FABBS Undergraduate Research Excellence Award

Eva McAlister López, University of Michigan, Department of Psychology

How Do a Bilingual’s Two Languages Interact? Cross-Linguistic Transfer of Morphological Awareness in Spanish-English Bilinguals

Eva McAlister López, University of Michigan, Department of Psychology

Abstract

How do bilingual children learn to read words? This study addresses a long-standing debate over what should inform teaching approaches to reading, especially when it comes to bilingual learners. More specifically, we examined morphological awareness and its relation to reading skill across languages in 100 Spanish-English bilingual children, ages 5-10 (M = 8.10, SD = 1.43). Children completed standardized language and reading games as well as our lab-designed measures of morphological awareness in English and Spanish (Marks et al., 2021). Analyses revealed that reading skills within and between each language were significantly correlated. Structural equation modeling revealed the inadequacy of exclusively looking to within-language models to explain literacy development, given that morphological awareness appears to transfer from Spanish to English through the mediator of word reading. This suggests that an accurate understanding of bilingual reading development must consider the cross-linguistic transfer effects of morphological awareness. In our increasingly bilingual world, understanding the nature of this transfer may provide valuable insight into the pedagogical implications of literacy education, and how to maximize bilingual learner outcomes.

Introduction/Background

The goal of reading is to be able to look at printed words, and interpret meaning. The meaningful units in language are called morphemes. Morphemes encode information about meaning, grammar, and pronunciation (Kuo & Anderson, 2006). Morphological awareness, therefore, refers to an individual’s sensitivity to these smallest units of meaning. Given their role in conveying information, how do children capitalize on their morphological awareness skills as they learn to read? How might growing up with two languages and two different morphemic systems complicate this question? 

Being able to recognize the morphemes in a word becomes increasingly important throughout literacy development. If the reader can recognize familiar morphemes, for instance, they will more efficiently access crucial information about words, even those that are unfamiliar (Kuo & Anderson, 2006). As an example, if a child fails to recognize the word unhelpful but can break it down into meaningful units: un + help + ful, they might recognize the root of the word, help, and be able to approximate the entire definition of the word. Additionally, recognizing morphemes in words can provide clarity in cases where pronunciation is ambiguous (Kuo & Anderson, 2006). For example, recognizing the distinct morphemes in cathouse versus cathedral may help a child disambiguate the pronunciation of the “th” letter combination when spelling alone cannot. Thus, morphological awareness helps to navigate single word reading and reading comprehension because it helps to decode meaning and pronunciation, especially when such information is not already provided by vocabulary knowledge.

But how do such metalinguistic skills operate in the mind of a bilingual, someone who communicates in both English and Spanish? Theories on bilingualism suggest that a bilingual’s two languages interact and influence one another. This means that a bilingual child learning to read in Language A may be influenced along the way by skills that they gained in Language B. This type of interaction has, for instance, been extensively observed with the sensitivity to sound segments and structures otherwise known as phonological awareness (Chung et al., 2019; Durgunoğlu et al, 1993; Wise et al., 2016). Similarly, research suggests that morphological awareness may also cross language boundaries. 

Recent findings propose a variety of ways in which morphological awareness in one language may affect literacy in a second language. Spanish speaking English Language Learners (ELLs) showed that morphological awareness measured in their first language, Spanish, could contribute to their word reading in English (Ramirez et al., 2010). Morphological awareness was also revealed to improve vocabulary across English and Spanish, in large part because these languages share a great deal of similar morphemes. For instance, a word like “inevitable” is structurally and meaningfully identical in both languages because it consists of morphemes found in both (in-, -evit-, -able). Finally, Bérubé and Marinova-Todd (2014) highlight the transfer of morphological awareness to reading comprehension. However, little is known about the nuances and directionality of this transfer. We pick up with the question: How does morphological awareness in a bilingual’s two languages contribute to their reading ability in each language (English and Spanish)? 

