Many people live in multilingual societies or speak more than one language, and linguistic diversity is a normal condition of human language. To understand language structure, representations, use, and development, it is becoming increasingly imperative that the language sciences bring language contact and multilingualism to the center of linguistic research. This book is about the linguistic properties of mature heritage languages, their acquisition, how they change, and their transmission. It is an example of how heritage languages enlighten our understanding of human language, the structure of different languages, and how languages change. Until now, the focus of works on heritage languages has been on vulnerability, language loss, and incomplete acquisition of aspects of the heritage language. In this book, I attempt to veer the discussion to seeing the structural properties of heritage languages as dynamic language change in different linguistic domains (syntax, semantics, morphology, phonology, interfaces), in different sociopolitical environments, and along their lifespans. Theories of diachronic language change have postulated (but not demonstrated) that bilingual individuals drive such changes, and I examine what role, if any, heritage language learners have in this process.

Defined by their minority status and transmission across a generation, heritage languages as discussed in this book are spoken by the bilingual children of immigrant parents. A heritage language (also ethnic, minority, community, third, non-official language) is a socio-politically minority and/or minoritized language acquired as a first, or one of the first, languages in a bilingual or multilingual context. Due to the societal pressure of the majority language, the heritage language typically develops under conditions of reduced exposure and is often used less than the majority language, especially after heritage speakers start school where the majority language is used. The term heritage language was initially coined in Canada (Cummins and Danesi, Reference Cummins and Danesi1990) to refer to languages other than the aboriginal languages of native and Inuit people and the two official languages of Canada (French and English). It rapidly expanded in the United States to refer to children and young adults with an immigrant background. Despite the different labels used in different countries, the terms heritage language and heritage speakers are now common in Europe and in other countries as well (Montrul and Polinsky, Reference Montrul and Polinsky2021). Although many linguistic situations give rise to heritage languages (immigration, colonization, territorial annexation, international adoption, among others), in this book heritage speakers refer to second-generation migrants: immigrant children or children of immigrants in the United States, who were exposed to Spanish, Hindi, and Romanian from birth as a first language or together with English. Their parents, first-generation immigrants, are native speakers of Spanish, Romanian, and Hindi who grew up in their homeland in monolingual or multilingual environments and moved to the United States in adulthood.

For over two decades, my scholarship has focused on questions of heritage language acquisition and loss as an individual phenomenon. The excitement of studying heritage languages is that the study of heritage languages invites a variety of different perspectives, and I have mainly approached it as a particular case of language acquisition. Language acquisition by an individual has a starting point or initial state, a long period of development spanning several years, and an endstate or ultimate attainment when the individual reaches linguistic maturity, after which certain aspects of the grammar (e.g., sound inventory, prepositions, determiners, pronouns, verb tenses, among many) no longer develop. Other aspects of language, such as vocabulary and stylistic variation, can be added to the basic grammatical repertoire of an individual throughout their lifespan. Seen in this way, in my early work on heritage languages, I tried to tease apart the difference between incomplete or partial acquisition and native language attrition in heritage speakers, and to understand differences and similarities between the acquisition process and the linguistic outcomes of heritage speakers and second language learners. The nature of heritage speakers’ grammatical competence, with respect to which specific aspects of their syntax and morphology are fully acquired and which ones remain variable and divergent into adulthood, continues to be the object of intense investigation. However, what remains particularly unclear is how biographical and situational factors, such as age of acquisition, length of exposure, frequency of use, and access to schooling, contribute to specific patterns of structural changes and variation observed in heritage language grammars as compared with the grammars of baseline speakers.

Most recent approaches to heritage languages have focused on heritage language varieties, and their structural properties in their own right, as products of language contact, microvariation, and language change independently of the environmental and situational factors that may have led to their current form (D’Alessandro, Reference D’Alessandro2021). When I started studying heritage languages, the structural properties of heritage grammars were compared to those of “prototypical” native speakers of the language (raised in a predominantly monolingual environment), who had also been the typical comparison group in studies of child language acquisition and in studies of second language acquisition by adults and children (see Chapter 1). There is consensus today that the input that heritage speakers are exposed to is critical to understanding their grammatical knowledge, and that first-generation immigrants constitute the parental generation: as such they are the most ecologically valid baseline to study and understand heritage language grammars, rather than the variety of language spoken in the homeland. The language of first-generation immigrants who have been living in a second/majority language environment for many years has been the focus of first language (L1) attrition, a subfield within linguistics and multilingualism that seeks to document and explain subtle grammatical language change at the individual level, due to long-term bilingualism and decreased native language use, and depending on their level of integration into their host society.

