Session 3G

A debated topic in psycholinguistics concerns the effects of bilingualism on cognition. The term 'bilingual advantage' refers to results that show that bilinguals may perform better than monolinguals in certain cognitive tasks, a claim has been confirmed across various populations. At the same time, a growing number of studies report a bilingual cognitive disadvantage (Gollan et al. 2002). In executive functioning (EF), different studies report contradictory findings, even when using the same tasks (Costa et al. 2008, Duñabeitia et al. 2014). While explanations have been offered (Leivada et al. 2021), the degree to which bilingual disadvantages are counterbalanced by bilingual advantages is unclear. The present systematic review addresses this question. 
Our starting point is the biological notion of trade-off, a negative correlation between processes that make use of the same finite resources within an organism (West-Eberhard 2003). We performed a systematic review of the literature on the bilingual disadvantage following the PRISMA Statement (Liberati et al. 2009) in PsycInfo, PsycExtra, APA Journals, and PubMed. Keywords were: "bilingual", "disadvantage". 
Results were split in the following categories: (i) studies that report both an advantage and a disadvantage; (ii) studies that report neither; (iii) studies that report only an advantage; (iv) studies that report a disadvantage for a population for which an advantage was also reported in the literature; (v) studies that report a disadvantage for a population for which no advantage was found in the literature.
Encompassing a variety of tasks and populations, less than half of the screened studies report only a bilingual disadvantage. Excluding category (ii) (i.e. null evidence), remaining studies either report both an advantage and a disadvantage or only an advantage. To obtain the overall picture, we performed a second search of the literature outside the PRISMA sample, which targeted the populations mentioned in studies that reported only a disadvantage. 100% of these populations had been linked to bilingual advantages in other studies. This equilibrium suggests that 'bilingual advantages' and 'bilingual disadvantages' should be conceived as inseparable parts of an overall fitness trade-off. 

References
Costa, A., Hernández, M. & Sebastián-Gallés, N. 2008. Bilingualism aids conflict resolution: Evidence from the ANT task. Cognition 106: 59-86.
Duñabeitia, J. A., J. A. Hernández, E. Antón, P. Macizo, A. Estévez, L. J. Fuentes & M. Carreiras. 2014. The inhibitory advantage in bilingual children revisited: Myth or reality? Experimental Psychology 61: 234-251.
Gollan, T. H., Montoya, R. I. & Werner G. 2002. Semantic and letter fluency in Spanish-English bilinguals. Neuropsychology 16: 562-576.
Leivada, E., Westergaard, M., Duñabeitia, J. A. & Rothman, J. 2021. On the phantom-like appearance of bilingualism effects on neurocognition: (How) should we proceed? Bilingualism: Language & Cognition 24: 197-210.
Liberati, A., Altman, D. G., Tetzlaff, J., Mulrow, C., Gøtzsche, P. C., Ioannidis, J. P. A., et al. (2009). The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate health care interventions: explanation and elaboration. PLoS Med. 6:e1000100.
West-Eberhard, M. J. 2003. Developmental Plasticity and Evolution. Oxford: OUP.

