Taking as a premise that phonological working memory (PWM) influences later language development, in their keynote article, Pierce, Genesee, Delcenserie, and Morgan aim to specify the relations between early language input and the development of PWM in terms of separable influences of timing, quantity, and quality of early language input. We concur that prior work has established that PWM and language development have reciprocal influences on one another during development (e.g., Baddeley, Gathercole, & Papagno, Reference Baddeley, Gathercole and Papagno1998; Gathercole, Reference Gathercole2006; Gathercole, Hitch, Service, & Martin, Reference Gathercole, Hitch, Service and Martin1997; Metsala & Chisolm, Reference Metsala and Chisholm2010). The goal of the keynote article was to describe how early language experience may influence the development of PWM. Pierce et al. argue that this can be done by comparing the development of PWM across groups of children with differing language experiences during early childhood, specifically (a) delayed exposure to language, (b) impoverished language input, or (c) enriched language input. The authors suggest that this comparison may contribute to establishing that individual differences in PWM are due, in part, to early language experience. Sensitive periods for phonological development that are open roughly in the first year of life are discussed, and it is suggested that the quantity and quality of early language input shapes the quality of phonological representations. Efforts to specify mechanisms by which early language input may influence the development of PWM have both theoretical and, potentially, clinical importance. Considering this, Pierce et al.’s article, which aims to create a platform for future research in terms of the timing, quantity, and quality of early language input, is a valuable contribution.

We were especially interested in the authors’ exploration of the role of early language experience on PWM and language outcomes. We are reminded of work in our laboratory, which has used “process” and “product” measures of speech and language performance. Process measures have included tasks that tap PWM, such as memory span for sequences of spoken digits, nonword repetition, and verbal rehearsal speed. We have used these process measures in order to better understand “product” measures, such as speech perception and expressive and receptive language outcomes, in deaf children who use cochlear implants (CIs; e.g., Pisoni, Cleary, Geers, & Tobey, Reference Pisoni, Cleary, Geers and Tobey1999). Our findings have been consistent with the principle that phonological processing, which includes the rapid phonological encoding and storage of incoming speech signals, is crucial to spoken language development (Baddeley et al., Reference Baddeley, Gathercole and Papagno1998; Gathercole, Reference Gathercole2006). Further, our findings support the generalization of this principle to the population of deaf children who receive CIs. It appears that “process” measures that tap PWM are related to and strongly predict speech and language development in children with CIs (e.g., Casserly & Pisoni, Reference Casserly and Pisoni2013; Kronenberger, Colson, Henning, & Pisoni, Reference Kronenberger, Colson, Henning and Pisoni2014; Pisoni & Cleary, Reference Pisoni and Cleary2003; Pisoni & Geers, Reference Pisoni and Geers2000). Like Pierce et al., we believe that understanding the ways in which the development of PWM may be influenced by early language experience is important to the field. In their keynote article, Pierce et al. present evidence that the phonological representations that are acquired during sensitive periods of phonological development in early childhood are underspecified when language input is delayed or impoverished, and suggest that less accurate encoding of phonological information may have the consequence of reducing the effectiveness of PWM.

In discussing underspecification in CI users, we believe it is important to be clear that CI users never receive undegraded auditory input, regardless of their amount of experience with the implant. Compared to a normal-hearing ear, a CI drastically lowers the resolution of information carried by the auditory nerve. The degradation of speech input through a CI has specific consequences in terms of what acoustic–phonetic cues can be perceived from the signal. For example, it is very difficult for CI users to discriminate between voices, in large part due to the lack of temporal fine structure, which contributes to fundamental frequency or voice pitch perception in listeners with normal hearing (Chatterjee, & Peng, Reference Chatterjee and Peng2008). With respect to phonetic features, the acoustic–phonetic cues received from a CI are underspecified in a way that appears to affect certain phonetic features more than others. For example, most CI users are able to discriminate minimal pairs of phonemes that differ only on a phonetic feature of the manner of articulation (e.g., fricatives vs. stops) with a high level of accuracy after a few years of experience with their implants, but are less accurate for minimal pairs that differ on a place of articulation feature (Pisoni et al., Reference Pisoni, Cleary, Geers and Tobey1999). CIs drastically reduce spectral information, which conveys most of the acoustic cues for place of articulation, but preserve the temporal envelope of the speech signal, which contains amplitude fluctuations that can convey manner of articulation cues (Shannon, Zeng, Kamath, Wygonski, & Ekelid, Reference Shannon, Zeng, Kamath, Wygonski and Ekelid1995; Wright, Frisch, & Pisoni, Reference Wright, Frisch, Pisoni and Webster1999). Thus, for CI users, underspecification is enforced by degradation of the auditory signal, such that acquisition of typical phonological representations cannot be expected, regardless of whether the CI is operational during early sensitive periods. Moreover, the degraded auditory signal that is conveyed by a CI affects some phonological representations more than others.

