2) semantically congruent and grammatically incongruent with S (men drinking juice), (3) semantically incongruent and grammatically congruent with S (a man eating oranges), and (4) semantically incongruent and grammatically incongruent (men eating oranges). To avoid participants anticipating the source of the semantic incongruency, about one-third of the semantic changes occurred at S (e.g., man for men, women for men, children for men), one-third at V (e.g., eating for drinking), and one-third at the object (e.g., oranges for juice). For information about how the picture verification task was scored, see the Scoring section.

Cognitive tests. Participants completed two cognitive tests in E-Prime: the letter-number sequencing test (verbal working memory) and the Flanker test (inhibitory control). The working memory test was adapted from a subtest from the revised version of the Wechsler Adult Intelligence Scale test (WAIS) (Wechsler Reference Wechsler1997), which measures working memory by showing participants a mixed series of letters and numbers (Q-1-B-3-J-2), and asking them to recall them, numbers first in ascending numerical order, then letters in alphabetical order (1-2-3-B-J-Q). Letters and numbers appeared one-by-one in the center of the screen and were preceded by a 500-ms fixation mark. When the word recall appeared, participants typed their answers, hit escape to end the trial, and then spacebar to prompt the next trial. While letters and numbers were visible for a short period of time, no time limit was imposed to recall them. Finally, the test had twenty-one letter-number series and began with series of two items (1 number and 1 letter) and continued to a maximum of eight items (4 numbers and 4 letters). Participants completed three practice trials and three trials at each series length. Participants received 1 point per correct series recalled, to a maximum of 21 points.

In the Flanker test, participants were instructed to respond to the direction of a red target chevron that was surrounded by various other symbols. The red target chevron appeared by itself (baseline trials), or in combination with black diamonds (neutral trials: no facilitation or interference effect), four black flanking chevrons pointing in the same direction as the target (congruent trials: facilitation effect), four black flanking chevrons pointing in the opposite direction to the target (incongruent trials: interference effect), or four black Xs directing participants to refrain from responding (no-go trials). The test had 40 baseline trials (20 trials with 2 chevron directions) and 480 experimental trials (20 trials with 4 conditions: 2 chevron directions and 3 chevron positions). For each trial, participants saw a 500-ms fixation mark, a 1000-ms stimulus, a 300-ms blank screen, and a 100-ms buffer screen to adjust for the monitor refresh rate. Finally, the test elicited four scores: inhibitory effects (incongruent minus neutral conditions) for accuracy and RTs, and facilitatory effects (congruent minus neutral conditions) for accuracy and RTs.

Grammaticality judgment task. For this paper-and-pencil task, participants classified thirty-two sentences (16 experimental, 16 fillers) as correct or incorrect, and identified the source of the error in incorrect sentences with a circle on the incorrect word(s). The thirty-two sentences were similar (same verbs and syntactic structure) but not identical to those used in the eye-tracking tasks, and half were grammatical and half ungrammatical. Although the sentences of the eye-tracking and the grammaticality judgment tasks were similar, they were never identical, to avoid practice effects.

Vocabulary and grammar tests. The vocabulary test assessed lexical knowledge of the target words by asking participants to match a list of Spanish nouns and verbs with their English translation equivalent. The grammar test evaluated grammatical knowledge of the target structure by asking participants to match conjugated Spanish regular verbs in third person singular and plural in the present and the preterit with their English translation equivalent.

Scoring

As mentioned earlier, this chapter will focus on the data of the reading eye-tracking task, for which participants read sentences (reading data) and chose one of four pictures per sentence (picture data). The dependent variables used for statistical analyses for both data sets are shown in Figure 5.2.

Figure 5.2 Dependent variables used in statistical analyses

The reading data measured early processing with first-pass duration on S and on V (henceforth gaze: time spent in the target word before moving on or looking back), and total processing with total duration on S and on V (henceforth total time: duration of all fixations in the target word) (see Eye-tracking methodology section for justification why eye-tracking methodology was used to investigate L2 morphological processing). The picture data measured sentence comprehension and subject bias based on picture selection. For sentence comprehension, participants received 1 point if they chose one of the two semantically congruent pictures. For example, for Figures 5.1a, they received 1 point if they chose the man drinking or the men drinking because the contrast is between drinking and eating. For subject bias, 1 point was awarded if they chose one of the two pictures grammatically congruent with the subject in incorrect sentences. For example, for Figure 5.1b (*the man drink), they received 1 point if they chose the man drinking or the man eating, because their choice followed the number of the subject (in this case, singular), and here the contrast was between singular S (man) and plural S (men).

Statistical analyses

Descriptive statistics are shown in Table 5.2 (reading data) and Table 5.3 (picture data).

Table 5.2 Descriptive statistics for the reading data

Table 5.3 Descriptive statistics for the picture data

For statistical analyses, the reading data followed a gamma distribution (frequency log link), and the picture data a binomial distribution (frequency logit link).3 The six variables from Table 5.2 generated twelve generalized linear mixed model (GLMM) analyses: six comparing the three monolingual groups (henceforth monolingual analyses), and six comparing the Romanian and English monolinguals with the Romanian and English learners (henceforth monolingual-learner analyses or MonoLearner). The GLMM analyses for the reading data and the sentence comprehension picture data had Agreement (agreement/disagreement), L1 (Romanian, English, and also Spanish for the monolingual analyses), and MonoLearner (Romanian and English learners vs. Romanian and English monolinguals), and all their possible interactions, as fixed factors, and participant as a random factor. The S bias picture data did not have Agreement because the bias is only visible in incorrect sentences. Also, MonoLearner was only a fixed factor in the analyses comparing learners with monolinguals. Inferential statistics (main effects and interactions) are shown in Table 5.4 (reading data for monolingual analyses), Table 5.5 (reading data for monolingual-learner analyses), Table 5.6 (picture data for monolingual analyses), and Table 5.7 (picture data for monolingual-learner analyses).

Table 5.4 Inferential statistics for the reading data: monolingual analyses

Table 5.5 Inferential statistics for the reading data: monolingual-learner analyses

Table 5.6 Inferential statistics for the picture data: monolingual analyses

Table 5.7 Inferential statistics for the picture data: monolingual-learner analyses

Results and discussion

Results and discussion are presented in a single section below organized by research question.

