Methodological othering through monolingual controls: How not to

Chapter 4 Methodological othering through monolingual controls: How not to Artemis Alexiadou Leibniz-Zentrum Allgemeine Sprachwissenschaft (ZAS); Humboldt Universität zu Berlin In a substantial body of work on a particular group of multilingual speakers, namely heritage speakers, these are often described as deviating from native-like mastery in a variety of ways. In comparison with monolingual controls, heritage speakers are viewed as lagging behind. This paper addresses this issue and argues that monolingual controls should not be dispensed with, but rather they can be used to gain insights into language variation and the processes that shape monolingual and heritage grammars alike. 1 Introduction Typically, in empirical research participants are divided into two groups: the experimental group (the group we are interested in) and the control group. In heritage language ressearch research, the control group is composed of agedmatched monolingual speakers, and also monolingual children, and the experimental group are the heritage speakers. Thus, based on an idealized first language (L1) competence, adopting the inclusion of a control group of monolingual speakers of the language is the default in empirical research, see Vulchanova et al. (2022) for a recent discussion. However, this perspective leads to othering heritage speakers, viewing them as some sort of special group: as in most cases, such speakers do not behave the exact same way as monolinguals, the explanations that are often given are deficit oriented. The grammar of heritage speakers is non-target grammar in comparison to the target, i.e., monolingual native grammar. This has been argued to be the result of incomplete acquisition or attrition Artemis Alexiadou. 2025. Methodological othering through monolingual controls: How not to. In Artemis Alexiadou, Claudio Scarvaglieri, Christoph Schroeder & Heike Wiese (eds.), The construction of multilinguals as Others: Do we practice what we preach?, 63–80. Berlin: Language Science Press. DOI: 10.5281/zenodo.17132445 Artemis Alexiadou or both depending on the phenomenon (Montrul 2016), the result of a bottleneck (Slabakova 2019), or simply heritage grammar is in some ways defective (Polinsky & Scontras 2020). The focus on non-target properties of heritage grammars have recently led Domínguez et al. (2019: 247ff.) to defend the term incomplete acquisition. As they state, in the field of L1 acquisition, it expresses the procedural character of language acquisition: grammars may fluctuate, change, develop and show sensitivity to maturational constraints. In the field of heritage language grammars, it is suggestive of a situation whereby heritage language speakers have not had the chance to learn all the available cues. The authors explicitly conclude that ’’investigating the grammatical knowledge of heritage speakers would benefit from including at least two control groups, a group of monolingual speakers and a group of speakers representing the parental/community input that the heritage speakers receive in that context. This could help clarify whether properties not instantiated in heritage speakers’ grammars were present in the input to start with.’’ Domínguez et al. (2019: 251). Thus, from this perspective two and not just one control groups are necessary to assess the properties of heritage grammars. The so-called deficit framing of heritage speakers has been criticized from a variety of perspectives, see e.g., Kupisch & Rothman (2018), Bousquette & Putnam (2020), Higby et al. (2023), Wiese et al. (2021), Wiese et al. (2022) to mention a few. In this paper, I will align with these perspectives and propose to abandon the idea that the monolingual speaker/native speaker is the model of all different types of language acquisition and competence. My focus is on heritage speakers (henceforth HSs) of Greek. I will adopt the idea that HSs are in fact part of the native speaker range, see e.g., Rothman & Treffers-Daller (2014); while I will not argue that one needs to dispense with the notion of the control group as such, I will suggest that whenever a control group is used, the point of departure should be to show how HSs opt for patterns that are also produced by monolinguals, and thus present in the input in some form, albeit in different settings. I will attempt to identify the mechanisms the different groups of speakers employ, see Higby et al. (2023), Luk & Bialystok (2013), Wiese et al. (2021) for insightful discussion from different perspectives. I will then discuss differences between HS and monolingual productions from the perspective of representational economy and analyticity, Scontras et al. (2018), and Putnam et al. (2021). The chapter is structured as follows: in section 2, I offer a definition of heritage speakers and their relation to the notion of the native speaker. In section 3, I report on several studies that show similarities and differences in the productions of HSs and native speakers of Greek. In section 4, I turn to some discussion of the findings, and in section 5 I conclude. 