With this recognition come several changes in perspective and advances in our fields: the comparisons of bilingual speakers with other bilingual speakers ceases to perpetuate the so-often problematized monolingual bilingual comparison, emphasizes similarities and continuity between the fields of heritage language acquisition and L1 attrition, and provides a unique window and microscopic view into the genesis and mechanisms of diachronic language change. If there is incipient optionality and variability already in the language of first-generation speakers, is it ignored or maximized in the heritage speaker generation? Do heritage speakers exhibit grammatical changes that do not occur in the parental generation? Do first-generation immigrants transmit directly structural changes to the heritage speakers?

In this book, I analyze the language of the parent generation and its effects on heritage language-speaking children to identify potential factors that may determine the wide range of linguistic variability found in heritage speakers. This book brings together the main findings from the project Differential Object Marking in Spanish, Hindi and Romanian Heritage Speakers funded by the National Science Foundation (BCS-0917593, ARRA 2009–2013), that I led as principal investigator (with Rakesh Bhatt and Roxana Girju, as co-principal investigators). Many individual aspects of this project have been published already over the years (Montrul, Reference Montrul, Davidson, De La Fuente and Foote2014; Montrul and Bateman, Reference Montrul and Bateman2020a,Reference Montrul, Bateman, Mardale and Montrulb; Montrul and Bhatia, Reference Bhatia, Montrul, Mardale and Montrul2020; Montrul et al., Reference Montrul, Bhatt and Bhatia2012, Reference Montrul, Gürel, Perpiñán and Judy2015, Reference Montrul and Yoon2019). with a focus on the individual languages. However, I focus on how the collective findings and the comparison of the three languages contributes to elucidate language change as an individual phenomenon in a bilingual context across the lifespan of an individual. What is the potential role of heritage language speakers in diachronic language change? The book also includes follow-up studies of the project that have not been published before.

Native Speakers, Interrupted aims to advance our understanding of heritage language development and change. The title was inspired by Language Interrupted, a book by McWhorter (Reference McWhorter2007) about language change, the process of linguistic simplification, and the agents of language change. McWhorter compares five standard languages and their sister languages (e.g., English and its Germanic sisters, Mandarin Chinese and its Altaic sisters, Persian and its early Indo-European languages sisters, Colloquial Arabic and its other Modern Arabic sisters and Malay, and its South Asian language relatives.) and argues that standard languages with many speakers have suffered a process of morphosyntactic simplification over the centuries: that is why they are interrupted. Uninterrupted, full languages with significant morphosyntactic complexity, by contrast, are those that continued their diachronic development without language contact, like Chechen (a northeast Caucasian language spoken in the Chechen Republic). McWhorter is not alone in his claim that non-native speakers, most often adult second language (L2) learners, contribute to language change of standard languages with millions of speakers (Bentz and Winter, Reference Bentz, Winter, Wichmann and Good2014; Dale and Lupyan, Reference Dale and Lupyan2012; Lupyan and Dale, Reference Lupyan and Dale2010; Thomason and Kaufman, Reference Thomason and Kaufman1988; Trudgill, Reference Trudgill, Breivik and Jahr1989, Reference Trudgill2001). Massive amounts of adult second language speakers “break down” standard languages because adults are notoriously known for being unsuccessful at learning second languages after a certain age, especially inflectional morphology, and for displaying incomplete acquisition (Thomason and Kaufman, Reference Thomason and Kaufman1988): “extensive non-native acquisition abbreviates complexity” (McWhorter, Reference McWhorter2007: p. 52).

Although some have argued that native speaking children are the agents of diachronic language change (Lightfoot, Reference Lightfoot1991), the accepted view in the study of language change since Weinreich (Reference Weinreich1953) has been that contact-induced innovations are introduced into a given language by speakers who have some degree of competence in two or more linguistic systems. Meisel et al. (Reference Meisel2013), also attribute a specific role to post-puberty second language (L2) learners, either as agents of diachronic change or as those who provide the input containing the structural triggers for parameter settings, differing from those grammars underlying the speech of the L1 learners’ parent generation.

