2. Something old, something new, something to look forward to

HSs represent a big proportion of the bilingual population, but linguistic research has so far focused on HLs that are part of the Indo-European family (e.g., Spanish and German) and on bilingual populations based in North America and Europe. The volume addresses these gaps by focusing on lesser-studied language varieties spoken in other regions. These varieties, in turn, become novel testing grounds for examining contact-induced variation and change. The wider geographical scope is made more apparent by Rao’s intentional geographical organisation of the volume (e.g., chapters 1 and 2 concern HLs in Asia, while chapters 3 and 4 explore HLs in Australia and New Zealand, respectively) – this thoughtful approach demonstrates that HLs are found in many parts of the world, reminding readers that multilingualism is the norm in most societies.

One of the many strengths of this volume lies in the studies’ analytical depth. Most studies adopt systematic comparisons of carefully constructed, robust data sets from multiple speaker groups. For example, in chapter 1 Kang, Han, Ryu, Schertz & Yun investigate phonetic realisations of Korean /y/ and /ø/ based on data from two HL varieties in Northern China (Dandong and Hunchun), two homeland varieties (Seoul and Northern Hamgyeong) and Mandarin. These multiple points of comparison allow them to identify variation between HL varieties, some of which can be attributed to cross-linguistic influence (CLI) from Mandarin. We also see studies adopting cross-generational comparisons to ascertain the extent of CLI and the presence or trajectory of on-going changes. This is evident, for example, in studies that showcase data from the Heritage Language Variation and Change (HLVC) Project (Nagy Reference Nagy2011). Nagy, Tse & Stanford in chapter 14, for example, investigate Cantonese tone mergers in spontaneous speech. Focusing on tone pairs (T2/T5, T3/T6 and T4/T6) that are reported to be undergoing merger in Hong Kong (the homeland variety), they find that the HSs in Toronto not only exhibit the expected tone mergers, but that the mergers are more advanced in spontaneous speech compared to lab-elicited speech as reported in previous work. Moreover, some studies examine both languages (i.e., HL and majority language) within the same sample of HSs, such as Kirkham & Zara’s investigation of Punjabi and English lateral consonants in chapter 6 and Jones’s examination of stop consonants in American English (the HL) and Modern Hebrew in chapter 2. Interestingly, both studies find evidence of CLI from the HL to the majority language – an outcome that is typically not examined in greater detail in HL research more generally. Taken together, the studies in this volume offer a richer, more nuanced understanding of HSs’ phonetic and phonological knowledge and, more broadly, the dynamics of cross-linguistic interactions.

It was encouraging to see that the studies (whether implicitly or explicitly) frame heritage speakers as language innovators. This view is taken up, for example, by Dehé & Rommel in chapter 12 by arguing that linguistic features stemming from CLI in Icelandic are not necessarily a replacement (of their homeland counterparts) but rather an addition to the (heritage) grammatical system. Such innovations, as Rao & Polinsky remind us in the introduction, are ‘indicative of their sophisticated knowledge of both language systems rather than a sign of a lack of knowledge or a lack of native speaker competence’ (p. 2). This reminder is critical, as heritage speakers are native speakers (Rothman & Treffers-Daller Reference Rothman and Treffers-Daller2014; Kupisch & Rothman Reference Kupisch and Rothman2018) with a full, coherent grammar – as full and coherent as the grammar of their homeland (or monolingual) counterparts. HSs can innovate and propel linguistic developments; in this way, I echo Rodríguez-Ordóñez & Sainzmaza-Lecanda (Reference Rodríguez-Ordóñez and Sainzmaza-Lecanda2021: 490) in saying HSs can also be harbingers of linguistic change.

Several studies forge strong connections with various theoretical models. In chapter 5, Shafi & Hellmuth investigate English loanword stress patterns in Mirpur Pahari. HSs display significant individual variability, which may come from the tension between Mirpur Pahari and English stress rules. The authors argue this is best understood in terms of a variable grammar, which is then formalised as stochastic OT constraints. In chapter 13, Tse tests the Functional Load Hypothesis on the maintenance of and contrasts among HSs of Cantonese in Toronto. There is evidence of a merger for , but not for . Tse shows that has fewer minimal pairs than and thus a lower functional load. Functional load may therefore be a mechanism driving CLI in HLs. Nagy, Łyskawa, Moran & Urban, in chapter 15, investigate whether there is evidence of assimilation or dissimilation in the phonetic realisations of voiced and voiceless stops in Polish among HSs in Toronto. Their comparative analysis shows cross-generational differences in how the phonetic cues (e.g., closure phonation, release duration and vowel duration) map onto the laryngeal contrasts. In relation to the framework of Flege & Bohn (Reference Flege, Bohn and Wayland2021), the authors argue that both assimilatory and dissimilatory processes may be found when the different phonetic cues of phonemes are considered. In sum, orienting towards different theoretical models and testing out their predictions can offer a better understanding of how and why certain variables (or any one of their parameters) are more or less susceptible to contact-induced effects.