64 4 Methodological othering through monolingual controls: How not to 2 Heritage speakers and the native speaker As stated in a post by Grammon & Babel (2021), ’’the idea of the “native speaker” originated within the context of European nationalism and colonialism in the 19th century. It proved useful both as a way of conceptualizing and labeling a particular linguistic identity tied to a nation and to differentiate between social groups within a colonial hierarchy. A closer examination reveals that the concept of the “native” speaker is tightly connected to discriminatory logics.’’, see also Higby et al. (2023) and references therein. Why should that be the case? As Grammon & Babel discuss, linguists and non-linguists use the term native speaker to describe an individual that grew up speaking a particular language and who is fully proficient in that language. The point is, however, that often a certain kind of authority as to how a language should be spoken is bestowed upon native speakers, thus othering multilingual speakers, see also Wiese et al. (2022) for extensive discussion and references. Typically, monolinguals are seen as native speakers of their language, as in their case one can a priori exclude language interference that may create a more complex linguistic behavior. For this reason, the monolingual speaker has been in the center of investigation in the history of theoretical and experimental linguistics. The origin of this view can be found in the early pages of Aspects of the Theory of Syntax: “Linguistic theory is concerned primarily with an ideal speaker-listener, in a completely homogeneous speech-community, who knows its language perfectly and is unaffected by such grammatically irrelevant conditions as memory limitations, distractions, shifts of attention and interest, and errors (random or characteristic) in applying his knowledge of the language in actual performance” (Chomsky 1965: 3). Thus, for a very long time, as stressed in Lohndal (2013), only monolingual speakers, treated as producing invariant speech were analyzed. This type of approach, however, leads to a situation where bilingual speakers cannot be viewed as a pure case of study, as there are too many factors interfering with the acquisition of their two grammars. The language of heritage speakers thus might be considered a rather complicated case. In other words, since there is a lot of variation among speakers in this domain, formal models of grammar cannot really be applied to the linguistic behavior of such speakers. However, we know by now that monolingual speakers do exhibit divergent outcomes, see e.g, Dąbrowska (2012) and subsequent work. Thus, if, as rightly suggested by an anonymous reviewer, such gradience is not only characteristic of bilingual populations, but is also found among monolinguals as well, the notion of the ideal speaker-hearer 65 Artemis Alexiadou requires a bit of a recalibration anyways (see Francis 2022 for discussion). Rothman et al. (2023) point out that focusing on monolingual controls has as a result that the bilingual data remain under-explored. From their perspective, it is rather unfortunate to employ this comparative aspect, especially since nowadays multilingualism is the norm across the globe. But who are heritage speaker and are they native speakers or not? According to Lohndal et al. (2019) and references therein, HSs are defined as minority language speakers in a majority language environment, they are bilinguals and by the time they are adults, they tend to be dominant in the language of their larger national community. Lack of formal education in the heritage language is associated with the low status of the heritage language. The question of whether HSs are native speakers has by now been addressed by many researchers in our field. In earlier research on HSs, we find terms such “vulnerability / resilience”, “attrition”, and “incomplete acquisition” to describe these speakers, see e.g., Montrul (2016) and Polinsky (2018) for extensive discussion and references. But clearly such terms suggest a perspective that is deficit oriented, as rightly pointed out in Higby et al. (2023), leading us to focus on aspects of heritage speakers’ grammar that are interpreted as lagging behind those of monolingual speakers, rather than the development of novel patterns, cf. the semi-lingualism controversies in the 1970s and 1980s. For instance, for Montrul (2016), HSs display characteristics of second language learners, i.e., not native like in some modules of grammar, while