On “Local Theory” Neutrality with Respect to “Meta-Theories” and Data from a Diversity of “Native Speakers”, Including Heritage Speaker Bilinguals: Commentary on Hulstijn (2024)

Abstract

This commentary critically engages with Hulstijn’s revised Basic Language Cognition (BLC) Theory, which aims to enhance explanatory power and falsifiability regarding individual differences (IDs) in language proficiency across native and non-native speakers. While commending BLC Theory’s emphasis on separating oral and written language cognition, we raise two key concerns. First, we question the theory’s exclusive alignment with usage-based approaches, arguing that its core constructs are, in principle, compatible with multiple meta-theoretical frameworks, including generative ones. As such, BLC Theory should remain neutral to maximize its cross-paradigmatic utility. Second, we address the theory’s treatment of heritage speaker bilinguals (HSs), particularly the implication that they may not typically acquire BLC. We contend that this position overlooks robust empirical evidence demonstrating that HSs develop systematic, rule-governed grammars influenced by their individual input and usage conditions. Moreover, we highlight how IDs among HSs can provide a valuable testing ground for BLC Theory, particularly regarding the role of input and literacy. We conclude that embracing theory neutrality and integrating diverse speaker data—especially from heritage bilinguals—can enhance BLC Theory’s generalizability, empirical relevance, and theoretical utility across language acquisition research.

In an effort to increase explanatory power and, crucially, empirical falsification potential, Hulstijn (2024) offers some refinements and clarifications to his Basic Language Cognition (BLC) Theory. At its core, BLC Theory provides a conceptual framework for investigating, predicting and explicating the spectrum of well-observed individual differences (IDs) in language proficiency in both native and non-native speakers, which cannot, we would agree, be neatly “mapped on a single proficiency scale” (p. 1). As the title of our commentary suggests, the points of discussion we raise are, by design, specific in scope. In other words, we do not delve into all the finer details Hulstijn (2024) articulates, notwithstanding that several others in addition to those we hone in on are equally worthy of serious independent comment, if not praise or deeper questioning.

Before getting into the main points of discussion, it is fitting to start with a synthesis of the updated BLC Theory as we understand it. To avoid circularity in argumentation and/or reduce the risk of (inadvertently) speaking beyond (as opposed to) each other, summarizing the main object of discussion is a critical step in any commentary. Doing so can bring to light from the outset any potential points of difference between what (we as) readers have understood and what the author intended to convey.

Basic Language Cognition (BLC) is part of the BLC-ELC framework (Hulstijn, 2011, 2015, 2019), which highlights a distinction between shared and non-shared language cognition. The construct of BLC itself refers to the implicitly acquired ability to understand and produce spoken language in everyday situations. As such, BLC is shared by all adult (native) speakers (of a given language); it is “characteristic of the human species” (p. 3). In contrast, Extended Language Cognition (ELC) is a unique component not necessarily shared entirely by any two speakers (natives or non-natives, in any combination thereof, alike). ELC is the complement/extension of BLC in the written domain or “the control of the written standard language, as taught in school” (p. 2, italicized in the original). This general approach further distinguishes between core linguistic cognition (CLC) and peripheral linguistic cognition (PLC). While overlapping with BLC and ELC, CLC and PLC do not have a one-to-one mapping relationship with them. CLC includes proficiency at all levels of grammar (lexis, phonetic–phonological, morpho-phonological, and morpho-syntactic), inclusive of how to use these grammatical forms in a semantically and pragmatically harmonious way to the communicative situation. PLC includes three important components: (i) the general ability for communicative interaction in monolingual and multilingual situations; (ii) strategic competence for linguistic communication performance under suboptimal conditions and (iii) conscious or unconscious metalinguistic knowledge for spoken and written language alike. While BLC maps neatly onto CLC, both BLC and ELC are relevant for PLC.