RQ1: Congruency effects

The first research question explored whether the participants were sensitive to SV agreement violations. As a reminder, only the reading data measured congruency effects. Our prediction that only the monolinguals would be sensitive to SV incongruencies was supported. In effect, monolinguals, but not learners, looked longer at S and V (gaze and total time) in incongruent than congruent sentences. This is in line with previous research indicating that, at early stages of acquisition, adult learners are insensitive to morphosyntactic violations. For example, Sagarra (Reference Sagarra and Han2008) found that beginners were insensitive to adverb-verb incongruencies unless they had high working memory. And Sagarra and Herschensohn (Reference Sagarra and Herschensohn2010) reported identical results with noun-adjective gender and number agreement.4

Monolinguals. Table 5.4 shows that there was a significant main effect for Agreement in gaze and total time at S and V: English, Romanian, and Spanish monolinguals looked longer at S and V in incongruent than congruent sentences (all p<0.01). Table 5.4 also reveals that there was a significant interaction of Agreement X L1 for V measures, indicating that the monolinguals looked longer at V in incongruent than congruent sentences except the Romanians (non-significant for gaze and total time) and the English monolinguals (non-significant for gaze but significant for total time). Overall, these results indicate that the three monolingual groups were sensitive to incongruencies (evident in S in all groups and V in the English and Spanish monolinguals), and that they all tended to look at V before S. This is why they all showed congruency effects in S gaze (they had seen S after V to realize they were incongruent at S times), but not all in V gaze. We already know that humans do not read in a word-by-word left-to-right way, but little is known about the order in which words are read. These findings indicate that humans process semantically heavier words, such as V, first. Finally, Table 5.6 shows that Agreement did not have a significant main effect or interaction for the sentence comprehension picture data.

Learners vs. monolinguals. Table 5.5 shows that S data generated a significant main effect of Agreement (disagreement > agreement) in total time, and a significant interaction of Agreement X MonoLearner (comparison of the monolinguals with the learners) in gaze and total time. Bonferroni pairwise comparisons revealed that the significant interactions were produced by the monolinguals, but not the learners, being sensitive to SV agreement violations (all p<0.01, except p<0.05 for S gaze).

RQ2: Transfer effects

The second research question had two parts. The first part focused on whether Romanian and Spanish monolinguals are more morphological and less lexical than English monolinguals. The goal of this question was to ensure that the three baseline groups process SV concord/discord in line with the language they speak: morphologically rich, in the case of Romanian and Spanish monolinguals; or morphologically poor, in the case of English monolinguals. The predictions for the monolinguals were supported for Romanian and English monolinguals, and partially supported for Spanish monolinguals. As predicted, Romanian monolinguals looked longer at V than did English monolinguals, and English monolinguals looked longer at S than did Romanian monolinguals. As for Spanish monolinguals, our hypothesis was supported for V data (they looked longer at V than did English monolinguals), but not for S data (both Spanish and English monolinguals looked longer at S than did Romanian monolinguals). We discuss these results in detail within the monolingual analyses subsection.

The second part of the second research question examined whether beginning Romanian learners of Spanish were more morphological and less lexical than English learners of Spanish (transfer effects). In other words, the question for the learners was whether earlier learned L1 cues (English vs. Romanian) can hinder later learned L2 cues at early stages of acquisition, in which L2 processing consumes a great amount of cognitive resources and L2 exposure is still minimal, to build up a morphological bias. Our prediction that the L1’s morphological richness would affect processing in the monolinguals, but not in the learners, was supported. We discuss these findings in detail within the learners vs. monolinguals analyses subsection.

Finally, for the picture data, all participants understood the meaning of the sentences well (minimum 90% accuracy), and most of them selected the picture that matched in number with the subject in incorrect sentences (i.e., lexically biased) (minimum 83% of the time). Table 5.6 displays the results of the inferential statistics for the picture data in the monolinguals, and Table 5.7 compares the monolinguals with the learners. Taking these findings as a whole, we argue that transfer can happen at any L2 developmental stage, but that, at early stages, beginners’ limited L2 knowledge forces them to be selective and economical. Thus, if there is lexical and morphological information encoding the same meaning (in this case, number) in the same sentence, they will process more salient information (like the unbound lexical cue of an overt subject) before less salient information (like the bound morphological cue of a verbal suffix), regardless of whether their L1 is morphologically rich or poor (cf. Han & Liu Reference Han and Liu2013; Park & Han Reference Park, Han and Han2008; Park, this volume).5 That is, the redundant and unsalient nature of the morphological cues in this study made them difficult to process by beginners. This is in line with Sagarra’s (Reference Sagarra and Han2008) findings that beginning English-Spanish learners are not sensitive to Adv-V incongruencies, but that higher working memory learners are more sensitive to incongruencies than are lower working memory ones. Also, Sagarra and Herschensohn (Reference Sagarra and Herschensohn2010) reported that beginning English-Spanish learners are insensitive to N-A gender and number discord, and that higher working memory intermediates are more sensitive to these agreement violations than are lower working memory ones. Finally, our beginning learners are adults, and there is plenty of evidence that processing an L2 as an adult is particularly effortful from a cognitive standpoint (e.g., Hasegawa, Carpenter & Just Reference Hasegawa, Carpenter and Just2002).

Monolinguals. As shown in Table 5.4, there were significant main effects for L1 in V total time and S gaze (all p<0.01) showing that Romanian monolinguals looked longer at V and less time at S than did English monolinguals. In other words, as expected, Romanians are more morphological and less lexical than English monolinguals. As for Spanish monolinguals, they seemed to be in the middle, that is, more morphological than English monolinguals but less so than Romanian monolinguals. We arrive at this conclusion based on (a) a significant interaction of Agreement X L1 showing that Spanish monolinguals looked longer at V than did English monolinguals (gaze and total duration in incorrect sentences, all p<0.05), and (b) a significant main effect of L1 for S revealing that Spanish monolinguals looked longer at S than Romanian monolinguals did (gaze and total duration, all p<0.01). Taken together, the results for the three monolingual groups indicate that Romanian and Spanish monolinguals are more morphological and less lexical than English monolinguals. These findings are in line with Sagarra and Ellis’s (Reference Sagarra and Ellis2013) findings with Romanian, English, and Spanish monolinguals processing adverb-verb congruencies (Romanians and Spaniards were more verbal and anglophones more adverbial). In addition, the finding that Spanish monolinguals are more lexical than Romanian monolinguals has important methodological implications for cross-linguistic studies, as it clearly confirms the need to collect baseline data not only with monolinguals of the target language (in the case of this study, Spanish monolinguals), but also with monolinguals of the L2 learners’ native language (in our case, Romanian and English monolinguals). That is, it is not sufficient to classify a language as morphologically rich or poor. In processing terms, we have demonstrated that morphological richness is not a binary but a continuous variable, and that there is ‘morphological gradation.’