they maintain native language mastery in other modules of grammar. As Montrul (op.cit.) details, HSs may exhibit some sort of ‘not target-like’ development (incomplete acquisition) or they show signs of language loss/decline (attrition). These two processes have rather different developmental paths: if changes are due to attrition, then we are dealing with properties that were acquired in childhood but were later subject to language loss or gradual decline; if changes are due to incomplete acquisition, then a particular area has not been mastered yet. As further stated in Montrul (2016), a speaker may show attrition in some areas that are acquired in pre-school age (e.g., gender), and incomplete acquisition in others that take several years to develop (e.g., passives). Polinsky (2018: 28) opts for the term divergent attainment and concludes: “the crucial point is that the divergence (or innovation) in the heritage grammar is systematic. In that regard, divergent attainment is different from attrition and transfer; the latter two may be less systematic, whereas the former results in a coherent grammar.’’ The reasons for this divergent attainment maybe diverse. Rothman & Treffers-Daller (2014) have already pointed out that while most linguists would agree that the terms monolingual and native should not in principle 66 4 Methodological othering through monolingual controls: How not to be synonymous, the fact remains that linguists often use these terms interchangeably. But if we understand a native language to be acquired from naturalistic exposure, in early childhood and in an authentic social context/speech community, then the authors stress these criteria of native-hood apply to HS bilingualism as well. More recently, Wiese et al. (2021) provide several arguments as to why HSs are native speakers of both their languages. Higby et al. (2023) problematize the way such ideologies appear across frameworks and advocate multicompetence and models that allow nuance and complexity. It is precisely this nuance and complexity that is quite hard to deal with from a generative grammar perspective, see the discussion of Lohndal (2013) above. In Alexiadou (2017), I have also suggested that we should move away from focusing on what HSs cannot do, but rather examine their grammars as independent linguistic systems. In fact, my point is that what we perceive as ’odd’ patterns are changes in the grammar, which emerge whenever there is variability in the realization of particular structures, in the sense of Adger (2014). While Domínguez et al. (2019), as we have already seen, suggest that lack of cues leads to incompleteness, I will argue that nuance emerges when there is more than one option for the realization of a particular feature or structure in the grammar. In such cases, HSs may opt for one of the options driven by general mechanisms, as we will discuss in section 4. The question now is where we find variability. Register variation is a good place to look at and this what I will do next by reviewing recent case studies on heritage-Greek. The point will be that whenever there is an option between two or maybe more strategies, one of the options will be associated with a particular register and because of this perhaps prevail in the heritage grammar. By contrast, in cases the grammar leaves no option, novel patterns may emerge as a result of re-structuring, see also Putnam et al. (2021) for a summary of finding in different heritage languages. 3 Case studies I will report here on two types of studies carried out within the frame of the research unit 2537 Emerging Grammars. I will discuss three case studies involving agreement mismatches in a) restrictive relative clauses (RRCs) (Alexiadou & Rizou 2022), b) gender agreement (Alexiadou et al. 2021), and c) adjectival agreement (Alexiadou et al. 2023). All these involve agreement in the context of DPs. I will contrast these results with the use of periphrastic constructions instead of lexical verbs in heritage-Greek (Alexiadou & Rizou 2023). In all those studies, 67 Artemis Alexiadou we compared HSs of Greek in the US, in Chicago and New York, to monolingual speakers of Greek residing in Athens. The studies were chosen as with respect to RRCs and periphrastic constructions, we have very good examples of options available to speakers of Greek for the realization of a particular structure that have been argued to be subject to register variation in the language. On the other hand, in the context of pronominal, gender and adjectival agreement, there is no option within the Greek grammar, adjectives, relative pronouns, and determiners all must agree with the head noun in gender and number. Moreover, there is a substantial body of work on heritage languages that shows that gender and DP internal agreement are domains that are affected