In this latest paper, Hulstijn (2024) argues that while cognitive factors (e.g., executive functions and general intelligence) have the potential to impact the acquisition of reading and writing skills irrespective of nativeness (PLC and, crucially, the ELC contribution to it), they do not play a (deterministic) role in the acquisition of (core) speech processing in either speaker type, i.e., they do not bring anything to bear on BLC/CLC. It is with this in mind that Hulstijn maintains that IDs in language proficiency cannot be meaningfully mapped to a single proficiency scale. Rather, IDs need to be contextualized and separated by two dimensions that only partially overlap: (i) the cognition of oral language (receptive and productive speech processing) and (ii) the cognition of written language (reading and writing). Given the refinements and clarifications offered, Hulstijn (2024) articulates several domains for empirical testing, that is, points for falsification that include predictions for when one should and should not expect the occurrence of IDs in the oral and written domains of language, for both native and non-native speakers.

There is much to be applauded with respect to the BLC Theory, not least its steadfast argumentation that IDs in oral and written domains of language should have a privileged focus in language acquisition/processing theoretical epistemology, as opposed to being deemphasized, ignored or even consciously treated as less interesting noise. We agree (i) that IDs should be a core focus of language research; (ii) that mapping them to a simplistic proficiency scale is futile; (iii) that they are not strict analogs of each other; (iv) that when properly investigated, they reveal systematicity and predictability; and (v) that providing falsifiable theorizing as regards (iv) and testing it can provide fundamental insights into the mechanisms and nature of language and cognition more generally. And yet, there are several points to Hulstijn’s positions as we understand them that would benefit from serious discussion.

Before narrowing the focus of our commentary to issues pertaining to BLC with respect to heritage language bilingualism (HLB), we ask an important question at a more general, higher theoretical level: Why is/must the BLC be predicated on a particular—reasonable as it might be—understanding of how language is acquired/learned? In other words, why is BLC Theory framed exclusively from a usage-based perspective of language acquisition/cognition?

Such a query is non-trivial on several grounds. In addition to our argumentation that doing so is both stipulative and, thus, unnecessarily limiting, it will have important consequences in our discussion related to HLB and BLC Theory in what follows downstream. Suffice it to say presently, we wish to make it clear that querying Hulstijn’s choice of alignment is not meant to question nor deny him any agency. It is, of course, his prerogative to apply the BLC as he sees fit in his own empirical practice. Clearly, he believes and might well be correct that a usage-based approach, in general, is on the best explanatory track. Our point is that, at least in our iterative readings of BLC Theory over time, inclusive of this 2024 refinement, there is/continues to be nothing that would make its application incommensurable with/adoptable by theories proposing (degrees of) linguistic domain-specificity. This is precisely because the BLC-ELC is a “local theory”—as Hulstijn contends—that astutely highlights factors that sit in the common ground for all researchers interested in multilingual language acquisition, irrespective of paradigm. As a “local theory”, BLC Theory is compatible with multiple “meta-theories”, and since BLC Theory does not make predictions that would enable us to adjudicate between competing meta-theoretical accounts of language, emphasizing one over the other is unnecessary. While Hulstijn’s steadfast alignment of BLC Theory to usage-based doctrine accords with his personal approach, it is stipulative to the extent that none of the BLC’s important highlights are, we believe, incompatible with core components of other meta-theoretical approaches, e.g., generative bilingual language acquisition theories, especially in their current formulations (see Rothman & Slabakova, 2018; Slabakova et al., 2020). While Hulstijn and his personal empirical application of the BLC need not be, BLC Theory itself should be agnostic on this point, rendering it more universally applicable.