Methods

Participants

This study recruited 100 bilingual children, ages 5-11 (M = 8.10, SD = 1.43), spanning a range of racial backgrounds, as long as at least one parent was a native Spanish speaker, and Spanish was spoken in the home. Children were only allowed to participate if they reported some proficiency in both English and Spanish and also typical language development. Unlike most English-Spanish bilingual participant groups, these children were not English Language Learners (ELLs)—which usually implicates a great disparity in proficiency between the two languages—but rather, were more equally capable in both languages. Researchers obtained assent from the participants and consent from their parents, as well as demographics and language background questionnaires. Participants were compensated both monetarily and with a toy prize at the end of the session. This study is approved by the institutional review boards. 

Measures of Language, Literacy & Cognitive Development 

Data collection involved a 3-4 hour session in which participants completed a series of standardized language and literacy assessments in both Spanish and English. These assessments were administered by native speakers of English and Spanish. 

English and Spanish vocabulary was assessed using the Peabody Picture Vocabulary Test Fifth Edition (PPVT-5; Dunn, 2018) and the Test en Vocabulario en Imagenes Peabody (TVIP; Dunn et al,. 1986), respectively. In this task, the child is shown a group of four different images and is asked to select the one that best illustrates the vocabulary word provided. English and Spanish Morphological awareness was assessed using an 

experimenter-developed measure called Early Lexical Morphology Measure (ELMM; Marks et al., 2021) and Early Lexical Morphology Measure—Spanish (ELMMS; Marks et al., 2021) respectively. Children listened as they were first given a specific prompt word, then an

incomplete sentence. They were then asked if they could modify the initial prompt word so that it would complete the sentence in a sensical way. For instance, the administrator would state a word, like “friendly” and then a sentence like, “She is my best ____.” and then ask what word could be used to finish the sentence. 

English and Spanish Single word reading was assessed using the Letter Word Identification subtest from the Woodcock Johnson IV battery in English (WJ-LWID; Schrank et al., 2014) and the Woodcock-Muñoz Batería III Pruebas de Aprovechamiento in Spanish (Muñoz-Sandoval et al., 2009), respectively. For this task, the participant was told to read words from a printed list, and they received a point for each word correctly articulated. 

English and Spanish reading comprehension was measured with the reading comprehension subtest of the Woodcock Johnson IV battery in English (WJ-LWID; Schrank et al., 2014) and the Woodcock-Muñoz Batería III Pruebas de Aprovechamiento in Spanish (Muñoz-Sandoval et al., 2009), respectively. During testing, the participant was shown a short passage with a missing word, asked to read the passage silently to themselves and then to respond with a one-word answer for what best would fit the blank.

Results

Descriptive Statistics 

Participants’ home environment was mostly Spanish-dominant and their school environment was nearly exclusively English. Table 1 presents means and standard deviations of children’s language and literacy scores in both languages. Notably, vocabulary was in the typical range for both languages (English: M = 99.5, SD = 19.2; Spanish: M = 107.4, SD = 17.7). Participants also had age-appropriate reading skills in each language (English word reading: M = 108.2, SD = 18.2; Spanish word reading: M = 110.5, SD = 24.9). 

Correlations 

Correlations between all Spanish and English measures are presented in Table 2. All correlations, both within and across language, are significant at p < .01. Exclusively within English measures, the strongest correlation was between word reading and reading comprehension (r = .77), and the same two measures correlated most strongly among Spanish measures (r = .84). Cross-linguistically, the most significant correlation was found between the word reading measures in English and Spanish. 

Morphological awareness in each language correlated with similar measures. Spanish morphological awareness was most highly correlated with Spanish reading comprehension (r = .66) and English reading comprehension (r = .62), while English morphological awareness was most highly correlated with English reading comprehension (r = .65). 

Structural Equation Models 

We ran two path analyses to examine the associations between morphological awareness, word reading, and reading comprehension both within and between languages. In the first model, we only allowed for within-language paths. In the second model, we opened various cross-linguistic paths, allowing for an effect of morphological awareness in one language on word reading in the other, for example.

Model A, the within-language model, was a poor fit for the data. The CFI and TLI statistics should be at least above 0.90, and for this model the CFI = .94 and the TLI = .86. The RMSEA and SRMR statistics should be at least below 0.10, and for this model the RMSEA = .22, and the SRMR = .13. The x2 statistic should be statistically insignificant for a well-fitting model (the higher the p-value the better) and in this model the x2(6, 100) = 35.163 p < .001. 