In the present book, I argue that heritage language speakers also qualify as potential agents of diachronic language change of the diasporic variety of their language in the language contact situation. Heritage speakers, like all speakers, are early bilinguals born with the cognitive ability to learn two or more languages fully and indeed retain native ability in specific grammatical areas of the heritage language due to their early exposure to the language. They are like native speakers because they were exposed to their home language from birth and grew up in a family environment where the language was spoken. They also learned the language implicitly, in a naturalistic setting. However, insufficient input and infrequent use of the heritage language during late childhood and adolescence often truncates – interrupts – the healthy development of the heritage language, profoundly affecting heritage speakers’ command of specific aspects of their grammar, such as vocabulary, morphosyntax, and other linguistic interfaces. What is interrupted in this case is not the language as a whole, as Darcy (Reference Darcy1953), Jespersen (Reference Jespersen1922), and McWhorter (Reference McWhorter2007) envisioned, but the individual language acquisition process itself, so that specific aspects of the heritage language, in some individuals, in some languages, and under some circumstances, show significant synchronic variability. Whereas McWhorter focuses on language and language change at the macro sociolinguistic level, my study is more in line with the approach taken by Meisel, Elsig, and Rinke (Reference Meisel, Elsig and Rinke2013), who link language acquisition and language change more directly, focusing on psycholinguistic and sociolinguistic aspects of language representation, use, and change of a specific linguistic phenomenon at the level of individuals. This book also considers the relationship between social and cognitive factors and timing in language acquisition, bilingualism, and language change.

The language of heritage speakers provides a unique testing case for identifying aspects of language that are particularly vulnerable to diachronic change. Indeed, a focus on a special group of native speakers allows me to address a fundamental issue at the heart of our understanding of language acquisition and bilingualism: how long does it take for a native language to develop and solidify (i.e., get entrenched), or cease to be flexible, adaptable, and prone to change with increased exposure to, and use of, a second language, under cognitive and social pressure? With these speakers, we can also analyze those social and cultural factors that affect the outcome of the language acquisition process from childhood into adulthood, as I do in the present book.

To accomplish these aims, I present major results from a comprehensive study of Spanish, Hindi, and Romanian heritage speakers’ knowledge of the linguistic phenomenon of Differential Object Marking (DOM). The in-depth, inter-generational, and comparative method I employ in this study produced the important finding that DOM is a vulnerable phenomenon in heritage speakers of Spanish, Hindi, and Romanian and in first-generation immigrants to different degrees. In Hindi and in Romanian, the results show broken transmission of DOM from the first to the second generation. The first generation, like the homeland native speakers tested in India and Romania, exhibit categorical command of DOM, while several speakers in the second generation show measurable variability. DOM is “lost in transmission” in these groups. However, Spanish as a heritage language in the US shows a different pattern: the first generation differs from the native speakers in Mexico, suggesting that the immigrants may have undergone attrition, and there are similarities between the first- and second-generation Spanish speakers in the United States. These results are replicated in a smaller-scale study with a group of other Spanish speakers from other Latin American countries. This finding is not consistent with broken transmission and could be interpreted as “direct transmission” of a contact variety already changed from the first to the second generation. Instead, I consider a different possibility: namely, that the erosion of DOM in the heritage speakers of Spanish spread and diffused to the first generation. I argue that second-generation heritage speakers can contribute to change in the grammar of the parental generation through their variable command of particular grammatical features, and thus lead the creation of a new variety of Spanish spoken in the United States. The results support the conclusion that transfer from English, insufficient input, and use of the heritage language by the second generation, potential attrition in the first generation, and other language specific factors appear to drive morphosyntactic erosion in Spanish, Hindi, and Romanian as heritage languages. Although McWhorter (Reference McWhorter2007) and Meisel, Elsig, and Rinke (Reference Meisel, Elsig and Rinke2013) suggest that post puberty L2 learners and their imperfect language learning capacity are the most obvious agents of diachronic language change, I make the case in this book that at least for Spanish in the United States, heritage speakers are also agents of language change, especially in the grammatical domains investigated in this book. Crucially, it is not only that the patterns of language attrition in the first-generation immigrants are transmitted to the heritage speakers: both groups converge on the same patterns through different developmental routes. If there is influence of one group over the other, I propose the direction is both ways, including from the second to the first generation as well. Follow-up studies of mothers and children led me to this conclusion.