Although deaf children implanted with CIs experience both delayed and degraded language input, Pierce et al. classified prelingual CI users as experiencing delayed exposure to language but not degraded exposure to language. Although not explicitly stated, the authors may have used this classification in consideration of sensitive periods in the first year of life for phonemic category acquisition (Kuhl, Reference Kuhl2004). Deaf children have not yet received their CIs at this early sensitive period for language, and as Pierce et al. correctly point out, their exposure to spoken language is thus necessarily delayed, even when current clinical guidelines that allow for implantation before the first birthday are followed (Dettman, Pinder, Briggs, Dowell, & Leigh, Reference Dettman, Pinder, Briggs, Dowell and Leigh2007). CI researchers have previously noted the potential influence of missed sensitive periods for phonemic category acquisition and sound pattern learning (Kuhl, Reference Kuhl2004; Saffran, Aslin, & Newport, Reference Saffran, Aslin and Newport1996) on language development in CI children (Niparko et al., Reference Niparko, Tobey, Thal, Eisenberg, Wang and Quittner2010). Because the perceptual specialization for the sound patterns of a native language that occurs in early childhood alters the neural representation of speech sounds, missing these sensitive periods likely results in suboptimal neural connectivity for spoken language processing (Kral, Kronenberger, Pisoni, & O'Donoghue, Reference Kral, Kronenberger, Pisoni and O'Donoghue2016). This may partially explain why age of implantation has a strong impact on individual difference in CI outcomes, and is consistent with the idea that sensitive periods for phonological input are important for specializing perception for the phonological features that are available from the signal that is transmitted to the auditory nerve by a CI. Pierce et al. correctly note at various points in their discussion that prelingual CI users experience degraded language input, although this is not reflected in their classification of CI users as delayed in language input but not as experiencing impoverished language input. Although we would have classified CI users differently than the authors, the classification may be superfluous to the general framework presented by Pierce et al. given that their analysis finds that delayed and degraded input have the same overall effect, namely, that of decreasing PWM and language abilities. For both delayed and impoverished language input, it is hypothesized that this occurs via the acquisition of underspecified representations during sensitive periods of phonological development.

In sum, Pierce et al. have made progress by setting up a broad framework for understanding the effects of the quantity and quality of early language input on phonological representations, particularly in terms of how underspecified phonological representations may affect working memory and, ultimately, language development. The authors aimed to provide a general framework for understanding the influence of early language experience on PWM. In our commentary, we note that this approach has limitations. For example, there is a lack of clarity in the terms underspecified, weak, and less accurate used to describe the phonological representations of children who receive delayed and/or degraded speech input early in life. Although the lack of clarity may be essential given the broad scope of this keynote article, we note that the type of signal degradation that is present for various populations, such as CI users, children who experience early bouts of otitis media, and signing deaf children born to nonsigning hearing parents, is likely to be different, and consequently to create different types of underspecification. We are hopeful that given the framework established by Pierce et al., future work may compare the ways in which phonological representations are underspecified for such populations. It is our view that this could potentially reveal new research directions in terms of understanding the relation(s) of phonological representations, working memory, and language outcomes across these populations.