A recurrent observation in HL research is that HSs seem to pattern like their monolingual/homeland counterparts in some respects but also diverge from them in others (Chang Reference Chang2021). This tension between similarity and difference is also apparent in this volume. Despite a tendency for HSs to shift towards increased majority language use, the studies in this collection generally demonstrate that the phonological representations of their HLs appear to be robust and similar with monolingual/baseline norms. Hanini, Spinu, Rafat & Alkhudidi’s study of Palestinian Arabic in chapter 11 shows that HSs with early but interrupted exposure still have strong perception and production skills. These findings support previous claims (e.g., Oh et al. Reference Oh, Jun, Knightly and Terry Kit-fong2003; Au et al. Reference Au, Oh, Knightly, Jun and Romo2008) that early childhood exposure imparts a lasting advantage in HL knowledge. Similarly, Chang, in chapter 9, finds that HSs of Korean can discriminate even with non-phonemic adjustments to VOT, suggesting that HSs have acquired the Korean laryngeal categories distinctly from their English (the HSs’ majority language) counterparts (but see Rato & Machado’s study of HSs’ perception of European Portuguese voiceless stops in chapter 16 for a countervailing example). Where we do see differences, which may be CLI-driven, is in the realisations of these categories. At the segmental domain, for example, Brehmer, Kurbangulova, Steinbach & Arifulin, in chapter 7, find HSs of Russian and Polish in Germany produce fortis and lenis stops with more German-like VOT values. At the suprasegmental domain, Feldhausen & del Mar Vanrell, in chapter 8, demonstrate that HSs of Peninsular Spanish in Germany employ different information- and focus-marking strategies compared to their homeland counterparts. Taken together, while HSs are able to establish separate representations for similar categories in their HL and majority language, they are not completely independent of each other – this can lead to ‘intermediate’ or divergent patterns.

HL scholars recognise that HSs have a wide range of social experiences and linguistic ecologies. To this end, a growing number of researchers, including the ones in this volume, acknowledge that variable patterns can be influenced not only by language-internal factors (e.g., Guzzo’s study of the role of prosodic factors in rhotic production by HSs of Brazilian Veneto in chapter 17), but also by language-external ones, such as literacy skills (in Oriyama’s study of phonological transfer among HSs of Japanese in chapter 3) and university instruction (e.g., Calhoun, Seals, Alefosio & Faamanatu-Eteuati’s study of Samoan focus-marking in chapter 4). Crucially, the myriad of ways these factors interact can also lead to much individual variability – another recurring finding in this volume. Acknowledging that multiple causation (Thomason Reference Thomason and Hickey2020) might be more fruitful in explaining linguistic outcomes, the field can move towards proposing theoretical models that consider universal linguistic processes ‘under the backdrop of potentially unique social settings’ (Natvig Reference Natvig2021: 19). Likewise, we can push forth innovative methodologies that can establish how multiple factors, including experiential and socio-demographic ones, conspire to drive observed linguistic patterns. This is exemplified by Chang & Yao’s use of multiple factor analysis (MFA) in chapter 10. The added strength of the MFA approach is its capacity to provide a more accurate way of categorising speaker groups and its ability to compare multiple phonetic variables while holding the bilingual population constant. Thus, while it seems that we are one step closer to being able to formulate generalisations about the universality of underlying processes (Polinsky & Scontras Reference Polinsky and Scontras2020), we cannot lose sight of locally situated factors that lead to important differences within and across groups. This is perhaps one of the most challenging and, at the same time, exciting parts of HL research.

3. Musing

One intriguing aspect of this volume is Oriyama’s use of written data in examining phonological transfer in Japanese (chapter 3). This approach might seem unusual at first, given that phonology is a field so focused on sound patterns of languages, but Zuraw (Reference Zuraw, Kilgarriff and Baroni2006: 59) notes that ‘a language’s phonology and orthography conspire to allow phonological data to be gleaned from text’. Zuraw (Reference Zuraw, Kilgarriff and Baroni2006, Reference Zuraw2007), for example, investigates intervocalic tapping and infixation in Tagalog using a Web-derived corpus of over 20 million Tagalog words. Such analyses of Tagalog are feasible largely because of its shallow orthography: users’ written forms often accurately represent their intended pronunciations. Similarly, Oriyama claims that Japanese kana and hiragana have ‘a consistent script–sound correspondence’ (p. 66), making it possible to examine phonological phenomena based on written evidence. In this context, I found Oriyama’s approach innovative and ensuing discussions compelling.

While appealing, this approach presents numerous limitations. Writing may be heavily influenced by prescriptive knowledge and often does not fully mimic spontaneous speech. More importantly, in the context of HLs, we must contend with varied levels of community and institutional support, including pedagogical initiatives for developing writing and literacy skills, impacting how much knowledge HSs have of their HL writing system. Some (heritage) languages may not even have a writing system to begin with, which precludes the use of written evidence in phonological analyses. Thus, while the use of written evidence presents advantages, including the potential for an abundance of data that can be gathered with fewer economic constraints, this approach may not apply widely to investigations of phonological phenomena across HLs – at least for now. Perhaps more importantly, Oriyama’s study inadvertently underscores some of the inequities surrounding heritage language use and education – something that we must also consider as part of our scholarship and in the context of giving back to the community.