Abstract
Kidd and Garcia provide a critical analysis of linguistic representation in child language acquisition. Their findings expose a rich irony: a field of study for which diversity of experience is a basic assumption has remained non-diverse in its empirical and theoretical pursuits. In this commentary, I discuss deep-rooted structural factors that have contributed to Western dominance in child language acquisition research. Future progress relies on our ability to enhance the visibility of under-represented participants, researchers, and evaluators.
Kidd and Garcia (2022) provide a much-needed chronicle of changing representation in language acquisition research. Their findings are both surprising and sobering. In exposing a long-standing pattern of sampling bias in language acquisition research, the authors articulate the clear scientific costs of narrow representation. Kidd and Garcia’s findings closely mirror broader sampling biases famously described by Henrich et al. (2010). Although there is currently vibrant discourse around these issues within the academic community, actual progress remains slow both within the field of language acquisition (Kidd & Garcia, 2022) and within the developmental sciences more generally (Nielsen et al., 2017; Singh et al., 2022). Kidd and Garcia’s (in press) very important work raises critical questions about the underlying source of skewed representation, which is crucial to identifying the levers of change. In this commentary, I discuss two structural barriers to diversification: intersectional visibility and the role of power and privilege. Identifying and dismantling major structural barriers is critical to redressing the current state of affairs and advancing towards a diversified science of language acquisition.
In comparing representation across languages, Kidd and Garcia report a persistent skew towards English, followed by French, Spanish, German, Dutch, and Italian, each of which outrank any language from the non-Western world. This does not just reflect linguistic imbalance, but also reflects sociocultural imbalance. Languages that predominate in Kidd and Garcia’s analysis are not the most widely spoken languages, but instead are widely spoken in sociocultural contexts that prevail in psychological research and are therefore highly visible. When viewed from an intersectional perspective, visibility in psychological research is co-determined by overlapping factors that can include race, culture, national origin, social position, and other factors (Cole, 2009; Syed et al., 2018). On account of intersectionality, some groups are rendered invisible in the research landscape. As noted by Kidd and Garcia, the scientific cost of intersectional invisibility is an incomplete narrative of language acquisition.
However, an additional damaging consequence of invisibility is that scholars may make false attributions to both under- and over-represented groups (Causadias et al., 2018). As described by Causadias et al. (2018), the behaviours of well-represented groups are often thought to reflect basic, acultural aspects of development. Kidd and Garcia provide several examples of findings in language acquisition that were deemed to be fundamental and universal based on data from a limited set of Western languages. In contrast, the behaviours of under-represented groups are often invoked as evidence for sociocultural variation rather than for fundamental processes (see also Kline et al., 2018). This elevates the position of widely studied populations within the research landscape to the foreground, while relegating studies with under-represented groups to the margins. This imbalance is at odds with the breadth of evidence needed for sound theory building in our field. As noted by Kidd and Garcia, ‘A comprehensive theory of any language phenomenon requires a representative sample that provides a solid foundation for theory building, and an even bigger sample for theory testing’.
Cultural misattribution in psychological research (Causadias et al., 2018) can permeate the review process. Scholars of language acquisition, working on under-represented languages, often report alarming anecdotes around the peer review process. When evaluating research from under-represented groups, reviewers may ask for data from a ‘control group’ from a more widely represented population to legitimize data from an under-represented sample. More fundamentally, they may ask for reasons why the under-represented population being studied is even relevant or worthy of a standalone contribution. These are standards to which scholars of English and other widely studied languages are typically not held, conferring a level of privilege upon these groups that is not granted to studies with under-represented populations. For these reasons, increasing the diversity of reviewers as well as of reviewers’ perspectives is critical to diversification. For example, the current Editor and Associate Editors for the four journals sampled by Kidd and Garcia are entirely drawn from the Global North. Only 6% of these scholars are based in Asia, 37% are based in the United Kingdom/European Union and 57% are based in North America. This introduces significant power imbalances, where those deciding on whether studies are published are concentrated within the same world regions from where participants most frequently originate. In this way, limited participant diversity operates within a larger ecosystem of limited evaluator diversity. Arguably, research infrastructure is often better established in the Global North. However, continuing to situate research evaluation in these contexts results in a cumulative privileging of scholars from these regions and only perpetuates imbalances in representation and visibility. Kidd and Garcia’s efforts to increase inclusiveness (e.g. the ‘Truly Global /L+/ Summer/Winter School on Language Acquisition’) represent much-needed efforts within the research community. In addition, larger disciplinary interactions need to take place to engage in the uncomfortable work of acknowledging intersectional invisibility and to initiate the hard work of correcting course.
In addition to invisibility, addressing Western centrality in language acquisition research requires us to examine power structures that exist within our discipline. The literature on decolonizing frameworks in psychological research provides valuable context for this type of self-examination (see Fernández et al., 2021). Decolonizing the research landscape involves a process of active resistance to long-standing norms and practices. New and equitable epistemic traditions in language acquisition research will only arise from a thorough self-critique of current and past practices that have so long prevailed in our field. To guide this process, Syed et al. (2018) proposed three questions that are of central importance. I address each in turn. First, from whose vantage point is research conducted? Research in language acquisition has traditionally been developed through a Western lens for a Western readership. This has limited our use of tools and methods to those developed by and for Western populations. Most often, these methods involve the quantitative study of carefully operationalized variables within controlled environments. Field studies remain relatively rare, but as Kidd and Garcia note, these studies may provide deeper insight into some under-represented languages. This requires us not only to increase the diversity of languages being studied but also to decenter research practices to reflect methodologies that befit the sociocultural context within which participants are sampled.
Second, what types of questions are valued? As Syed et al. note, questions about majoritized groups are presumed to be relevant and valuable by default. Ideas that take hold are ideas that resonate with majoritized groups. This provides no assurance that these ideas are scientifically more valuable than those applied to under-represented groups. A greater diversity of individuals and perspectives among those who evaluate research (e.g. editors and reviewers) are likely to democratize our conceptualizations of scientific value.
Finally, who gets left out? As Kidd and Garcia note, those who get left out are those from developing countries, largely from the Global South. However, change on this front is likely not within the purview of a single researcher. The reward structures of academic research generally do not incentivize breadth of representation over scientific depth. Diverse representation is often viewed as a useful adjunct to the science, rather than intrinsic to the science. I suggest that three groups are key agents of change in incentivizing diversity: funding agencies, professional organizations and journal editors. In different ways, these three groups are in a unique position to grant greater visibility to those who are left out of our scientific narrative and to reward collaborative and inclusive efforts to broaden participation in language acquisition research.
On a final note, our field will not be enriched by theoretically unguided, symbolic efforts to study under-represented languages. Instead, true progress requires first, that we identify the specific scientific opportunities of studying a particular language and second, that we generalize faithfully from the language groups being sampled. In a recent paper, Christiansen et al. (2022) articulate distinct scientific gains associated with researching widely studied languages (which are generally well documented), comparing widely studied and under-represented languages (to examine overarching theories of language acquisition) and within-language comparisons (to investigate the effects of sociocultural context on acquisition). This framework is a useful guide as we reform research practices in language acquisition.