Frameworks/theories/models are not a reflection of the people who propose them, but rather the articulation of a set of then current ideas offered by them. Once offered in the public domain, the frameworks/theories/models exist as a communal tool independent of their authors. While the original authors retain some privileges with some level of authority (e.g., iteratively revising or “updating” their frameworks/theories/models), this does not equate to absolute ownership that would be permissive of unnecessary stipulation. Make no mistake, frameworks/theories/models are fraught with stipulations. However, to meet the ultimate criterion of parsimony—accord with Occam’s razor—stipulations must be a necessary part of the theory itself. We do not see how stipulating a usage-based understanding of language constitutes a necessary stipulation, but rather submit that it is peripheral and thus ratcheted on top.

Let us consider an example of our line of reasoning from within our own group. While Rothman’s body of empirical work in which he applies his Typological Primacy Model (TPM) (Rothman, 2011, 2015) is arguably generative in nature for additional reasons, the TPM itself is definitively not a generative theory (see Rothman et al., 2019, for discussion). A careful reading of the epistemological writings related to the model itself will show that it has always been presented as paradigm impartial. This is precisely because the TPM is a framework that deals with a theory-neutral topic, one of great relevance to all who study non-native language acquisition/processing: cross-linguistic influence/transfer at the initial stages of additive multilingual acquisition. The fact that the TPM is authored by an assumed generative scholar and/or irrespective of what he might specify in his empirical application—e.g., what he assumes is the mental constitution of transfer itself within his specific empirical studies—does not make the overarching framework of the TPM a generative theory per se. To have stipulated it must be, would have been to detract from its higher application as a bona fide framework. As such, it can and has been used by scholars with otherwise seemingly irreconcilable differences in paradigmatic assumptions and ideologies.

Thus, we simply ask: why should BLC Theory exclude itself by inessential design from being used by or useful as a guiding force to a subset of theories when what it independently adds or focuses attention on is not in actual opposition to them? Doing so unnecessarily limits the coverage and explanatory power of BLC Theory, if not its potential utility as a bridge between competing cognitive theories of language learning in the space where they can and do agree to agree. Indeed, there is good reason to label specific theories/models/frameworks as exclusively usage-based or generative when they, in part or in whole, are predicated on tenets that only apply to one or the other. In our view, what the BLC Theory highlights is simply not such a case in point. If there are reasons that in fact make BLC Theory only commensurable with usage-based frameworks beyond preferred stipulation, which we do not preclude, we look forward to Hulstijn’s explanation of what we have missed or misunderstood in this regard. In any case, the larger field could benefit from more frameworks that are theory-neutral where possible such that the same guiding principles could be applied by researchers with distinct paradigmatic persuasions in pursuit of answering universal questions where theory-specific differences are less important. The BLC Theory could be a significant tool, then, in such a light.

In the remainder of the commentary, we draw focus to some questions pertaining to where facts regarding HLB sit with respect to tenets of the BLC Theory. We wish to acknowledge from the outset footnotes 3 and 4 in Hulstijn (2024, p. 9):

Footnote 3: Thus, so-called heritage speakers (Rothman, 2009) can be considered native speakers (Rothman & Treffers-Daller, 2014). BLC Theory is targeted at adult native speakers who have acquired BLC.

Footnote 4: In response to a clarification question of one reviewer, I would like to clarify that native speaker and BLC are orthogonal constructs. Typical adult native speakers attain BLC but heritage native speakers may not do so. Learners of non-native languages may or may not attain BLC, depending on a number of factors.