by restructuring, see e.g., (Benmamoun et al. 2013, Bolonyai 2007, Montrul et al. 2012, Albirini et al. 2013, Fuchs 2019, Putnam et al. 2021 and references therein). As we will see, the results support the view that heritage grammars do not subsume to one type of re-organization only, see also Bousquette & Putnam (2020), and Putnam et al. (2021) for a similar claim, cf. Montrul (2016): HSs pursue different directions depending on the domain of investigation of the type described in Bousquette & Putnam (2020) for Heritage German in the US and in Putnam et al. (2021) for fusional languages in general, see also Scontras et al. (2018). Thus, since different directions are chosen, the question is what this can be attributed to. 3.1 Methodology The methodology we employed was the following: we elicited naturalistic data in oral and written modality in two distinct communication settings from the aforementioned groups of speakers. This methodology is referred to as “Language situations” methodology, (see Wiese 2020). This set-up provides comparable data in both the heritage and majority language in two different levels of formality and in two modalities (data sets 2x2). The data were elicited by native speakers of Greek and American English paying attention to the respective setting. Four communication settings were created (formal spoken, formal written, informal spoken, informal written). Half of our HSs were tested first in their heritage language and afterwards in their majority while the other half of our HSs were tested first in their majority language and then in their heritage. This particular design prevents our participants from biases. Monolingual controls took part only in one language. A short video (00:42 minutes) of a fictional event was shown to every participant. A non-severe car accident was taking place in a parking lot and the task was to retell what happened to a friend and a police officer respectively imagining that they witnessed the incident. They had to produce both an oral 68 4 Methodological othering through monolingual controls: How not to (voice messages) and a written narration (whats app and written police report) in two distinct communication settings. 3.2 Restrictive relative clauses in heritage-Greek It has been reported that HSs of several languages encounter difficulties with relative clauses, potentially related to processing difficulties, see Polinsky (2018) for an extensive summary. Alexiadou & Rizou (2022) examined how HSs of Greek in the US would use RRCs and whether Greek RRCs would show signs of restructuring. The research question was motivated by discussion in the literature, according to which in both languages there is variation with respect to the realization of the form of the relative pronoun, arguably conditioned by formality. Greek RRCs appear post-nominally and are introduced either by a) the relative pronoun o opios lit. ‘the who’ which agrees in gender and number with the nominal head it modifies or b) the complementizer pu ‘that’, which is an un-inflected element. It has been argued that in Greek, the choice between pu and the agreeing form is regulated by modality and formality. Specifically, Holton et al. (1997) signal that Greek RRCs introduced by pu appear predominantly in informal, while o opios RRCs in formal registers.Greek RRCs appear post-nominally and are introduced either by a) the relative pronoun o opios lit. ‘the who’ which agrees in gender and number with the nominal head it modifies or b) the complementizer pu [that], which is an un-inflected element. It has been argued that in Greek, the choice between pu and the agreeing form is regulated by modality and formality. Specifically, Holton et al. (1997) signal that Greek RRCs introduced by pu appear predominantly in informal, while o opios RRCs in formal registers. English RRCs in the standard language are introduced by either a) a whelement, who for persons, and which for animals/things, both non-agreeing, or b) the complementizer that. There is also a third option, the zero form, which Greek lacks, e.g., This is the house which/that/Ø I told you about, from Guy & Bayley (1995: 148). Guy & Bayley (1995) report that the choice of form shows typical patterns of socio-linguistic variation and depends on the type of antecedent (reference to humans), the mode of communication (speech vs. written) and the adjacency of the antecedent. Specifically, they point out that there seems to be an inhibition towards using that to refer to humans. Wh-elements are used mainly in formal writing, while that is favoured in informal settings. In theory then, pu and that seem to have similar distributional patterns. In Alexiadou & Rizou (2022), we found out that Greek monolingual controls produced numerically more pu RRCs compared to HSs, see table 1 from Alexiadou & Rizou (2022: 