Rethinking Language Proficiency: Commentary on Hulstijn (2024)

Abstract

We enjoyed reading Jan Hustijn’s update of his Basic Language Cognition (BLC) theory in the pages of Languages (Hulstijn, 2024), and we are honored to provide a commentary here. Researchers, language learners, and educators hold varying definitions and understandings of language proficiency. Innovative work that helps researchers in the field rethink proficiency is therefore of utmost importance. Hulstijn’s BLC theory offers a clear, well-motivated, and comprehensive framework for investigating what language proficiency may be, what ceilings there may be in what can be attained, and what causes there may be for attainment differences. Hulstijn tells us that BLC rests on a metatheory (i.e., a foundation of assumptions) of language and language learning grounded in two areas familiar to most scholars in second language acquisition (SLA): usage-based linguistics and complex systems. He offers BLC as a theory, that is, a tool to empirically investigate the causation and development of a phenomenon, in this case, proficiency, through a well-articulated research program. In this commentary, we highlight some aspects of the updated BLC theory that we find most useful. Following the Popperian spirit embraced by Hulstijn, we also introduce some critical probes subjects for refinements that may benefit BLC theory in the future.

1. Usage-Based Understanding of Nativeness and Non-Deficit Comparisons

We particularly appreciate that a ‘(non)native speaker’ is defined in BLC theory as having had (or not had) massive exposure to a given language from birth. This is in keeping with its usage-based foundations. Importantly, this definition allows BLC theory to ask the same set of questions about proficiency for native and non-native-speaking adults alike, and to investigate the same set of variables for both types of language users. Indeed, BLC theory makes predictions of a meaningful lack of difference in some areas (e.g., BLC attainment within native speakers; ELC attainment between native and non-native speakers) and meaningful differences in others (e.g., BLC attainment within non-native speakers depending on the amount and quality of exposure; ELC attainment for native and non-native speakers alike, depending on education and profession). In this regard, Hulstijn’s account of proficiency is well positioned to become an empirical tool to investigate Cook’s (2012) multi-competence perspective of second-language competence as non-deficient and worthy of study on its own.

Ideally, in the future, we would like to see more recognition in BLC theory that native speakers, too, can be bi/multilingual. Clearly, Hulstijn knows this well, as he rightly says “A person cannot simultaneously be the world’s best chess player and the world’s best violin player. But the literature on child bilingualism shows robust evidence that children can acquire oral control in at least two typologically different languages” (Hulstijn, 2024, p. 7). Yet, BLC, thus far, seems to treat child bilingualism as an interesting phenomenon that lies outside the scope of the theory. For example, ‘typological differences between L1 and L2’ is named as a plausible cause for individual differences within only non-native speakers for both BLC and ELC (in Predictions 3 and 4, p. 6), whereas this variable is left out of consideration for native speakers altogether (in Predictions 1 and 2, p. 5). However, crosslinguistic influence is relevant as a potential variable for native speakers as well, in a world where many adults grow up with bi/multilingual exposure from birth or at very young ages prior to entering school. The former is a situation of simultaneous bilingualism that is as much as three times more frequent than the latter, sequential child bilingualism, according to an estimation by De Houwer (2021, p. 4). The overall incidence of bilingualism in the world makes the bilingual condition expected, not a rarity. As for the world’s population, Grosjean (2024) provides a plausible estimatin that approximately 43% of the world’s population is bilingual.

Despite this bilingual population growth, we welcome Hulstijn’s formulation of nativeness as a mere function of language experience and history of exposure, not an inherent characteristic of language users. This is a healthy departure from imaging proficiency as a unitary construct that all native speakers possess and most non-native speakers lack. Simply put, the BLC theory of proficiency sends a strong anti-deficit message that proficiency is indeed attainable for non-native speakers, and that individual differences in at least some dimensions of proficiency (i.e., those related to ELC) are possible and systematic (that is, predictable and explainable) for native speakers as well. BLC theory makes it an empirical matter to gauge the extent to which native and non-native speakers vary individually in their proficiency. The theory leaves the door open to the possibility, for now, that the two types of users are not fundamentally different—an SLA debate that remains unresolved (e.g., Bley-Vroman, 2009; MacWhinney, 2022).

2. The Roles of Schooling and Literacy in Shaping and Structuring Proficiency

BLC theory proposes a dichotomous model, i.e., Basic Language Cognition (BLC) and Extended Language Cognition (ELC, formerly known as HLC, Higher Language Cognition in previous publications). This separation of proficiency into “basic” and “extended” realms is the most distinctive element of the theory, and it coincides with the separation of oral (naturalistic) and literate (schooled) language use. In the updated BLC theory, ELC is precisely defined “solely in terms of literacy skills” as “the control of the written standard language, as taught in school” (Hulstijn, 2024, p. 4). For native and non-native speakers alike, ELC will vary among individuals. Hulstijn offers causal variables such as differences in cognitive abilities (e.g., intelligence, reasoning, problem-solving skills, or world knowledge) and environmental factors such as variation in written language exposure, educational background, and professional or cultural profile. We think the central roles of schooling and literacy in BLC theory are strengths. Here, we offer some observations.