Producing written or spoken output is widely believed to facilitate second language (L2) development (Gass, 2010), possibly because output promotes noticing (Swain, 2005), that is, the conscious registration of a mismatch between what L2 learners want to produce and what they are able to produce. Empirical support for the noticing function of output has been mixed (compare, e.g., Uggen, 2012, and Izumi & Bigelow, 2000), in part due to the use of measures of noticing with disputable construct validity (Uggen, 2012). This study takes advantage of eye tracking, the concurrent registration of participants' eye movements, to operationalize noticing as a hybrid construct involving both attention and awareness (Godfroid et al., 2013). We explore whether producing written output (as opposed to receiving more written input) promotes attention to the English past hypothetical conditional during reading, which is a difficult target structure that lacks a first language (L1) counterpart for Chinese learners of English (e.g., Bassetti, 2021). 
Sixty-eight L1 Chinese speakers with intermediate English proficiency were randomly assigned to an experimental/output group (EG, n = 34) and a control/non-output group (CG, n = 34). Both groups read an essay (adapted from Uggen [2012]) twice as input, during which time their eye-movements were recorded. After each reading, they completed a post-reading task, which was essay reconstruction (output) for the EG and reading comprehension (non-output) for the CG. 
Results showed that the output group but not the non-output group improved in their written production accuracy from pretest to posttest, with a medium effect size. Interestingly, the eye-tracking data showed selective effects of output on the late measure of rereading only, which is associated with controlled processing and conscious reanalysis of input (e.g., Maie & Godfroid, 2021). These findings adduce support for the noticing function of output and help to refine it by showing that the benefits of written output accrue when L2 learners are given an opportunity to engage deeply with task-relevant input and engage in additional, conscious analysis of the relevant linguistic structure.
References
Bassetti, B. (2021). Language and counterfactual reasoning in Chinese, English and ChineseL1-EnglishL2 reasoners. International Journal of Bilingualism, 1-22.
Gass, S. M. (2010). The relationship between L2 input and L2 output. In E. Macaro (Ed.), The Bloomsbury companion to second language acquisition (pp. 194–219). London, GBR: Continuum International Publishing.
Godfroid, A., Boers, F., & Housen, A. (2013). An eye for words. Gauging the role of attention in Incidental L2 vocabulary acquisition by means of eye-tracking. Studies in Second Language Acquisition, 35, 483–517.
Izumi, S., & Bigelow, M. (2000). Does output promote noticing and second language acquisition? TESOL Quarterly, 34(2), 239–278.
Maie, R., & Godfroid, A. (2021). Controlled and automatic processing in the acceptability judgment task: An eye‐tracking study. Language Learning, 1-40.
Swain, M. (2005). The Output Hypothesis: Theory and research. In E. Hinkel (Ed.), Handbook of research in second language teaching and learning (pp. 471–483). Mahwah, NJ: Lawrence Erlbaum Associates.
Uggen, M. S. (2012). Reinvestigating the noticing function of output. Language Learning, 62(2), 506–540.

Learning second (L2) or third (L3) language grammar is influenced by a constellation of learner characteristics (e.g., cognitive aptitudes), language features (e.g., complexity, typology), and exposure-related variables (e.g., type of instruction). Regarding explicit knowledge of foreign language grammar, different studies found that it can be predicted by rule inferencing skills, fluid intelligence, or memory capacity. Still, only few studies have considered those variables jointly in controlled laboratory experiments. 
In the present study, a novel miniature language based on standard Italian (MiniItaliano) was used to investigate which subject-related and item-related variables contribute to the incidental learning of L3 grammar in controlled exposure conditions. The training took the form of a computerized game designed as an employee training at an Italian supermarket. After initial exposure to Italian nouns, adjectives and verbs, the participants responded to utterances in Minitaliano (e.g., customer orders) by choosing the right picture, after which they received feedback. Unbeknownst to them, the covert goal of the training was to incidentally learn the target syntactic structures in L3 (SVA, noun agreement, and word order). For noun agreement, the stimuli were balanced in terms of number, gender and congruence with L1 patterns. Moreover, up to 15% of the input was based on the rule that was alternative to the default morphological regularity coding gender agreement in MiniItaliano.
The participants were 45 native speakers of Polish, with L2 English and no prior knowledge of any Romance language. Apart from the training, the experimental procedure included rule-awareness questionnaires and cognitive tasks of: (1) non-verbal intelligence (Raven's Advanced Progressive Matrices), (2) grammar inductive abilities (the Artificial Language task from the TUNJO battery [Rysiewicz, 2012]) and (3) working memory (the Polish Reading Span [Biedroń & Szczepaniak, 2012]). During the final session, the participants performed a surprise posttest, which included a written untimed Grammaticality Judgment Task (GJT) tapping into explicit knowledge of grammar, and a timed auditory GJT tapping into automatized explicit knolwedge. The untimed GJT involved judgment attributions (i.e., whether each response stemmed from a linguistic rule or intuition). 
Mixed effects logistic regression models showed that variability in explicit knowledge of L3 grammar was primarily predicted by grammar inductive abilities, and, to a lesser degree, by fluid intelligence and working memory. Moreover, reliance on rules (rather than intuitions) was a strong predictor of accuracy in the GJT. Regarding item-related variables, morphological patterns of gender agreement were much easier to acquire in the case of plural noun phrases (NPs) than singular NPs. This might be attributed to the fact that the non-default rule implemented into MiniItaliano interfered more with the singular forms than with the plural forms. The relationships found for explicit knowledge were largely replicated for automatized explicit knowledge, but the effects were weaker, which reflects limitations of applying explicit rules caused by time constraints. Overall, the results suggest that developing knolwedge of L3 grammar in incidental learning conditions is primarily contingent on the ability to infer and apply metalinguistic rules, and can be hampered by unreliable input.