With regard to the picture data shown in Table 5.6, the inferential statistics revealed a significant main effect for both sentence comprehension (p=0.025) and subject bias (p=0.040), but a look at the descriptive statistics from Table 5.3 revealed clear ceiling effects. In effect, all participants understood the sentences correctly 90–98% of the time, and all participants chose the picture that matched in number with the subject in incorrect sentences 83–94% of the time. Consequently, the standard deviation values were extremely low (i.e., very few participants made comprehension errors or chose the picture that matched in number with the verb for incorrect sentences), and the GLMM analyses were unable to run pairwise comparisons. To be conservative, and to ensure that there were no pairwise significant differences, two additional GLMM analyses were carried out with the five groups (the three monolingual groups and the two learner groups): one for sentence comprehension and one for S bias. As expected, none of the GLMMs revealed significant differences: for sentence comprehension, there was no main effect for Agreement, F(1, 1349)=1.067, p=0.302, or Group, F(4, 196)=1.343, p =0.255, and no interaction of Agreement X Group, F (1, 1349)=0.862, p =0.486; and for S bias, there was no main effect for Group, F(4, 154)=1.171, p=0.326. Therefore, the conclusions for the picture data are that all participants understood the meaning of the sentences very well, and that all participants preferred pictures that matched in number with S rather than with V in incorrect sentences. This is so because processing incorrect sentences is cognitively more effortful than processing correct sentences. This also explains why lower proficiency learners are more lexical and less morphological than higher proficiency ones (e.g., Sagarra & Ellis Reference Sagarra and Ellis2013), and why low proficiency learners are insensitive to morphosyntactic violations unless they have high working memory span (e.g., Sagarra Reference Sagarra and Han2008; Sagarra & Herschensohn Reference Sagarra and Herschensohn2010).

Learners vs. monolinguals. As shown in Table 5.5, there was a significant main effect for L1 for all V measures (all p<0.01), generated by Romanian monolinguals looking longer at V than did English monolinguals (these results are discussed above in the analyses comparing Romanian, English, and Spanish monolinguals). With regard to S measures, the addition of the learner groups washed out L1 effects in the monolinguals, but the significant interaction of L1 X MonoLearner in gaze and total duration (all p<0.01) revealed that English monolinguals looked longer at S than did Romanian monolinguals (again, these results are discussed above in the monolingual analyses). Most importantly, there were no transfer effects for the learners. The only exception, a main effect for S total duration, cannot be taken as evidence for transfer for several reasons. First, there are no other measures pointing at transfer. Second, the p value is very close to being non-significant: p=0.049. Third, there was a significant interaction of L1 X MonoLearner and all pairwise comparisons were the same (meaning lack of interaction) except for English monolinguals looking longer at S than did Romanian monolinguals, and Romanian and English learners being alike (this difference is the only way to explain the interaction, given that the other pairwise comparisons had a p<0.01). Finally, the trend (Romanian learners > English learners) is the opposite of the pattern consistently found in Romanian and English monolinguals (English monolinguals > Romanian monolinguals).

RQ3: Learner effects

The last research question examined whether beginning learners process SV agreement/disagreement in the same way as monolinguals. Because this question only applies to the comparison of the monolinguals with the learners, monolingual (only) analyses are not applicable. Our prediction that learners would read slower and regress longer than monolinguals was supported. However, as discussed below, our hypothesis that learners would be more lexical than monolinguals was not.

Learners. As can be observed in Table 5.5, there was a significant main interaction of MonoLearner (monolinguals vs. learners) in all reading measures (all p<0.01): learners looked at S and V longer than monolinguals did (i.e., they are slower readers, a logical finding). However, learners did not seem to be more lexical than monolinguals. Thus, for V data, the interaction of L1 X MonoLearner for V total time was due to L1 (Romanian and Spanish monolinguals looked longer at V than English monolinguals did, but learners behaved similarly). In the same line, for S data, the interaction of Agreement X MonoLearner for gaze and total time was a result of Agreement (monolinguals being sensitive and learners insensitive to SV disagreement), and the interaction of L1 X MonoLearner for gaze and total time was due to L1 (English monolinguals looked longer at S than Romanians did, while English and Romanian learners behaved similarly). Because the finding that learners process information more slowly than monolinguals is not theoretically interesting, we exclude these results from the conclusion.

Conclusion

We investigated the processing of Spanish sentences with SV agreement and disagreement by adult monolinguals (Romanian, Spanish, English) and beginning adult learners (Romanian-Spanish, English-Spanish), using eye tracking. The results confirm a distinction between monolinguals and beginning learners in terms of sensitivity and L1 effects, manifested in S and V gaze and total time reading measures. First, monolinguals, but not beginners, are sensitive to SV violations: Romanian, Spanish, and English monolinguals looked longer at S and V in incorrect than correct sentences, but beginners processed correct and incorrect sentences alike. Second, L1 modulates morphological processing in monolinguals, but not in learners: Romanian and Spanish monolinguals spent more time looking at V (more morphological) and less looking at S (less lexical) than English monolinguals did, but Romanian and English learners processed information in the same manner. These findings are in line with studies reporting insensitivity to morphosyntactic violations at beginning but not later stages of L2 development (e.g., Sagarra Reference Sagarra and Han2008; Sagarra & Herschensohn Reference Sagarra and Herschensohn2010). Most importantly, these findings suggest that beginning learners may be immune to morphological transfer effects. It is important to note that this does not contradict models in favor of transfer, as we argue that transfer effects emerge at intermediate proficiency levels, when morphological processing is cognitively more assimilable and sensitivity to morphological agreement violations appears. We arrive at this conclusion following (a) Han and Liu’s (Reference Han and Liu2013) findings suggesting that transfer requires a certain level of L2 knowledge (cf. Park & Han Reference Park, Han and Han2008), (b) Sagarra and Ellis’s (Reference Sagarra and Ellis2013) results that Romanian monolinguals and Romanian-Spanish learners look longer at V than did English monolinguals and English-Spanish learners in sentences with Adv-V/V-Adv congruencies/incongruencies, and (c) statistical analyses comparing the beginners reported in this chapter with intermediate and advanced learners (work in progress), showing that intermediate and advanced learners transfer but beginners do not.