137). However, contrary to our expectation and Holton et al. (1997), 69 Artemis Alexiadou both groups used more pu RRCs in the formal register. Modality did not seem to affect the use of pu: Table 1: Production of pu RRCs across registers and modalities HSs Monolingual controls Formal spoken 135 123 Formal written 93 129 Informal spoken 101 111 Informal written 61 85 390 448 With respect to o opios RRCs, in Alexiadou & Rizou (2022), we noted that monolingual speakers of Greek produce quantitatively more such clauses in contrast to heritage speakers (see table 2, from Alexiadou & Rizou 2022: 138). Monolingual controls seemed to favor the production of this type of clauses in the formal register, while modality again did not seem as influential. Table 2: Production of o opios RRCs across registers and modalities HSs Monolingual controls Formal spoken 3 88 Formal written 3 61 Informal spoken 0 39 Informal written 2 23 8 211 We concluded that pu is not uniquely associated with informal settings, across groups. For the HSs, o opios RRCs are avoided, and we attributed this to the fact that the pronoun agrees in phi-features (gender and number) with the head noun it modifies. As we will see in the next case study, Greek HSs in the US avoid producing agreement dependencies within the NP (Alexiadou et al. 2021), thus, as a result, RRCs introduced by the pronoun o opios are avoided so that our speakers do not find themselves in a position, where they need to track agreement. This seems to support e.g., Slabakova’s (2019) view that agreement constitutes a potential ’bottleneck’. This is further corroborated by the fact that, 70 4 Methodological othering through monolingual controls: How not to as we showed in Alexiadou & Rizou (2022), they have no problem using who and which forms that do not exhibit any (gender/number) agreement in their English productions. In conclusion, in this case study a particular structure is avoided, and a less complex form, which is void of phi features, is used. 3.3 Gender agreement in heritage-Greek The above results are correlated with the observations made in Alexiadou et al. (2021) concerning gender agreement in US Greek, which we conducted with the same group of speakers. We noted that some of our speakers fail to assign gender to Greek nouns, Greek being a three-gender language, either by avoiding using the determiner or by resorting to default neuter. Not surprisingly from the perspective of the study on RRCs the most affected category in terms of agreement was the relative pronoun, where basically the form was rarely produced, as detailed in the previous section. Importantly, however, in Alexiadou et al. (2021) we found no correlation with formality or modality and avoidance of gender assignment. While it can be argued that interference with English, a language that lacks grammatical gender, is the culprit, a surprising finding was the inconsistency in patterns of DP internal agreement. Since, as discussed in Polinsky (2018), gender is a category subject to re-structuring in heritage grammars, the Greek data further align with results on other languages, which show that the processing of agreement within noun phrases signals a processing difficulty. Monolinguals do not show any gender agreement mismatches. 3.4 Adjectival agreement in heritage-Greek The fact that US speakers of heritage-Greek have problems with DP internal agreement was further corroborated by the results in Alexiadou et al. (2023), where we found differences in the production of mismatches between monolinguals and US HSs: prenominal adjectival agreement was indeed more vulnerable than subject verb agreement. Specifically, adolescent speakers in the US show 17.5% of errors. For those speakers that produced errors, we further noted that some of them repeated the mismatched agreement structure in the second narration and then avoided using the same structure. Others, however, produced a mismatch in their first narration and then produced the correct form in the following narrations. The fact that we do find errors DP internally, points to the same conclusion: once more, these results align with previous literature pointing to issues with agreement within the nominal domain. Monolinguals do not show any agreement mismatches. 71 Artemis Alexiadou Gharibi, Khadij, Fatih Bayram & Gustavo Guajardo. 2023. Lexical and morphosyntactic variation in Persian heritage language outcomes. Linguistic Approaches to bilingualism 14(6). 886–914. DOI: 10.1075/lab.21052.gha. Grammon, Devin & Anna Babel. 2021. What does native speaker mean anyway? Source: Language Log blog. Blog post. https://languagelog.ldc.upenn.edu/nll/ ?p=51108. Guy, Gregory R. & Robert Bayley. 1995. On the choice of relative pronouns in English. American Speech 70(2). 148–162. Higby, Eve, Evelyn Gámez & Claudia Holguín Mendoza. 