The roles of schooling in BLC theory are interesting. Schooling is predicted to iron out any pre-existing differences in BLC among native-speaking children (which will stem from differences in the quantity and quality of input in the home) and to gradually augment differences in ELC among native-speaking adults. It may also serve, it seems, to flatten out ELC differences between native and non-native users, removing any ceilings to what can be attained by non-native-speaking adults. It is difficult to separate schooling from literacy, since reading and writing are taught and cultivated in schools. Perhaps BLC theory assumes that the onset of literacy (when a person first learns to read and write) creates ELC, schooling further shapes it, and self-motivated literacy engagement over an adult’s life (i.e., well beyond schooling, for leisure, work, or both) continues to expand ELC proficiency. In relation to this, in passing, Husltijn made an interesting distinction (on p. 4) between illiterate adults, who have never learned to read and write, and aliterate adults, who function as basic readers and writers. We assume that in future investigations of BLC theory, both illiteracy (which is a categorical variable) and aliteracy (which is a gradient variable on a very wide continuum) may be important to investigate, as both can shed light on what exactly can be counted as ELC.

In his updated definition of ELC, Hulstijn introduces one issue that gave us pause: the stipulation that ELC requires control of not just written language, but also standard language. We understand that the key distinctions between BLC and ELC are the ability to use a high register or formal language, and that this is usually the standard language. This is very notable in many societies; for example, in mainland China, the ability to understand and use Standard Mandarin, as compared to the many dialects spoken in different regions, is expected of advanced and highly educated language users. But we wonder if invoking standard language in ELC is conceptually necessary or profitable. It clashes with mounting critiques against negative ideologies that penalize minoritized speakers of vernacular varieties as lacking in linguistic competence, simply because they do not use the variety arbitrarily chosen by the powerful as correct or appropriate. This is a practice experienced during schooling and painfully remembered by minoritized language users—for example, second-generation speakers of Turkish and Dutch in the Netherlands in Sevinç and Backus (2019). We invite further reflection and reconsideration of the necessity for the theory of emphasizing standard language as part of the definition of ELC.

We believe the theory is right in singling out ELC as an important dimension of proficiency and in linking ELC to levels of education and literacy. For one, we now have robust evidence that once phenomena beyond frequent, ordinary, basic, and simple language are probed, first-language (native) skills are subject to large individual differences. This challenges the idea that all native speakers converge to the same grammar, and by corollary, to the same proficiency. Ewa Dąbrowska is to be credited for bringing this finding to light in the mainstream of linguistics and SLA, starting with Dąbrowska and Street (2006). Since then, new studies have continued to corroborate their findings (e.g., in Turkish L1, Gedik, 2024). We also have research demonstrating that education and literacy make distinct contributions to levels of attainment in a second language as well, beyond the also important variable of differential exposure to language input (e.g., Andringa, 2014; Janko et al., 2019; Peyton & Young-Scholten, 2020; Tarone et al., 2009). If we put together these two strands of work, the inevitable conclusion is that differences in education and literacy shape language and create differences in knowledge, attainment, and proficiency for all users, regardless of their nativeness status. Any theory of proficiency should therefore account for these two variables.

3. Too Much of a Dichotomy?

Finally, we wonder whether the theory may have placed too stringent demands on the delineation of what separates oral (naturalistic) and literate (schooled) language use, and whether the conceptualization may be too dichotomous to be workable empirically. According to Biber (1995), the spoken–written modality is multidimensional. This multidimensionality presents a challenge to the clear-cut dichotomy of BLC and ELC. If we add digital communication, which today engages millions of language users for increasingly more hours each day—for leisure, language learning, schooling, and work (Demata et al., 2018)—the challenge of neatly separating spoken and written linguistic resources, and high-frequency and high-simplicity BLC, from sophisticated or complex ELC can seem even more daunting. More generally, it may be challenging to reach solid conclusions about what constitutes BLC in a speech community and where to draw distinctions between BLC and ELC. Shared and non-shared knowledge of native speakers varies across time, regions, individuals, and languages. Emerging trends, such as the frequent use of texting and posting on social media, growing reliance on AI, different language policies, and the length of obligatory education in regions using the same language, can all influence what constitutes BLC and where the distinctions should be between BLC and ELC.