Based on the L1 results discussed above, we argue that when lexical and morphological information encoding the same meaning (e.g., temporal reference or number) coexists in the same sentence (redundancy), and the morphological information is less salient (e.g., bound suffixes) than lexical information (e.g., unbound temporal adverbs or explicit subjects), the extent to which humans rely more on either morphological or lexical information depends on whether their L1 is morphologically rich or poor. However, if the task is cognitively taxing, they become lexical regardless of their L1. Two pieces of evidence from this study and from previous studies (e.g., Han & Liu Reference Han and Liu2013; Sagarra Reference Sagarra and Han2008; Sagarra & Herschensohn Reference Sagarra and Herschensohn2010) support this statement: (1) in this study, all participants were more S than V biased in picture selection in incorrect sentences, which are more difficult to process than correct ones; and (2) beginners are insensitive to morphosyntactic violations because processing an L2 at early stages of development is cognitively demanding. Considering these findings and the results of previous studies reporting that lower proficiency learners with higher working memory span are more sensitive to morphosyntactic violations than those with lower working memory span, we conclude that transfer occurs as long as processing is not cognitively strenuous.6 When processing is cognitively taxing, lexical information is given priority at the expense of morphological information, which is ignored or processed to a lesser extent.

Finally, this study contributes to the field of SLA with two additional findings. First, all monolinguals looked longer at S (gaze times only, that is, excluding regressions) in incongruent than congruent sentences, indicating that they looked at S after V (incongruencies are evident when processing the second element), even though S physically preceded V in the sentence. We interpret these findings as indicative of monolinguals’ tendency to process semantically heavier words, such as V, before semantically lighter ones, such as S. Interestingly, the beginning learners showed the opposite pattern: although they were insensitive to SV incongruencies (as indicated by non-significant differences in processing congruent vs. incongruent conditions), the means for S gaze in congruent sentences were higher than those in incongruent ones. This can be suggestive of beginners’ preference to read SV sentences in an SV order, and monolinguals in a VS one. The second additional finding refers to Spanish monolinguals being more lexical than Romanian monolinguals, even though both are native speakers of morphologically rich Romance languages. This finding is relevant for the interpretation of previous cross-linguistic studies and the design of new ones, because it confirms the need to collect data from not only monolingual speakers of the learners’ target language (in this case, Spanish) but also monolinguals of the native language of each language group under investigation (in this case, Romanian and English). It also indicates that morphological richness is not a discrete (null, poor, or rich) variable but a continuous one, with an infinite number of values between the ‘points’ of no morphology and rich morphology.

Future research will shed more light on how monolinguals and learners of different proficiency levels process morphological non-adjacent dependencies. First, we need to examine the effects of redundancy (VanPatten Reference VanPatten1996). This study focuses on the processing of redundant morphological information (V is always accompanied by S). A study we have in progress collects data with and without the explicit S, to determine whether morphology is processed in the absence of additional cues to number. Second, it is important to explore the effects of salience. This study concentrates on lexical cues that are salient (unbound explicit S) and morphological cues that are non-salient (bound suffixes). A study in progress compares bound morphology (verbal suffixes) with unbound morphology (auxiliary verbs). Third, it is relevant to examine the effects of input modality. The stimuli of this study are delivered in the written mode. Humans rely on prosody to understand language (Carroll, this volume; Fodor Reference Fodor1998; Pynte & Colonna Reference Pynte and Colonna2002) and extract rules (e.g., De Diego-Balaguer & López-Barroso Reference de Diego-Balaguer and López-Barroso2010; Peña et al. Reference Peña, Bonatti, Nespor and Mehler2002). Prosody also takes place during reading (e.g., Kitagawa & Fodor Reference Kitagawa, Fodor, Fanselow, Féry, Vogel and Schlesewsky2006), but a listening experiment may yield different results, considering that learners cannot process at their own pace or re-listen to previous information. We are currently analyzing the data of the listening eye-tracking task. Fourth, it is significant to confirm that linguistic and cognitive constraints lead to lexical bias. Previous studies reveal that working memory is correlative with sensitivity to morphology, and this study shows that all participants are lexical when choosing the pictures in incongruent sentences, which are more difficult to process than congruent (correct) ones. We are currently analyzing the data of a study ascertaining whether lexical bias correlates with working memory and inhibitory control in monolinguals and learners, and whether lexical bias is associated with higher cognitive demands. Fifth, it is crucial that future studies differentiate between beginning and intermediate learners, instead of grouping the two together as low proficiency learners. Last, it bears pointing out that the conclusions of this study apply only to the languages investigated. Additional studies with other L1s will help determine whether such conclusions are generalizable to other languages with different typological profiles, such as Russian (rich inflectional morphology, but lexically more distant from Spanish than Romanian) and Chinese (no inflectional morphology).

Appendix

\ indicates a line break

Sentences in Spanish

  1. Por la mañana los empleados firman el contrato en\ la reunión.

  2. Por la mañana los empleados firma el contrato en\ la reunión.

  3. Por la noche el chico factura su guitarra en\ el aeropuerto.

  4. Por la noche el chico facturan su guitarra en\ el aeropuerto.

  5. Por la mañana la tía corta el pelo del niño en\ el baño.

  6. Por la mañana la tía cortan el pelo del niño en\ el baño.

  7. Por la tarde los ancianos aprenden alemán en\ la academia.

  8. Por la tarde los ancianos aprende alemán en\ la academia.

  9. Por la mañana las chicas barren la cocina con\ la escoba.

  10. Por la mañana las chicas barre la cocina con\ la escoba.

  11. Por la noche las fans preguntan el nombre de\ la cantante.

  12. Por la noche las fans pregunta el nombre de\ la cantante.

  13. Porla noche la novia cocina pollo para\ el novio.

  14. Por la noche la novia cocinan pollo para\ el novio.

  15. Por la noche el periodista imprime la noticia en\ la oficina.

  16. Por la noche el periodista imprimen la noticia en\ la oficina.

  17. Por la tarde la abuela limpia el vestido en\ el baño.

  18. Por la tarde la abuela limpian el vestido en\ el baño.

  19. Por la noche el hijo escucha el cuento de\ la madre.

  20. Por la noche el nieto escuchan el cuento de\ la abuela.

  21. Por la mañana las mujeres cogen libros de\ la biblioteca.

  22. Por la mañana las mujeres coge libros de\ la biblioteca.

  23. Por la tarde los jefes esperan el vuelo en\ el aeropuerto.

  24. Por la tarde los jefes espera el vuelo en\ el aeropuerto.

  25. Por la tarde la hermana gasta dinero en\ zapatos nuevos.

  26. Por la tarde la hermana gastan dinero en\ zapatos nuevos.

  27. Por la noche las niñas reciben un regalo de\ sus abuelos.

  28. Por la noche las niñas recibe un regalo de\ sus abuelos.

  29. Por la tarde los padres cancelan la fiesta para\ su hijo.

  30. Por la tarde los padres cancela la fiesta para\ su hijo.

  31. Por la tarde el hombre bebe un refresco en \ el bar.

  32. Por la tarde el hombre beben un refresco en \ el bar.

Translation in English

  1. ‘In the morning the employees sign the contract in the meeting.’