Chapter 1 discusses the ideology and operationalization of the contentious term native speaker, because the concept of native speaker is used in the academic world and in the real world as a benchmark to judge non-native speakers, including many bilinguals. The aim of this chapter is to provide a typology of native speakers, some of whom may be “interrupted.” I cover how the notion of native speaker is constructed socially and measured psycholinguistically, while also discussing the range of variability related to education and knowledge of another language that is found in monolingually raised native speakers. Despite exhibiting high variability in degree of ultimate attainment compared to monolingually raised native speakers, I argue that heritage speakers become native speakers because they were exposed to the heritage language since birth and grew up speaking it to some extent. Also because of their early language experience with the language, they show a much higher incidence of native ability in several aspects of language, especially areas that are extremely hard for L2 learners to master at native levels, like phonology.

Chapter 2 illustrates some of the most common grammatical changes exhibited in the heritage languages from one generation to the next and presents the theoretical and circumstantial possibilities that have been argued to determine the characteristics of heritage language grammars. From the perspective of language acquisition, many of the linguistic changes and patterns exhibited by heritage speakers in several linguistic domains (but not all), most notably in morphosyntax and semantics, have been characterized as arising from incomplete, partial, or interrupted acquisition of the heritage language under reduced input conditions. But they could also be due to language attrition in childhood or later, or to changes in the linguistic input that are transmitted to the second generation. This is the idea that we test in the rest of the book.

Chapter 3 is about the specific linguistic focus of this book – DOM. It describes how DOM is manifested in Spanish, in Hindi, and Romanian at present and how the phenomena evolved diachronically with emphasis on the semantic and syntactic features that regulate DOM in each language. This chapter presents the most recent syntactic accounts of DOM in Spanish, Romanian, and Hindi that inform the linguistic analyses and interpretation of the results.

Chapter 4 reviews first, second, and bilingual acquisition studies of DOM, with a focus on which situation leads to variability. I also discuss the connection between language acquisition and language change and I motivate the research questions and hypotheses investigated in the rest of the book.

Chapter 5 describes the research questions, hypotheses, and common methodology employed in the three studies discussed in Chapters 6–8, including a description of the main instruments used to collect the data.

Chapter 6 (Spanish), Chapter 7 (Hindi), and Chapter 8 (Romanian) descriptively summarize the patterns of production and comprehension of obligatory DOM with animate, specific direct objects obtained in the three languages in the context of the general research questions and specific hypotheses formulated in Chapter 5. Each chapter begins with a brief description of the sociolinguistic situation of the language in the United States, including general characteristics of the populations, followed by a detailed summary of most of the questions in the language background questionnaire that all participants who participated in the studies completed for each language. Each chapter presents the main results from all the linguistic tasks.

Chapter 9 brings together results from the three languages to compare the nature of language change in the participant groups that participated in the studies (heritage speaker and the first-generation immigrants) and follow-up results from another study, which is discussed in Chapter 9, of Spanish in the United States. At issue is why the first-generation Spanish-speaking immigrants show language change as well (L1 attrition), like all the heritage language groups, but not the Hindi-speaking and the first-generation Romanian-speaking immigrants. We consider linguistic and situational factors that could explain the structural differences found, and the results largely support the Linguistic Niche Hypothesis (Lupyan and Dale, Reference Lupyan and Dale2010), which relates structural changes to sociopolitical changes.

Finally, Chapter 10 considers language transmission, and whether the changes observed in Spanish in the United States in the second-generation heritage speakers have been directly transmitted by the first-generation immigrants. Although a positive answer to this question is common, I present acquisition data from recent studies that challenge this directionality. By showing how age of acquisition plays a role in when changes are most likely to happen at the individual level, I argue that if one generation is influencing the other more, it is the heritage speakers who are influencing the first-generation speakers. The conclusion is that when certain social and environmental circumstances are given, heritage language speakers can also be the main agents of language change of their heritage language and can contribute to create a new variety of the original language. Changes observed in the grammars of child heritage speakers are unlikely to be directly transmitted by the parental generation.

Coming from a psycholinguistic, experimental approach, this study confirms the observation in sociolinguistics that young speakers, i.e., adolescents, contribute to language change. I consider the implications of these findings for linguistics, language change, and for language policies.