Homophone priming refers to the effect that words with an identical phonological form but unrelated meaning (homophones) can co-activate each other, so that a homophone automatically supports contextually irrelevant meaning. Consumer research has identified homophone priming as a promising subliminal persuasion strategy, for example with the word bye increasing the willingness to buy (Davis & Herr, 2014). Theoretically, the effect requires insufficient suppression of the irrelevant meaning (Gernsbacher, 1993), which has been found to occur under cognitive load and in children. This leads to interesting yet unexplored predictions for bilinguals. As L2 comprehension requires higher processing load but is also associated with greater cognitive control (Thoma & Baum, 2019), homophone priming could be stronger or weaker than in L1. 
Against this background, we investigated if homophone priming biases preference formation under cognitive load differentially in L1 and L2. We used a 2 (prime: control vs. homophone) x 2 (cognitive load: no load vs. load) x 2 (language: L1 German vs. L2 English) design, with prime and cognitive load condition as between-subjects factors and language as within-subjects variable. The online-survey study included 138 participants and 12 control/homophone pairs per language. In each trial, participants first read an advertisement. Its pictorial background equally biased comprehension towards (a) meaning of a non-ambiguous control word and (b) the contextually irrelevant meaning of a homophone word, e.g., a picture of a piggybank with the slogan "Leisure program {(a) deal / (b) sale}! All activities 50% off" (Figure 1). Next, participants rated their attitude towards the advertisement to increase attention. On the subsequent target-screen, they rated their preference towards one of two pictorial choice options, one representing the unsupported homophone meaning and the other the meaning of a distractor (Figure 1). The cognitive load groups had to remember a 5-letter string throughout and was tested on their memory after each trial. Phonological awareness was used as a co-variate. Participants' degree of bilingualism was assessed with several self-report measures. Data were analyzed with mixed effects regression models.
Figure 1. Examples of stimuli from a trial in the English homophone condition.pdf

The results (see Figure 2) showed that the presence of homophones in advertising slogans biased participants' preference towards the contextually unsupported homophone meaning, yet only under cognitive load (straight line). Surprisingly, this interaction was almost identical in L1 and L2. We discuss presumably reciprocal effects of load and control in L2 processing which have methodologically and theoretically interesting implications.
Figure 2. Co-variate adjusted interaction plots for preference.pdf



References
Davis, D. F., & Herr, P. M. (2014). From bye to buy: Homophones as a phonological route to priming. Journal of Consumer Research, 40(6), 1063–1077. https://doi.org/10.1086/673960
Gernsbacher, M. A. (1993). Less skilled readers have less efficient suppression mechanisms. Psychological Science, 4(5), 294–298. https://doi.org/10.1111/j.1467-9280.1993.tb00567.x
Thoma, D., & Baum, A. (2019). Reduced language processing automaticity induces weaker emotions in bilinguals regardless of learning context. Emotion, 19(6), 1023–1034. https://doi.org/10.1037/emo0000502