  2. ‘In the morning the employees signs the contract in the meeting.’

  3. ‘At night the boy checks in his guitar at the airport.’

  4. ‘At night the boy check in his guitar at the airport.’

  5. ‘In the morning the aunt trims the boy’s hair in the bathroom.’

  6. ‘In the morning the aunt trim the boy’s hair in the bathroom.’

  7. ‘In the afternoon the elders learn some German in the language academy.’

  8. ‘In the afternoon the elders learns some German in the language academy.’

  9. ‘In the morning the ladies clean the kitchen with the broom.’

  10. “In the morning the ladies cleans the kitchen with the broom.”

  11. ‘At night the boys ask the name of the singer.’

  12. ‘At night the boys asks the name of the singer.’

  13. ‘At night the girlfriend cooks chicken for the boyfriend.’

  14. ‘At night the girlfriend cook chicken for the boyfriend.’

  15. ‘At night the journalist prints the news in the office.’

  16. ‘At night the journalist print the news in the office.’

  17. ‘In the afternoon the grandmother cleans the dress in the bathroom.’

  18. ‘In the afternoon the grandmother clean the dress in the bathroom.’

  19. ‘At night the son listens to the mother’s story.’

  20. ‘At night the son listen to the mother’s story.’

  21. ‘In the morning the ladies get books from the library.’

  22. ‘In the morning the ladies gets books from the library.’

  23. ‘In the afternoon the managers await the flight in the airport.’

  24. ‘In the afternoon the managers awaits the flight in the airport.’

  25. ‘In the afternoon the sister spends money on new shoes.’

  26. ‘In the afternoon the sister spend money on new shoes.’

  27. ‘At night the girls get a present from their grandparents.’

  28. ‘At night the girls gets a present from their grandparents.’

  29. ‘In the afternoon the parents cancel the party for their son.’

  30. ‘In the afternoon the parents cancels the party for their son.’

  31. ‘In the afternoon the man drinks soda at the bar.’

  32. ‘In the afternoon the man drink soda at the bar.’