In Model B, we allowed for cross-linguistic paths, resulting in a much better fit. The CFI and TLI statistics are both 1.00, which is the best score of fit for these indices. The RMSEA is .00, which is also the best measure of fit for this statistic. The SRMR = .03, which is also notably close to 0. The p-value for the x2 statistic is no larger than .756, all reflecting a particularly well-fitting model. 

Model B revealed that within-language contribution patterns were similar in both languages—word reading partially mediated the association between morphological awareness and reading comprehension. In addition, we observed cross-linguistic influences (marked in blue) of Spanish morphological awareness to English word reading (β = .36, p < .001), of English word reading to Spanish reading comprehension (β = .22, p < .001), and finally, an influence of Spanish reading comprehension on English reading comprehension (β = .28, p < .01). 

Discussion

The focus of this study was to better understand how morphological awareness in a bilingual’s two languages contributes to their reading ability within and across languages. Our findings suggest that morphological awareness plays an important role in reading outcomes within language and additionally appears to transfer cross-linguistically from Spanish to English via word reading. 

In the cross-linguistic transfer model, analyses revealed similar within-language patterns of associations in English and Spanish. In both languages, morphological awareness had a direct influence on word reading, and word reading partially mediated the association between morphological awareness and reading comprehension within language. This finding is aligned

with prior successful models for English predicting that morphological awareness influences reading comprehension both indirectly through word reading, and directly through the language system itself (Deacon, Kieffer & Laroche, 2014). It stands partially in contrast to parallel research in Spanish, however, as D’Allessio, Jiachenco & Wilson (2019) found that morphological awareness only directly contributes to word reading in Spanish. Unlike our model, their path analyses revealed no significant mediating influence of word reading. 

One possible reason for this discrepancy has to do with the demographic differences between participant groups. D’Allessio, Jiachenco & Wilson’s participants were Spanish-speaking monolinguals, and our study tested Spanish-English bilinguals. In a language like Spanish that is already so transparent, we theorize that morphological awareness does not directly rely on word reading to aid in reading comprehension. However, because our bilingual participants have already had to establish this mediated path (morphological awareness to word reading to reading comprehension) in English, it is possible that, by extension, this tendency is also reflected in their Spanish. 

We also examined between-language paths to explore possible bilingual transfer effects. We observed cross-linguistic influences of Spanish morphological awareness to English word reading, of English word reading to Spanish reading comprehension, and finally, an influence of Spanish reading comprehension on English reading comprehension. Our model also suggests that Spanish morphology influences English word reading and that English word reading in turn contributes to Spanish reading comprehension. This cross-linguistic effect is consistent with previous research revealing a contribution of Spanish morphological awareness to English reading comprehension. Ramírez et al. (2013) found that awareness of Spanish derivational morphology indirectly contributed to English reading comprehension, specifically by way of

English cognate vocabulary and English morphological awareness. Cognate knowledge thus appears to be a bridge by which Spanish morphological awareness can transfer to aid in English reading comprehension. 

Finally, Spanish reading comprehension skill also contributes to English reading comprehension skill. We don’t observe this reciprocal relationship of transfer in the inverse direction (namely English MA to Spanish word reading to English reading comprehension). This may be due to a common tendency for skills in the more transparent language (i.e. Spanish) to benefit use of the more opaque language (i.e. English). On the whole, what this model does tell us is that morphological awareness contributes indirectly to reading comprehension within both languages and Spanish morphological awareness contributes cross-linguistically to English reading comprehension when mediated by English word reading. 

Impact Statement

In an increasingly globalized, shifting world, our young learners are exposed to more languages than likely ever before. The children of this generation carry all kinds of hyphenated identities and varying degrees of bilingualism. As such, it is our responsibility to attempt to understand how literacy development may differ in bilingual children and apply this insight to pedagogical settings where monolingual-informed theories for maximizing learning have always dominated. With our findings, we can also begin to validate the importance of maintaining heritage bilingualism, which is so valued by rapidly growing communities around the world.