But perhaps there is virtue in dichotomizing, as it makes the theory more widely accessible and more practically applicable. In our work on elicited imitation (Wu & Ortega, 2013; Wu et al., 2022, 2023), we found Hulstijn’s BLC theory useful in explaining why this method offers an effective shortcut to measuring L2 proficiency. Elicited imitation taps into speakers’ second-language BLC while eliminating any demands for literacy, i.e., without having to include ELC in the measure. In our design, a 30-sentence elicited imitation task holistically assesses a wide range of integrative and reconstructive language abilities in different BLC dimensions (e.g., implicit knowledge of phonetics, prosody, morphology, and syntax of the target language; automaticity in speech perception and production). There are now parallel instruments (all available on IRIS, iris-database.org), all translated based on the same English stimuli with minor adjustments to retain comparable length and plausibility across languages. We managed to offset the typological variation in degrees of complexity in morphology and syntax and features of phonetics and phonology in each target language, and the instrument versions effectively measure L2 proficiency in seven different languages. Based on the accumulated findings, a psycholinguistic tool like elicited imitation offers a modest but useful tool for the investigation of BLC theory and can help us explore how the age of onset; degree of exposure and productive use of the language; learners’ education background and intelligence; and crosslinguistic transfer affect the attainment of BLC and ELC.

Hulstijn’s BLC-ELC proficiency model also sheds light on proficiency frameworks widely adopted in language education. We have in mind, in particular, the ACTFL [American Council on the Teaching of Foreign Languages] (2024) in the United States, which specify five major levels of proficiency, including Novice, Intermediate, Advanced, Superior, and Distinguished, and the Common European Framework of Reference for Languages (CEFR) (Council of Europe, 2009) in Europe, which organizes language proficiency into six levels, from A1 to C2. These proficiency scales have a tremendous educational impact. They guide and inform learners about the specific language skills and communicative abilities required for each level of study, and they are often used to differentiate language users for purposes such as placing learners into specific language classes, recruiting for professional positions, and investigating or controlling for proficiency in research studies. The BLC-ELC model offers helpful theoretical explanations for the distribution of language users’ proficiency levels on these scales. Only a few language users fall into the highest few levels of language proficiency, regardless of whether they are native or non-native users, because those levels are heavy on ELC demands, which, according to BLC theory, vary considerably across native and non-native speakers and are modulated by level of education, control of written standard language, intelligence or academic and analytical predispositions, and so on (Hulstijn, 2024). For example, speakers rated at the level of Distinguished according to the ACTFL Proficiency Guidelines can “negotiate, use persuasive and hypothetical discourse, comment thoughtfully on a range of general topics, and tailor language to a variety of audiences”. Similarly, those regarded as proficient users at C2 of the CEFR can “summarize information from different spoken and written sources, reconstructing arguments and accounts in a coherent presentation” and can express themselves “spontaneously, very fluently and precisely, differentiating finer shades of meaning even in more complex situations”. The most advanced levels in both proficiency scales, therefore, require users to be equipped with BLC and, more importantly, to demonstrate ELC proficiency. In sum, BLC theory is an antidote to the pitfalls of using monolingual or highly educated speakers as standards or comparisons for second-language users and sets reasonable and attainable goals for language learning. The BLC-ELC proficiency construct should be an essential reference for language learners, educators, and researchers.

4. Future Work

We close with three observations worthy of consideration when thinking of future work on BLC theory. First, a good research practice in this research program will be to collect language exposure and use data for all study participants (native and non-native) on a gradient scale, for example, according to the suggestions put forth by Brown and Gullberg (2008, pp. 230–231). Second, a considerable task for scholars pursuing the research program outlined by BLC theory is to work out how “education” and “literacy” should be operationalized empirically moving forward, given that these two variables are implemented across studies in often poorly motivated and non-comparable ways. Third, as Husltijn rightly identifies, it will be imperative to conduct research with diverse participants representative of the full socioeconomic and educational spectrum of a given society or population of interest and to investigate languages other than English. We add here that the strategic choice of language for the study of BLC theory might help advance knowledge, particularly about ELC, in leaps and bounds. For example, literacy in Mandarin Chinese, whether L1 or L2, is notoriously different from literacy in English or other European languages (Wang et al., 2018), and these differences can help us understand what might or might not be included in ELC, particularly since there is a robust empirical SLA base from which to draw (Ke, 2018; Wu et al., 2024). In addition, languages like Arabic, traditionally classified as diglossic, can be particularly useful when investigating the contents and nature of ELC, because of the heteroglossic continuum they offer along the formal–informal and spoken–written axes (Saiegh-Haddad & Joshi, 2014; Saiegh-Haddad et al., 2022).

In the end, as Hulstijn argues, it is all an empirical matter. And BLC theory offers theoretically well-motivated and methodologically firmly grounded guidance for the interesting empirical work ahead that can help us rethink proficiency for native and non-native language users alike.