References

Bardovi-Harlig, K. (2000). Tense and Aspect in Second Language Acquisition: Form, Meaning, and Use. Oxford: Blackwell.
Bardovi-Harlig, K. (2009). The use of adverbials and natural order in the development of temporal expression. IRAL, 30(4), 299–320.
Bates, E. and MacWhinney, B. (1982). Functionalist approaches to grammar. In E. Wanner & L. Gleitmann (eds.), Language Acquisition: The State of the Art (pp. 173–218). New York: Cambridge University Press.
Bordag, D. and Pechmann, T. (2007). Factors influencing L2 gender processing. Bilingualism: Language and Cognition, 10, 299–314.
Clahsen, H. and Felser, C. (2006). Grammatical processing in language learners. Applied Psycholinguistics, 27, 3–42.
Dale, P. S. and Fenson, L. (1996). Lexical development norms for young children. Behavioral Research Methods, Instruments, & Computers, 28, 125–127.
de Diego-Balaguer, R., Fuentemilla, Ll., and Rodríguez-Fornells, A. (2011). Brain dynamics sustaining rapid rule extraction from speech. Journal of Cognitive Neuroscience, 23(10), 3105–3120.
de Diego-Balaguer, R. and López-Barroso, D. (2010). Cognitive and neural mechanisms sustaining rule learning from speech. Language Learning, 60(2), 151–187.
de Graff, R. (1997). The eXperanto experiment: Effects of explicit instruction on second language acquisition. Studies in Second Language Acquisition, 19, 249–276.
DeKeyser, R. M. (1997). Beyond explicit rule learning: Automatizing second language morphosyntax. Studies in Second Language Acquisition, 19, 195–222.
Dietrich, R., Klein, W., and Noyau, C. (1995). Acquisition of Temporality in a Second Language. Amsterdam: John Benjamins.
Ditman, T., Holcomb, P. J., and Kuperberg, G. R. (2009). An investigation of concurrent ERP and self-paced reading methodologies. Psychophysiology, 44, 927–935.
Dracos, M. (2009). On the use of Spanish subject personal pronouns by instructors of Spanish as a second language. Unpublished manuscript, Pennsylvania State University, University Park.
Dussias, P. E. (2010). Uses of eye-tracking data in second language sentence processing research. Annual Review of Applied Linguistics, 30, 149–166.
Ellis, N. C. (2006). The Associative-Cognitive CREED. In B. VanPatten and J. Williams (eds.), Theories in Second Language Acquisition: An Introduction (pp. 77–96). Mahwah, NJ: Lawrence Erlbaum Associates.
Ellis, N. C. (2007). Blocking and learned attention in language acquisition. CogSci 2007: Proceedings of the Twenty-Ninth Cognitive Science Conference. Nashville, Tennessee, August 1–4, 2007.
Ellis, N., Hafeez, K., Martin, K. I., Chen, L., Boland, J., and Sagarra, N. (2012). An eye-tracking study of learned attention in adult language acquisition. Applied Psycholinguistics, First View, 1–33. doi:10.1017/S0142716412000501.
Ellis, N. C. and Sagarra, N. (2010a). The bounds of adult language acquisition: Blocking and learned attention. Studies in Second Language Acquisition, 32(4), 1–28.
Ellis, N. C. and Sagarra, N. (2010b). Learned attention effects in L2 temporal reference: The first hour and the next eight semesters. Special issue “The earliest stages of language learning” (ed. P. Indefrey and M. Gullberg). Language Learning, 60(2), 85–108.
Ellis, N. C. and Sagarra, N. (2011). Blocking and learned attention in language acquisition: A replication and generalization study. Studies in Second Language Acquisition, 33(4), 589–624.
Evans, V. (2003). The Structure of Time: Language, Meaning and Temporal Cognition. Amsterdam: John Benjamins.
Fodor, J. D. (1998). Learning to parse?Journal of Psycholinguistic Research, 29, 25–36.
Foote, R. (2011). Integrated knowledge of agreement in early and late English-Spanish bilinguals. Applied Psycholinguistics, 32,187–220.
Frenck-Mestre, C. (2005). Eye-movement recording as a tool for studying syntactic processing in a second language. Second Language Research, 21, 175–198.
Giacalone-Ramat, A. (1992). Grammaticalization processes in the area of temporal and modal relations. Studies in Second Language Acquisition, 14, 297–322.
Goldschneider, J. M. and DeKeyser, R. (2001). Explaining the “natural order of L2 morpheme acquisition” in English: A meta-analysis of multiple determinants. Language Learning, 51, 1–50.
Goodall, G. (2008). Poverty/richness of the stimulus in instructed L2 acquisition of Spanish. Paper presented at the Hispanic Linguistics Symposium, Université de Laval, Quebec.
Grosjean, F. and Lee, P. (2013). The Psycholinguistics of Bilingualism. Malden, MA: Wiley-Blackwell.
Gurzynski-Weiss, L., Daidone, D., and Long, A. Y. (2014). Instructor oral input in Spanish L2 classrooms: Examining patterns of subject expression use. GURT. Georgetown University.
Han, Z.-H. and Liu, Z. (2013). Input processing of Chinese by ab initio learners. Second Language Research, 29(2), 145–164.
Hasegawa, M., Carpenter, P. A., and Just, M. A. (2002). An fMRI study of bilingual sentence comprehension and workload. Neuroimage, 15, 647–660.
Herschensohn, J. & Young-Scholten, M. (2013a). The Cambridge Handbook of Second Language Acquisition. Cambridge University Press.
Herschensohn, J. & Young-Scholten, M. (2013b). Workshop on native language influence in second language acquisition. Nineteenth International Congress of Linguists, Geneva, Switzerland, July 21–27.
Holme, R. (2013). Roundtable: differing perspectives on native language influence in second language acquisition. Nineteenth International Congress of Linguists, Geneva, Switzerland, July 21–27.
Hopp. H. (2013). The development of L2 morphology. Second Language Research, 29(1), 3–6.
Jegerski, J. (2010). Ultimate attainment in SLA: Near-native sentence processing in Spanish. Unpublished doctoral dissertation, University of Illinois at Chicago.
Jiang, N. (2004). Morphological insensitivity in second language processing. Applied Psycholinguistics, 25, 603–634.
Jiang, N. (2007). Selective integration of linguistic knowledge in adult second language learning. Language Learning, 57(1), 1–33.
Jiang, N., Novokshanova, E., Masuda, K., and Wang, X. (2011). Morphological congruency and the acquisition of L2 morphemes. Language Learning, 61(3), 940–967.
Juffs, A. (2005). The influence of first language on the processing of wh-movement in English as a second language. Second Language Research, 21, 121–151.
Kamin, L. J. (1969). Predictability, surprise, attention, and conditioning. In B. A. Campbell and R. M. Church (eds.), Punishment and Aversive Behavior (pp. 276–296). New York: Appleton-Century-Crofts.
Kitagawa, Y. and Fodor, J. D. (2006). Prosodic influence on syntactic judgements. In G. Fanselow, C. Féry, R. Vogel, and M. Schlesewsky (eds.), Gradience in Grammar: Generative Perspectives (pp. 336–358). Oxford University Press.
Klein, W. (1994). Time in Language. London: Routledge.
Kruschke, J. K. and Blair, N. J. (2000). Blocking and backward blocking involve learned inattention. Psychonomic Bulletin & Review, 7, 636–645.
LaBrozzi, R. (2010). Processing of Lexical and Morphological Cues in a Study Abroad Context. Unpublished dissertation, Pennsylvania State University, University Park.
Lado, R. (1957). Linguistics Across Cultures. Ann Arbor: Michigan University Park.
