The Ontogenesis Model (OM) of the L2 lexical representation usefully brings together disparate strands of research on the development and representation of second language (L2) words in a single coherent framework. We discuss the potential utility of the framework, challenge some of the authors’ current assumptions, and suggest areas where further elaboration and research will enhance the OM's impact.

Central to the OM is the notion of “fuzziness” in lexical representations, which Bordag, Gor, and Opitz (Reference Bordag, Gor and Opitz2021) describe as “inexact or ambiguous”. It is necessary to determine the situations where this term is useful from those where it obscures crucial differences among types of learning problems. For example, “fuzziness” has been used to describe the phono-lexical representation of novel phonological categories that are effectively ambiguous due to overlap with familiar (native) categories. In these cases, the learner maintains a phono-lexical contrast via the relative fuzziness of the new and familiar phonemes, resulting in asymmetric lexical activation because fuzzier representations are contacted by a wider range of acoustic events than are less fuzzy representations (as in scenarios reported by e.g., Barrios & Hayes-Harb, Reference Barrios and Hayes-Harb2021; Darcy, Daidone & Kojima, Reference Darcy, Daidone and Kojima2013; Escudero, Hayes-Harb & Mitterer, Reference Escudero, Hayes-Harb and Mitterer2008; Weber & Cutler, Reference Weber and Cutler2004). In other cases, learners’ phono-lexical representations are not inexact or ambiguous, but rather neutralize a novel contrast to the familiar category. In yet other cases, learners may map an L2 category to multiple L1 categories, resulting in spurious L2 lexical representations (e.g., Elvin, Williams & Escudero, Reference Elvin, Williams, Escudero, Molsing, Perna and Ibaños2020; Escudero, Smit & Reference Escudero, Smit and MulakMulak, in preparation; van Leussen & Escudero, Reference Van Leussen and Escudero2015). These and additional possible scenarios are not clearly differentiated in the OM, though they may differ in origin, nature, learning problem, and solution for development.

The OM aims to describe ontogenetic development, but does not presently endorse specific theories of the genesis (initial state) of learners’ representations, the role that input plays in development, or how learning proceeds, and as such does not produce predictions regarding expected developmental paths or the influence of differential input on these paths. It is further unclear what role perception plays in mappings from acoustic, written, or semantic input to representations. Future iterations of the OM should incorporate enlightening proposals regarding the initial state, mappings between inputs and representations, and learning mechanisms (e.g., Second Language Linguistic Perception, L2LP; Elvin & Escudero, Reference Elvin, Escudero, Gutierrez-Mangado, Martínez-Adrián and Gallardo-del-Puerto2019; Elvin et al., Reference Elvin, Williams, Escudero, Molsing, Perna and Ibaños2020; van Leussen & Escudero, Reference Van Leussen and Escudero2015).

Crucially, the ways that initial representations relate to optimal L2 representations (e.g., ambiguous/partially overlapping, neutralized/fully overlapping, or extra/spurious, etc.) may impact the ways in which acoustic input can influence development. A theory of the nature of these representations is crucial for an explanatorily adequate account: in what way are these types of information stored as part of L2 lexical representations (abstractly, e.g., Cutler, Reference Cutler2015, or in an exemplar-based manner)? As well, the OM captures the interconnectedness of phonological, orthographic and semantic representations, but does not provide an account of how this interconnectedness facilitates, impedes, or otherwise causes change in L2 lexical representations. In other words, how does the OM formalise the role of the visual information, auditory information, or knowledge of the world that a learner accumulates? A model that aims at explaining L2 lexical development must provide for hypotheses concerning the relative importance of the relevant domains, as is done with the relevant phonetic dimensions for phonological representations in previous proposals (e.g., the L2LP model). For example, Cutler (Reference Cutler2015) argues that while semantic or orthographic input may provide clues to the nature of optimal L2 representations, without the ability to perceive an L2 lexical contrast, added competition and processing delay in word recognition may ultimately prevent L2 development.

As a framework, the OM offers value by unifying multiple domains of lexical representations as well as multiple traditions of research, and importantly provides a way for researchers in the relevant fields to think about and to characterize the independent and interconnected properties of phonological, orthographic, and semantic representations in the L2 lexicon. It does not yet provide for testable predictions or explanations, due in part to the absence of theories of the initial state, how input is integrated into representations, mechanisms for learning, or how the domains of lexical representations interact and resolve conflicting information. As the OM evolves towards predictive power, we expect it will additionally serve to inform methodological choices and innovations. Models can help to clarify what tasks are likely to be most informative in selecting among competing hypotheses, and can also provide explanations that inform language teaching (e.g., Elvin & Escudero, Reference Elvin, Escudero, Gutierrez-Mangado, Martínez-Adrián and Gallardo-del-Puerto2019). We look forward to seeing how the authors and others integrate proposals of other more explicit models of L2 perceptual and lexical development in their path to explaining L2 lexical representations.