It is accurate, in our view, that the starting point of departure of Hulstijn is one where heritage speaker (HS) bilinguals are accepted as native speakers of their heritage language, as seemingly advocated in footnote 3. One could find some internal contradiction, however, by Hulstijn given the wording of footnote 4. If “typical adult native speakers” attain BLC, then HSs who “may or may not” are not being conceived of as equal in their nativeness. We surmise Hulstijn is simply using “typical native speaker” as synonymous with L1-dominant speaker, a position that confuses/confounds resulting dominance with nativeness (see Rothman et al., 2023, for discussion). Of more importance is to question what is meant by “adult native speakers who have acquired BLC”. One reasonable reading, especially in light of what is written in footnote 4, is that adult HS bilinguals are assumed to not “typically” acquire BLC. Therefore, BLC Theory is simply disinterested or inapplicable to HLB. Really? If this is the case, the slope towards losing coverage can be rather slippery. To start, this begs the question of how one determines what BLC is in any particular context; that is, what is the resulting outcome of BLC in any specific context that renders it “acquired”? Reading between the lines, it seems that if HSs—or at least some—do not acquire BLC, then gauging that failure relies on reported differences in HS linguistic competence in domains of grammar that contribute to BLC assessment to something else where BLC would be unquestionably acquired. Our best guess here would be that the relevant comparison is to an idealized L1-dominant user, the referenced “typical native speakers”. If our understanding is on the right track, would that not also entail that some L1-dominant users have not acquired BLC, those that also differ in interesting ways from so-called standard baselines in domains that contribute to BLC (Dąbrowska, 2012)? Thus, would BLC Theory not fail to apply to them as well?

We refer the reader to Ortega (2019), De Houwer (2023) and Rothman et al. (2023), amongst others, for a more complete discussion of the many reasons why standardized monolingual baseline assumptions are problematic. Suffice it to say for now that it is overly simplistic, if not outright spurious, to dismiss a consideration of HLB data outright as it pertains to assessing the BLC Theory and its constructs. It is more or less accepted presently that HS grammatical competence outcomes, distinct as they often are from other sets of native speakers (Montrul, 2016, 2022; Polinsky, 2018), are universally compliant, rule-governed, systematic grammars whose individual variation is neither random nor evidence of incompleteness, but rather correlate with experience with the HL in terms of input quality and quantity as well as opportunity for use over time (Kupisch & Rothman, 2018; Paradis, 2023; Kubota et al., 2025). In this light, HSs most certainly attain BLC in their HL or the very notion of BLC is, in our view, meaningless. In fact, suggesting that they do not seem to us to run in disaccord with the tenets of a usage-based perspective that Hulstijn advocates in the first place and the facts of HLB studies overall, unless he is suggesting that some HSs simply do not receive enough input to acquire BLC. HLB under this logic would constitute an ideal context for testing the “exposure cost” for BLC that Hulstijn (2024) refers to in Section 5.2. However, while it might be true that some individual HSs lack sufficient input to acquire BLC—if existent at all, we surmise this would be a small minority of HSs—this is not representative of the common HS’s experience with the HL and certainly not indicative of the data that populate the majority of the published literature.

It is also interesting to point out that much of the recent literature that has sought to explain IDs in HLB outcomes is highly compatible—although not exclusively so—with a usage-based perspective showing that regressors reducible to exposure, usage opportunity and education/literacy training are highly predictive of constrained inter-group and intra-individual HS variation in areas that should fall under the remit of BLC. This same evidence, thus, seems to suggest that BLC and ELC are not as separable as the BLC Theory would contend. It is true that for some HSs, those individuals with only oracy in their HL (a minority of HSs, to be sure), ELC is simply irrelevant. And yet, increasing evidence showing that instruction/degree of literacy in the HL correlates with closer aligning grammatical outcomes with standard L1-dominant users (e.g., Kupisch & Rothman, 2018; Bayram et al., 2019; Hao et al., 2024; Kubota et al., 2025)—those HSs, we might assume, would be considered to have acquired BLC following the above logic—seems to suggest that there is a relationship between reading and writing skills (or at least training therein) and areas of grammar that would fall under the domain of BLC as described. If such is true of HS bilinguals, then we might assume this is also true of L1-dominant users.

With the above points in mind, we look forward to hearing more from our colleague on the issues we have raised. To be sure, there is promise and usefulness to the revised BLC Theory. In our view, the BLC-ELC framework and thus the larger field will equally benefit as the issues we have raised are engaged. We have always found exchanges with Jan Hulstijn to be highly productive in the past and are excited in anticipation of his reply to what we have respectfully drawn attention to herein.