Lardiere, D. (2007). Ultimate Attainment in Second Language Acquisition. Mahwah, NJ: Lawrence Erlbaum Associates.
Lee, J. F. (2002). The incidental acquisition of Spanish future morphology through reading in a second language. Studies in Second Language Acquisition, 24, 55–80.
Leeser, M. (2004). Learner proficiency and focus on form during collaborative dialogue. Language Teaching Research, 8, 55–81.
MacWhinney, B. (2011). The logic of the Unified Model. In S. Gass and A. Mackey (eds.), Handbook of Second Language Acquisition (pp. 85–112). New York: Routledge.
MacWhinney, B. and Bates, E. (1989). The Crosslinguistic Study of Sentence Processing. Cambridge University Press.
MacWhinney, B., Pléh, C., and Bates, E. (1985). The development of sentence interpretation in Hungarian. Cognitive Psychology, 17, 178–209.
McDonald, J. L. and MacWhinney, B. (1991). Levels of learning: A microdevelopmental study of concept formation. Journal of Memory and Language, 30, 407–430.
McLaughlin, J., Osterhout, L., and Kim, A. (2004). Neural correlates of second language word learning: minimal instruction produces rapid change. Nature Neuroscience, 7, 703–704.
Nelson, K. (1996). Language in Cognitive Development. New York: Cambridge University Press.
Park, E. S. and Han, Z.-H. (2008). Learner spontaneous attention in L2 input processing: An exploratory study. In Z.-H. Han (ed.), Understanding Second Language Process (pp. 106–132). Clevedon: Multilingual Matters.
Parodi, T., Schwartz, B., and Clahsen, H. (2004). On the L2 acquisition of the morphosyntax of German nominals. Linguistics, 42, 669–705.
Pawlak, A., Oehlrich, J., and Weist, R. (2006). Reference time in child English and Polish. First Language, 26, 281–297.
Peña, M., Bonatti, L. L., Nespor, M., and Mehler, J. (2002). Signal driven computations in speech processing. Science, 298, 604–607.
Pynte, J. and Colonna, S. (2002). Competition between primary and non-primary relations during sentence comprehension. Journal of Psycholinguistic Research, 30, 569–599.
Rast, R. (2008). Foreign Language Input: Initial Processing. Clevedon: Multilingual Matters.
Rayner, K. and Clifton, C. (2002). Language processing. In D. Medin (ed.), Stevens Handbook of Experimental Psychology: Memory and Cognitive Processes (3rd edn.), Vol. II (pp. 261–316). New York: Wiley.
Rescorla, R. A. and Wagner, A. R. (1972). A theory of Pavlovian conditioning: Variations in the effectiveness of reinforcement and nonreinforcement. In A. H. Black and W. F. Prokasy (eds.), Classical Conditioning II: Current Theory and Research (pp. 64–99). New York: Appleton-Century-Crofts.
Sagarra, N. (2007). From CALL to face-to-face interaction: The effect of computer-delivered recasts and working memory on L2 development. In A. Mackey (ed.), Conversational Interaction in Second Language Acquisition: A Series of Empirical Studies (pp. 229–248). Oxford University Press.
Sagarra, N. (2008). Working memory and L2 processing of redundant grammatical forms. In Z.-H. Han (ed.), Second Language Processing and Instruction: Broadening the Scope of Inquiry (pp. 142–159). Clevedon: Multilingual Matters.
Sagarra, N. and Ellis, N. C. (2013). From seeing adverbs to seeing verbal morphology. Language experience and adult acquisition of L2 tense. Studies in Second Language Acquisition, 35, 1–30.
Sagarra, N. and Herschensohn, J. (2010). The role of proficiency and working memory in gender and number agreement processing in L1 and L2 Spanish. Lingua, 20, 2022–2039.
Sagarra, N. and Herschensohn, J. (2013). Processing of gender and number agreement in late Spanish bilinguals. International Journal of Bilingualism, 17(5), 607–627.
Sagarra, N. and Seibert Hanson, A. (2011). Eyetracking methodology: A user’s guide for linguistic research. Hispanic and Lusophone Linguistics, 4, 543–555.
Santilli, M. (1996). Teacher talk and written materials in the Spanish as a second language classroom and their importance as input for Second Language Acquisition. Unpublished doctoral dissertation, University of Illinois at Urbana-Champaign.
Sanz, C. (1999). What form to focus on? Linguistics, language awareness, and the education of L2 teachers. In J. Lee and A. Valdman (eds.), Meaning and Form: Multiple Perspectives (pp. 3–23). Boston: Heinle & Heinle.
Sato, M. and Felser, C. (2008). Sensitivity to morphosyntactic violations in English as a second language. Ms., University of Essex, Colchester.
Schwartz, B. D. and Sprouse, R. A. (1994). Word order and Nominative Case in nonnative language acquisition: A longitudinal study of (L1 Turkish) German Interlanguage. In T. Hoekstra and B. D. Schwartz (eds.), Language Acquisition Studies in Generative Grammar (pp. 317–368). Amsterdam: John Benjamins.
Skiba, R. and Dittmar, N. (1992). Pragmatic, semantic, and syntactic constraints and grammaticalization: A longitudinal perspective. Studies in Second Language Acquisition, 14, 323–349.
Slabakova, R. (2008). Meaning in the Second Language. Studies in Language Acquisition Series, Berlin/New York: Mouton de Gruyter.
Slobin, D. I. (1971). Data for the symposium. In D. I. Slobin (ed.), The Ontogenesis of Grammar (pp. 3–14). New York: Academic Press.
Slobin, D. I. (1992). Psycholinguistics. Glenview, IL: Scott, Foresman and Company.
Starren, M. (2001). The Second Time: The Acquisition of Temporality in Dutch and French as a Second Language. Utrecht: LOT.
Tokowicz, N. and MacWhinney, B. (2005). Implicit and explicit measures of sensitivity to violation in second language grammar: An event-related potential investigation. Studies in Second Language Acquisition, 27, 173–204.
Tolentino, L. C. and Tokowicz, N. (2011). Across languages, space, and time: A review of the role of cross-language similarity in L2 (morpho)syntactic processing as revealed by fMRI and ERP methods. Studies in Second Language Acquisition, 33, 31–125.
Trueswell, J. C., Tanenhaus, M. K., and Garnsey, S. M. (1994). Semantic influences on parsing: Use of thematic role information in syntactic ambiguity resolution. Journal of Memory and Language, 33, 285–318.
Valian, V. (2006). Young children’s understanding of present and past tense. Language Learning and Development, 2, 251–276.
VanPatten, B. (1996). Input Processing and Grammar Instruction: Theory and Research. Norwood, NJ: Ablex.
VanPatten, B. (2007). Input processing in adult second language acquisition. In B. Van Patten and J. Williams (eds.), Theories in Second Language Acquisition: An Introduction (pp. 115–136). Mahwah, NJ: Lawrence Erlbaum Associates.
VanPatten, B., Keating, G. D., and Leeser, M. (2012). Missing verbal inflections as a representational problem: Evidence from self-paced reading. Linguistic Approaches to Bilingualism, 2(2), 109–140.
Wechsler, D. (1997). Wechsler Adult Intelligence Scale – 3rd Edition (WAIS-3®). San Antonio, TX: Harcourt Assessment.
Wilson, M. P. and Garnsey, S. (2009). Making simple sentences hard: Verb bias effects in simple direct object sentences. Journal of Memory and Language, 60, 368–392.
Zobl, H. and Liceras, J. M. (1994). Functional categories and acquisition orders. Language Learning, 44, 159–180.