2023. Challenging deficit frameworks in research on heritage language bilingualism. Applied Psycholinguistics 44. 417–430. DOI: 10.1017/S0142716423000048. Holton, David, Peter Mackridge & Irene Philippaki-Warburto. 1997. Greek: A comprehensive grammar of the modern language. London: Routledge. Kupisch, Tanja & Jason Rothman. 2018. Terminology matters! Why difference is not incompleteness and how early child bilinguals are heritage speakers. International Journal of Bilingualism 22(5). 564–582. DOI: 10 . 1177 / 1367006916654355. Lohndal, Terje. 2013. Generative grammar and language mixing. Theoretical Linguistics 39(3–4). 215–224. DOI: 10.1515/tl-2013-0013. Lohndal, Terje & Michael T. Putnam. 2024. The importance of features and exponents: Dissolving feature reassembly. Linguistic Approaches to Bilingualism 14(1). 1–36. DOI: 10.1075/lab.23023.loh. Lohndal, Terje, Jason Rothman, Tanja Kupisch & Marit Westergaard. 2019. Heritage language acquisition: What it reveals and why it is important for formal linguistic theories. Language and Linguistics Compass 13(12). e12357. DOI: 10.1111/lnc3.12357. Luk, Gigi & Ellen Bialystok. 2013. Bilingualism is not a categorial variable: Interaction between language proficiency and usage. Journal of Cognitive Psychology 25(5). 605–621. DOI: 10.1080/20445911.2013.795574. Montrul, Silvina. 2016. The acquisition of heritage languages. Cambridge: Cambridge University Press. DOI: 10.1017/CBO9781139030502. Montrul, Silvina, Rakesh Mohan Bhatt & Archna Bhatia. 2012. Erosion of case and agreement in Hindi heritage speakers. Linguistic Approaches to Bilingualism 2(2). 141–76. DOI: 10.1075/lab.2.2.02mon. Polinsky, Maria. 2018. Heritage languages and their speakers. Cambridge: Cambridge University Press. DOI: 10.1017/9781107252349. Polinsky, Maria & Gregory Scontras. 2020. Understanding heritage languages. Bilingualism: Language and Cognition 23(1). 4–20. DOI: 10 . 1017 / S1366728919000245. 78 4 Methodological othering through monolingual controls: How not to Putnam, Michael T., Lara Schwarz & Andrew D. Hoffman. 2021. Morphology of heritage languages. In Silvina Montrul & Maria Polinsky (eds.), The Cambridge Handbook of Heritage Languages and Linguistics, 613–643. Cambridge: Cambridge University Press. Ralli, Angela. 2004. Stem-based vs word-based morphological configurations: The case of Modern Greek preverbs. Lingue e Linguaggio 3(2). 241–275. Rothman, Jason, Fatih Bayram, Vincent DeLuca, Grazia Di Pisa, Jon Andoni Duñabeitia, Khadij Gharibi, Jiuzhou Hao, Nadine Kolb, Maki Kubota, Tanja Kupisch, Tim Laméris, Alicia Luque, Brechje van Osch, Sergio Miguel Pereira Soares, Yanina Prystauka, Deniz Tat, Aleksandra Tomić, Toms Voits & Stefanie Wulff. 2023. Monolingual comparative normativity in bilingualism research is out of “control”: Arguments and alternatives. Applied Psycholinguistics 44(3). 316–329. DOI: 10.1017/S0142716422000315. Rothman, Jason & Jeanine Treffers-Daller. 2014. A prolegomenon to the construct of the native speaker: Heritage speaker bilinguals are natives too! Applied Linguistics 35(1). 93–98. DOI: 10.1093/applin/amt049. Scontras, Gregory, Maria Polinsky & Zuzanna Fuchs. 2018. In support of representational economy: Agreement in heritage Spanish. Glossa: A journal of general linguistics 3(1). 1–29. Slabakova, Roumyana. 2019. The bottleneck hypothesis updated. In T. Ionin & M. Rispoli (eds.), Three streams of generative language acquisition research: Selected papers from the 7th GALANA meeting, 319–346. Amsterdam: John Benjamins. DOI: 10.1075/lald.63.16sla. Vulchanova, Mila, Valentin Vulchanov, Antonella Sorace, Cristina Suarez-Gomez & Pedro Guijarro-Fuentes. 2022. Editorial: The notion of the native speaker put to the test: Recent research advances. Frontiers in Psychology 13. DOI: 10.3389/ fpsyg.2022.875740. Wiese, Heike. 2020. Language situations: A method for capturing variation within speakers’ repertoires. In Asahi Yoshiyuki (ed.), Methods in dialectology XVI, 105–117. München: Peter Lang. Wiese, Heike, Artemis Alexiadou, Shanley Allen, Oliver Bunk, Natalia Gagarina, Kateryna Iefremenko, Maria Martynova, Tatiana Pashkova, Vicky Rizou, Christoph Schroeder, Anna Shadrova, Luka Szucsich, Rosemarie Tracy, Wintai Tsehaye, Sabine Zerbian & Yulia Zuban. 2021. Heritage speakers as part of the native language continuum. Frontiers in Psychology 12. Pedro Guijarro-Fuentes, Cristina Suarez-Gomez, Mila Vulchanova, Antonella Sorace & Valentin Vulchanov (eds.). 209–227. DOI: 10.3389/fpsyg.2021.717973. 79 Artemis Alexiadou Wiese, Heike, Artemis Alexiadou, Claudio Scarvaglieri & Christoph Schroeder. 2022. Multilinguals as Others in society and academia: Challenges of belonging under a monolingual habitus. Working Papers in Urban Language & Literacies 302. https://wpull.org/wp-content/uploads/2022/10/WP302-Wiese-et-al2022-Multilinguals-as-others.pdf. 80