This research was partially supported by a National Science Foundation Grant BSC-0717557 and a Research Council Grant from Rutgers University. Special thanks to the Instituto Cervantes de Bucarest, Romania.

1 While Foote (Reference Foote2011) reported no working memory effects in L2 processing of SV or NAdj agreement, her results could be due to her participants (advanced vs. Sagarra’s beginning/intermediate learners), scoring (words recalled vs. Sagarra’s composite score), statistical analyses (correlations vs. Sagarra’s ANCOVAs), and manipulation of distance between the agreement source and the target (distance increases syntactic complexity and Waters & Caplan’s claim that working memory does not affect syntactic processing).

2 The notion that there are “morphology rich” (cum free word order) parsing strategies from the L1 and “lexical biased” ones dates back to the early days of the Competition Model (e.g., Bates & MacWhinney Reference Bates, MacWhinney, Wanner and Gleitmann1982). We conclude that native speakers of morphologically rich languages look longer at verbs than native speakers of morphologically poor languages following the Competition Model’s assumption that there are competing cues (morphology, word order, animacy, etc.) with differential weighting according to the native language (which will bias learner attention initially), and that transfer can only occur if some level of similarity is present (see MacWhinney Reference MacWhinney, Gass and Mackey2011 for a review of the model and its tenets). English, Romanian, and Spanish have SV number agreement in the third person singular of the present tense, so transfer is theoretically possible in both English and Romanian learners of Spanish. Thus, the difference between English and Romanian learners lies in the degree of morphological richness of their native language, rather than in the possibility that transfer takes place.

3 Gamma distribution refers to a two-parameter family of continuous probability distributions, whereas binomial distribution is the discrete probability distribution of the number of successes in a sequence of n independent yes–no answers. Log link is the exponent to which another fixed value, the base, must be raised to produce that number, whereas the logit of a probability is the logarithm of the odds.

4 Lack of congruency effects in behavioral studies should be taken with caution, as there is ERP evidence of beginners’ sensitivity to morphological agreement violations (e.g., Tokowicz & MacWhinney Reference Tokowicz and MacWhinney2005). This discrepancy could be due to behavioral techniques not being fine-grained enough to capture novices’ weak sensitivity.

5 Han and Liu’s argument (see also Park & Han Reference Park, Han and Han2008) is that transfer presupposes linguistic readiness, or some knowledge of the target language.

6 Sagarra and Ellis (Reference Sagarra and Ellis2013) investigated the possibility that learners focus more on lexical cues because they normally precede morphological cues by asking the learners to read both AdvV and VAdv sentences. They found that all participants spent more time on their preferred cue, regardless of its location.

Footnotes

This research was partially supported by a National Science Foundation Grant BSC-0717557 and a Research Council Grant from Rutgers University. Special thanks to the Instituto Cervantes de Bucarest, Romania.

1 While Foote (Reference Foote2011) reported no working memory effects in L2 processing of SV or NAdj agreement, her results could be due to her participants (advanced vs. Sagarra’s beginning/intermediate learners), scoring (words recalled vs. Sagarra’s composite score), statistical analyses (correlations vs. Sagarra’s ANCOVAs), and manipulation of distance between the agreement source and the target (distance increases syntactic complexity and Waters & Caplan’s claim that working memory does not affect syntactic processing).

2 The notion that there are “morphology rich” (cum free word order) parsing strategies from the L1 and “lexical biased” ones dates back to the early days of the Competition Model (e.g., Bates & MacWhinney Reference Bates, MacWhinney, Wanner and Gleitmann1982). We conclude that native speakers of morphologically rich languages look longer at verbs than native speakers of morphologically poor languages following the Competition Model’s assumption that there are competing cues (morphology, word order, animacy, etc.) with differential weighting according to the native language (which will bias learner attention initially), and that transfer can only occur if some level of similarity is present (see MacWhinney Reference MacWhinney, Gass and Mackey2011 for a review of the model and its tenets). English, Romanian, and Spanish have SV number agreement in the third person singular of the present tense, so transfer is theoretically possible in both English and Romanian learners of Spanish. Thus, the difference between English and Romanian learners lies in the degree of morphological richness of their native language, rather than in the possibility that transfer takes place.

3 Gamma distribution refers to a two-parameter family of continuous probability distributions, whereas binomial distribution is the discrete probability distribution of the number of successes in a sequence of n independent yes–no answers. Log link is the exponent to which another fixed value, the base, must be raised to produce that number, whereas the logit of a probability is the logarithm of the odds.

4 Lack of congruency effects in behavioral studies should be taken with caution, as there is ERP evidence of beginners’ sensitivity to morphological agreement violations (e.g., Tokowicz & MacWhinney Reference Tokowicz and MacWhinney2005). This discrepancy could be due to behavioral techniques not being fine-grained enough to capture novices’ weak sensitivity.

5 Han and Liu’s argument (see also Park & Han Reference Park, Han and Han2008) is that transfer presupposes linguistic readiness, or some knowledge of the target language.

6 Sagarra and Ellis (Reference Sagarra and Ellis2013) investigated the possibility that learners focus more on lexical cues because they normally precede morphological cues by asking the learners to read both AdvV and VAdv sentences. They found that all participants spent more time on their preferred cue, regardless of its location.

Figure 0

Table 5.1 Descriptive statistics for the cognitive tests

Figure 1

Figure 5.1a Sample trial of the SV agreement condition in the eye-tracking task

Figure 2

Figure 5.1b Sample trial of the SV disagreement condition in the eye-tracking task

Figure 3

Figure 5.2 Dependent variables used in statistical analyses

Figure 4

Table 5.2 Descriptive statistics for the reading data

Figure 5

Table 5.3 Descriptive statistics for the picture data

Figure 6

Table 5.4 Inferential statistics for the reading data: monolingual analyses

Figure 7

Table 5.5 Inferential statistics for the reading data: monolingual-learner analyses

Figure 8

Table 5.6 Inferential statistics for the picture data: monolingual analyses

Figure 9

Table 5.7 Inferential statistics for the picture data: monolingual-learner analyses

Accessibility standard: Unknown

Why this information is here

This section outlines the accessibility features of this content - including support for screen readers, full keyboard navigation and high-contrast display options. This may not be relevant for you.

Accessibility Information

Accessibility compliance for the HTML of this chapter is currently unknown and may be updated in the future.

Save book to Kindle

To save this book to your Kindle, first ensure no-reply@cambridge.org is added to your Approved Personal Document E-mail List under your Personal Document Settings on the Manage Your Content and Devices page of your Amazon account. Then enter the ‘name’ part of your Kindle email address below. Find out more about saving to your Kindle.

Note you can select to save to either the @free.kindle.com or @kindle.com variations. ‘@free.kindle.com’ emails are free but can only be saved to your device when it is connected to wi-fi. ‘@kindle.com’ emails can be delivered even when you are not connected to wi-fi, but note that service fees apply.

Find out more about the Kindle Personal Document Service.

Available formats
×

Save book to Dropbox

To save content items to your account, please confirm that you agree to abide by our usage policies. If this is the first time you use this feature, you will be asked to authorise Cambridge Core to connect with your account. Find out more about saving content to Dropbox.

Available formats
×

Save book to Google Drive

To save content items to your account, please confirm that you agree to abide by our usage policies. If this is the first time you use this feature, you will be asked to authorise Cambridge Core to connect with your account. Find out more about saving content to Google Drive.

Available formats
×